{"id": "arxiv:1602.02261", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67f8697c21dabc2c4781eba46bfb60671506b413471efcdb28864c11649b3a2c", "submitted_at": "2016-02-06T14:53:02Z", "updated_at": "2016-05-20T16:26:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1602.02261v2"}, "authors": ["Rodrigo Nogueira", "Kyunghyun Cho"], "canonical_source_url": "https://arxiv.org/abs/1602.02261", "citation_url": "https://awesomecua.github.io/resources/arxiv-1602-02261/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67f8697c21dabc2c4781eba46bfb60671506b413471efcdb28864c11649b3a2c", "submitted_at": "2016-02-06T14:53:02Z", "updated_at": "2016-05-20T16:26:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1602.02261v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:46Z", "role": "primary", "sha256": "5f99bcea7c1585129af46fa5c978002778dedc3d98e3a907e92030f4a5588f00", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/1602.02261v2"}], "identifiers": {"arxiv": "1602.02261"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1602.02261"}], "name": "End-to-End Goal-Driven Web Navigation", "publication": null, "published_at": "2016-02-06", "record_sha256": "f86c73e02bd0aa44b2fff5312837e5903b8dbc16ae0e05eab9c088a6c3774e9f", "relations": [], "resource_id": "arxiv:1602.02261", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/1602.02261v2", "topics": ["web-navigation", "evaluation", "planning"], "updated_at": "2016-05-20"}, "text": "End-to-End Goal-Driven Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebNav for converting websites into goal-directed hyperlink navigation tasks and WikiNav for evaluating language-conditioned navigation and question answering on Wikipedia.\nPlatforms: web\nTopics: web-navigation, evaluation, planning\nPublished: 2016-02-06\nAuthors: Rodrigo Nogueira, Kyunghyun Cho\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/1602.02261v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-1602-02261/"}
{"id": "arxiv:1802.08802", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75d19fa97ce85a93bb9665a8ae06607bc0b90fde1c8376cdd60e99b55e90729c", "submitted_at": "2018-02-24T05:32:47Z", "updated_at": "2018-02-24T05:32:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.08802v1"}, "authors": ["Evan Zheran Liu", "Kelvin Guu", "Panupong Pasupat", "Tianlin Shi", "Percy Liang"], "canonical_source_url": "https://arxiv.org/abs/1802.08802", "citation_url": "https://awesomecua.github.io/resources/arxiv-1802-08802/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75d19fa97ce85a93bb9665a8ae06607bc0b90fde1c8376cdd60e99b55e90729c", "submitted_at": "2018-02-24T05:32:47Z", "updated_at": "2018-02-24T05:32:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/1802.08802v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "1802.08802"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1802.08802"}], "name": "Reinforcement Learning on Web Interfaces Using Workflow-Guided Exploration", "publication": null, "published_at": "2018-02-24", "record_sha256": "8e68e80f6e809fbc7f0d4ab55042c7758019e4f49b46fc91c996f305204783f0", "relations": [], "resource_id": "arxiv:1802.08802", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/1802.08802v1", "topics": ["reinforcement-learning", "workflow"], "updated_at": "2018-02-24"}, "text": "Reinforcement Learning on Web Interfaces Using Workflow-Guided Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nUses workflows extracted from demonstrations to constrain reinforcement-learning exploration for interactive web tasks.\nPlatforms: web\nTopics: reinforcement-learning, workflow\nPublished: 2018-02-24\nAuthors: Evan Zheran Liu, Kelvin Guu, Panupong Pasupat, Tianlin Shi, Percy Liang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/1802.08802v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-1802-08802/"}
{"id": "arxiv:1808.09132", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f368116f70e987cde111835a05b5b36e0d345dabb4e8351610e2bbfb691b3440", "submitted_at": "2018-08-28T06:09:39Z", "updated_at": "2018-10-01T03:22:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1808.09132v2"}, "authors": ["Panupong Pasupat", "Tian-Shun Jiang", "Evan Zheran Liu", "Kelvin Guu", "Percy Liang"], "canonical_source_url": "https://arxiv.org/abs/1808.09132", "citation_url": "https://awesomecua.github.io/resources/arxiv-1808-09132/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f368116f70e987cde111835a05b5b36e0d345dabb4e8351610e2bbfb691b3440", "submitted_at": "2018-08-28T06:09:39Z", "updated_at": "2018-10-01T03:22:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1808.09132v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "1808.09132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1808.09132"}], "name": "Mapping Natural Language Commands to Web Elements", "publication": null, "published_at": "2018-08-28", "record_sha256": "9801b6dc68f704d00437dc919fa03a4d57f34a7df1d63a240b86f405e9e46d66", "relations": [], "resource_id": "arxiv:1808.09132", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/1808.09132v2", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2018-10-01"}, "text": "Mapping Natural Language Commands to Web Elements\nResource type: paper\nReview status: curated\nScope: core\n\nDefines language grounding on webpages as choosing the UI element targeted by a natural-language command. A collection of more than 50,000 commands and three baseline models study functional references, relationships between elements, and visual position cues.\nPlatforms: web\nTopics: grounding, data-generation, evaluation\nPublished: 2018-08-28\nAuthors: Panupong Pasupat, Tian-Shun Jiang, Evan Zheran Liu, Kelvin Guu, Percy Liang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/1808.09132v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-1808-09132/"}
{"id": "arxiv:1909.00031", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4cef465f53abcbaad50c83f24e0ce54fb7f23390955e9f7e5592ec0b8287839e", "submitted_at": "2019-08-30T18:35:01Z", "updated_at": "2020-01-07T00:57:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1909.00031v2"}, "authors": ["Toby Jia-Jun Li", "Marissa Radensky", "Justin Jia", "Kirielle Singarajah", "Tom M. Mitchell", "Brad A. Myers"], "canonical_source_url": "https://arxiv.org/abs/1909.00031", "citation_url": "https://awesomecua.github.io/resources/arxiv-1909-00031/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4cef465f53abcbaad50c83f24e0ce54fb7f23390955e9f7e5592ec0b8287839e", "submitted_at": "2019-08-30T18:35:01Z", "updated_at": "2020-01-07T00:57:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/1909.00031v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:46Z", "role": "primary", "sha256": "ee42fc28e088432e79837641c310dbd2d72f260a2a01d86f5f7f8b9c1b544a39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/1909.00031v2"}], "identifiers": {"arxiv": "1909.00031"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/1909.00031"}], "name": "Interactive Task and Concept Learning from Natural Language Instructions and GUI Demonstrations", "publication": null, "published_at": "2019-08-30", "record_sha256": "f43c1273f7d55aab61cf4db5854032a1742697227814a89bac2dbef1cf95ed90", "relations": [], "resource_id": "arxiv:1909.00031", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/1909.00031v2", "topics": ["human-in-the-loop", "mobile-navigation", "imitation-learning"], "updated_at": "2020-01-07"}, "text": "Interactive Task and Concept Learning from Natural Language Instructions and GUI Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PUMICE, an agent that learns mobile-app procedures and conditional concepts through natural-language conversation and GUI demonstrations, recursively resolving ambiguous instructions with users.\nPlatforms: unspecified\nTopics: human-in-the-loop, mobile-navigation, imitation-learning\nPublished: 2019-08-30\nAuthors: Toby Jia-Jun Li, Marissa Radensky, Justin Jia, Kirielle Singarajah, Tom M. Mitchell, Brad A. Myers\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/1909.00031v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-1909-00031/"}
{"id": "arxiv:2005.03776", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af7b45455ffbd616e8cdf5612b19275ecb7bf07b1065af9f5b724ae19e96fb4", "submitted_at": "2020-05-07T21:41:40Z", "updated_at": "2020-06-05T02:11:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2005.03776v2"}, "authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "canonical_source_url": "https://arxiv.org/abs/2005.03776", "citation_url": "https://awesomecua.github.io/resources/arxiv-2005-03776/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af7b45455ffbd616e8cdf5612b19275ecb7bf07b1065af9f5b724ae19e96fb4", "submitted_at": "2020-05-07T21:41:40Z", "updated_at": "2020-06-05T02:11:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2005.03776v2"}, "publication": {"authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "date_basis": "Explicit publisher metadata: 2020/7; precision=month.", "date_parts": [2020, 7], "doi": "10.18653/v1/2020.acl-main.729", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.acl-main.729/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "8198-8210", "pdf_url": "https://aclanthology.org/2020.acl-main.729.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "type": "conference-paper", "url": "https://aclanthology.org/2020.acl-main.729/", "venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "c196294e19378d08d71a7cfd41c2e368f5a6b49ee7c3601257537b9ae5c42add", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.acl-main.729/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2005.03776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2005.03776"}, {"rel": "publication", "url": "https://aclanthology.org/2020.acl-main.729/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.acl-main.729.pdf"}], "name": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "publication": {"authors": ["Yang Li", "Jiacong He", "Xin Zhou", "Yuan Zhang", "Jason Baldridge"], "date_basis": "Explicit publisher metadata: 2020/7; precision=month.", "date_parts": [2020, 7], "doi": "10.18653/v1/2020.acl-main.729", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.acl-main.729/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "8198-8210", "pdf_url": "https://aclanthology.org/2020.acl-main.729.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Mapping Natural Language Instructions to Mobile UI Action Sequences", "type": "conference-paper", "url": "https://aclanthology.org/2020.acl-main.729/", "venue": "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics", "volume": null}, "published_at": "2020-05-07", "record_sha256": "95eb6614cbedee6c08b36740a4060735f90e0e43b35264941f52b9cdf335d0b4", "relations": [], "resource_id": "arxiv:2005.03776", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2005.03776v2", "topics": ["mobile-navigation", "grounding", "data-generation", "action-prediction"], "updated_at": "2020-06-05"}, "text": "Mapping Natural Language Instructions to Mobile UI Action Sequences\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PIXELHELP for predicting mobile UI action sequences from instructions, together with action-phrase annotations and synthetic UI action descriptions for training. Separate Transformers extract instruction steps and ground object descriptions using interface content and screen position.\nPlatforms: unspecified\nTopics: mobile-navigation, grounding, data-generation, action-prediction\nPublished: 2020-05-07\nAuthors: Yang Li, Jiacong He, Xin Zhou, Yuan Zhang, Jason Baldridge\nRecorded arXiv version: v2\n\nPublication edition: Mapping Natural Language Instructions to Mobile UI Action Sequences\nPublication authors: Yang Li, Jiacong He, Xin Zhou, Yuan Zhang, Jason Baldridge\nPublication edition date (recorded precision): 2020-07\nPublication venue: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics\nPublication publisher: Association for Computational Linguistics\nPublication pages: 8198-8210\nPublication doi: 10.18653/v1/2020.acl-main.729\nPublication date basis: Explicit publisher metadata: 2020/7; precision=month.\nPublication notes: Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://aclanthology.org/2020.acl-main.729/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2005.03776v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2005-03776/"}
{"id": "arxiv:2008.05132", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6352985532945fd562ce0382aef0e977ce1ec5909b6872e5fe048167e9732a0", "submitted_at": "2020-08-12T06:36:33Z", "updated_at": "2020-09-07T12:57:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2008.05132v2"}, "authors": ["Jieshan Chen", "Mulong Xie", "Zhenchang Xing", "Chunyang Chen", "Xiwei Xu", "Liming Zhu", "Guoqiang Li"], "canonical_source_url": "https://arxiv.org/abs/2008.05132", "citation_url": "https://awesomecua.github.io/resources/arxiv-2008-05132/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6352985532945fd562ce0382aef0e977ce1ec5909b6872e5fe048167e9732a0", "submitted_at": "2020-08-12T06:36:33Z", "updated_at": "2020-09-07T12:57:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2008.05132v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2008.05132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2008.05132"}], "name": "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?", "publication": null, "published_at": "2020-08-12", "record_sha256": "7731d93e5abe0b67cb10ed4e7a0a3a7c306dd8b8f46a9ce993cf4961bf3171b8", "relations": [], "resource_id": "arxiv:2008.05132", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2008.05132v2", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2020-09-07"}, "text": "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?\nResource type: paper\nReview status: curated\nScope: core\n\nCompares seven traditional and learned GUI element detectors on more than 50,000 interface images. The analysis motivates a coarse-to-fine non-text detector combined with a learned text detector to address GUI-specific localization requirements.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, evaluation\nPublished: 2020-08-12\nAuthors: Jieshan Chen, Mulong Xie, Zhenchang Xing, Chunyang Chen, Xiwei Xu, Liming Zhu, Guoqiang Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2008.05132v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2008-05132/"}
{"id": "arxiv:2010.04295", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68af0a0a4b605d0c968e8913232fde16e3eadf6361099e8fc3d015dcdb26302b", "submitted_at": "2020-10-08T22:56:03Z", "updated_at": "2020-10-08T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2010.04295v1"}, "authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "canonical_source_url": "https://arxiv.org/abs/2010.04295", "citation_url": "https://awesomecua.github.io/resources/arxiv-2010-04295/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68af0a0a4b605d0c968e8913232fde16e3eadf6361099e8fc3d015dcdb26302b", "submitted_at": "2020-10-08T22:56:03Z", "updated_at": "2020-10-08T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2010.04295v1"}, "publication": {"authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "date_basis": "Explicit publisher metadata: 2020/11; precision=month.", "date_parts": [2020, 11], "doi": "10.18653/v1/2020.emnlp-main.443", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.emnlp-main.443/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The retained arXiv abstract reports 162,859 phrases; the ACL abstract reports 162,860. This publication proposal does not overwrite the arXiv annotation or assert a unified dataset count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "5495-5510", "pdf_url": "https://aclanthology.org/2020.emnlp-main.443.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "type": "conference-paper", "url": "https://aclanthology.org/2020.emnlp-main.443/", "venue": "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "750e60b4d28e5b97a46595ea2280b520bf9edcc445b9931bad225e0028935c5e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2010.04295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2010.04295"}, {"rel": "publication", "url": "https://aclanthology.org/2020.emnlp-main.443/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2020.emnlp-main.443.pdf"}], "name": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "publication": {"authors": ["Yang Li", "Gang Li", "Luheng He", "Jingjie Zheng", "Hong Li", "Zhiwei Guan"], "date_basis": "Explicit publisher metadata: 2020/11; precision=month.", "date_parts": [2020, 11], "doi": "10.18653/v1/2020.emnlp-main.443", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2020.emnlp-main.443/"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The retained arXiv abstract reports 162,859 phrases; the ACL abstract reports 162,860. This publication proposal does not overwrite the arXiv annotation or assert a unified dataset count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "5495-5510", "pdf_url": "https://aclanthology.org/2020.emnlp-main.443.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements", "type": "conference-paper", "url": "https://aclanthology.org/2020.emnlp-main.443/", "venue": "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)", "volume": null}, "published_at": "2020-10-08", "record_sha256": "94a14799f767a59f5075eded64f7643ea8fba7f605eb92a91f29c4b859ea6d99", "relations": [], "resource_id": "arxiv:2010.04295", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2010.04295v1", "topics": ["screen-understanding", "grounding"], "updated_at": "2020-10-08"}, "text": "Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces crowdsourced descriptions of mobile UI elements and studies multimodal models that generate useful widget captions.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\nPublished: 2020-10-08\nAuthors: Yang Li, Gang Li, Luheng He, Jingjie Zheng, Hong Li, Zhiwei Guan\nRecorded arXiv version: v1\n\nPublication edition: Widget Captioning: Generating Natural Language Description for Mobile User Interface Elements\nPublication authors: Yang Li, Gang Li, Luheng He, Jingjie Zheng, Hong Li, Zhiwei Guan\nPublication edition date (recorded precision): 2020-11\nPublication venue: Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 5495-5510\nPublication doi: 10.18653/v1/2020.emnlp-main.443\nPublication date basis: Explicit publisher metadata: 2020/11; precision=month.\nPublication notes: The retained arXiv abstract reports 162,859 phrases; the ACL abstract reports 162,860. This publication proposal does not overwrite the arXiv annotation or assert a unified dataset count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://aclanthology.org/2020.emnlp-main.443/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2010.04295v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2010-04295/"}
{"id": "arxiv:2012.12350", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51fdddb9a3d8a6090875719a51b22008b1f43bc38c6cfe1713a49a52f7e8434c", "submitted_at": "2020-12-22T20:49:52Z", "updated_at": "2021-01-25T20:37:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2012.12350v2"}, "authors": ["Zecheng He", "Srinivas Sunkara", "Xiaoxue Zang", "Ying Xu", "Lijuan Liu", "Nevan Wichers", "Gabriel Schubiner", "Ruby Lee", "Jindong Chen", "Blaise Agüera y Arcas"], "canonical_source_url": "https://arxiv.org/abs/2012.12350", "citation_url": "https://awesomecua.github.io/resources/arxiv-2012-12350/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51fdddb9a3d8a6090875719a51b22008b1f43bc38c6cfe1713a49a52f7e8434c", "submitted_at": "2020-12-22T20:49:52Z", "updated_at": "2021-01-25T20:37:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2012.12350v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "139375513dafe5d7cfd0f5843b2a3465294e2b54051cf1c964a0d0751166a469", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2012.12350v2"}], "identifiers": {"arxiv": "2012.12350"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2012.12350"}], "name": "ActionBert: Leveraging User Actions for Semantic Understanding of User Interfaces", "publication": null, "published_at": "2020-12-22", "record_sha256": "bfd9d45ad03f3a44eedba18e7adac45aced9f8bd0a1503182434462a9705616b", "relations": [], "resource_id": "arxiv:2012.12350", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2012.12350v2", "topics": ["screen-understanding", "grounding", "action-prediction"], "updated_at": "2021-01-25"}, "text": "ActionBert: Leveraging User Actions for Semantic Understanding of User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nPretrains ActionBert on user interaction traces to combine visual, linguistic, and UI-structure features. It evaluates the resulting UI representations on tasks including icon classification and description-based component retrieval.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, action-prediction\nPublished: 2020-12-22\nAuthors: Zecheng He, Srinivas Sunkara, Xiaoxue Zang, Ying Xu, Lijuan Liu, Nevan Wichers, Gabriel Schubiner, Ruby Lee, Jindong Chen, Blaise Agüera y Arcas\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2012.12350v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2012-12350/"}
{"id": "arxiv:2101.04893", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6a5305254ea405b6297c0a3eee43d58fc028abaa1a538d03b175b39b9daccca", "submitted_at": "2021-01-13T05:56:15Z", "updated_at": "2021-01-13T05:56:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2101.04893v1"}, "authors": ["Xiaoyi Zhang", "Lilian de Greef", "Amanda Swearngin", "Samuel White", "Kyle Murray", "Lisa Yu", "Qi Shan", "Jeffrey Nichols", "Jason Wu", "Chris Fleizach", "Aaron Everitt", "Jeffrey P. Bigham"], "canonical_source_url": "https://arxiv.org/abs/2101.04893", "citation_url": "https://awesomecua.github.io/resources/arxiv-2101-04893/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6a5305254ea405b6297c0a3eee43d58fc028abaa1a538d03b175b39b9daccca", "submitted_at": "2021-01-13T05:56:15Z", "updated_at": "2021-01-13T05:56:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2101.04893v1"}}, "environments": ["ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2101.04893"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2101.04893"}], "name": "Screen Recognition: Creating Accessibility Metadata for Mobile Applications from Pixels", "publication": null, "published_at": "2021-01-13", "record_sha256": "8ba3dbd3966fe5ba1f07a81805f4698c70686b7bc6cbeb444b57345b1b789761", "relations": [], "resource_id": "arxiv:2101.04893", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2101.04893v1", "topics": ["screen-understanding"], "updated_at": "2021-01-13"}, "text": "Screen Recognition: Creating Accessibility Metadata for Mobile Applications from Pixels\nResource type: paper\nReview status: curated\nScope: adjacent\n\nInfers UI elements and accessibility metadata from mobile screenshots to support screen-reader interaction with iOS applications.\nPlatforms: ios\nTopics: screen-understanding\nPublished: 2021-01-13\nAuthors: Xiaoyi Zhang, Lilian de Greef, Amanda Swearngin, Samuel White, Kyle Murray, Lisa Yu, Qi Shan, Jeffrey Nichols, Jason Wu, Chris Fleizach, Aaron Everitt, Jeffrey P. Bigham\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2101.04893v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2101-04893/"}
{"id": "arxiv:2101.09465", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5fdf4bc31edd633fd0858835db0afedeef5dbc2694c4511903976c338bd19b1b", "submitted_at": "2021-01-23T09:43:44Z", "updated_at": "2021-11-08T08:31:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2101.09465v2"}, "authors": ["Xingyu Chen", "Zihan Zhao", "Lu Chen", "Danyang Zhang", "Jiabao Ji", "Ao Luo", "Yuxuan Xiong", "Kai Yu"], "canonical_source_url": "https://arxiv.org/abs/2101.09465", "citation_url": "https://awesomecua.github.io/resources/arxiv-2101-09465/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5fdf4bc31edd633fd0858835db0afedeef5dbc2694c4511903976c338bd19b1b", "submitted_at": "2021-01-23T09:43:44Z", "updated_at": "2021-11-08T08:31:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2101.09465v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2101.09465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2101.09465"}], "name": "WebSRC: A Dataset for Web-Based Structural Reading Comprehension", "publication": null, "published_at": "2021-01-23", "record_sha256": "3abb63a2bcfb82fe82e437c3f3faac4ac3a1421608746545688483a138ffc766", "relations": [], "resource_id": "arxiv:2101.09465", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2101.09465v2", "topics": ["screen-understanding", "evaluation"], "updated_at": "2021-11-08"}, "text": "WebSRC: A Dataset for Web-Based Structural Reading Comprehension\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces question answering over webpage text and structure, with aligned HTML, screenshots, and metadata for structural reading comprehension.\nPlatforms: web\nTopics: screen-understanding, evaluation\nPublished: 2021-01-23\nAuthors: Xingyu Chen, Zihan Zhao, Lu Chen, Danyang Zhang, Jiabao Ji, Ao Luo, Yuxuan Xiong, Kai Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2101.09465v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2101-09465/"}
{"id": "arxiv:2103.16057", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e00b8f1b34097455ece2cf97efabbb64f398e3fffdec13394f731096672dc654", "submitted_at": "2021-03-30T04:02:34Z", "updated_at": "2021-04-04T23:51:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2103.16057v2"}, "authors": ["Nancy Xu", "Sam Masling", "Michael Du", "Giovanni Campagna", "Larry Heck", "James Landay", "Monica S Lam"], "canonical_source_url": "https://arxiv.org/abs/2103.16057", "citation_url": "https://awesomecua.github.io/resources/arxiv-2103-16057/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e00b8f1b34097455ece2cf97efabbb64f398e3fffdec13394f731096672dc654", "submitted_at": "2021-03-30T04:02:34Z", "updated_at": "2021-04-04T23:51:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2103.16057v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2103.16057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2103.16057"}], "name": "Grounding Open-Domain Instructions to Automate Web Support Tasks", "publication": null, "published_at": "2021-03-30", "record_sha256": "1e8cc9a0efb8876e5bd4b4101428b142cda49eb56652ce8dafd0058dffd06bb9", "relations": [], "resource_id": "arxiv:2103.16057", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2103.16057v2", "topics": ["grounding", "web-navigation"], "updated_at": "2021-04-04"}, "text": "Grounding Open-Domain Instructions to Automate Web Support Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces RUSS, which translates support instructions into a domain-specific language and grounds them in actionable web elements.\nPlatforms: web\nTopics: grounding, web-navigation\nPublished: 2021-03-30\nAuthors: Nancy Xu, Sam Masling, Michael Du, Giovanni Campagna, Larry Heck, James Landay, Monica S Lam\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2103.16057v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2103-16057/"}
{"id": "arxiv:2105.11941", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8788e419d1a901f8ad9dd6089a12fe129baee764b1c285bbeb4ea60d7046afa2", "submitted_at": "2021-05-25T13:45:54Z", "updated_at": "2022-04-14T11:00:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2105.11941v2"}, "authors": ["Jingwen Fu", "Xiaoyi Zhang", "Yuwang Wang", "Wenjun Zeng", "Sam Yang", "Grayson Hilliard"], "canonical_source_url": "https://arxiv.org/abs/2105.11941", "citation_url": "https://awesomecua.github.io/resources/arxiv-2105-11941/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8788e419d1a901f8ad9dd6089a12fe129baee764b1c285bbeb4ea60d7046afa2", "submitted_at": "2021-05-25T13:45:54Z", "updated_at": "2022-04-14T11:00:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2105.11941v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "41c4da71e457ea23725a8070b0da64c234f03613841b95edea3a5f3c95dd8c68", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2105.11941v2"}], "identifiers": {"arxiv": "2105.11941"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2105.11941"}], "name": "Understanding Mobile GUI: from Pixel-Words to Screen-Sentences", "publication": null, "published_at": "2021-05-25", "record_sha256": "da36b1613c7d1c2b644d40d0a8d067cfa95d07cc06fb834eb46f1dacf1fa2db3", "relations": [], "resource_id": "arxiv:2105.11941", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2105.11941v2", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-04-14"}, "text": "Understanding Mobile GUI: from Pixel-Words to Screen-Sentences\nResource type: paper\nReview status: curated\nScope: core\n\nModels mobile screens as atomic text and graphic components whose relationships are learned by a Screen Transformer. The work introduces RICO-PW annotations and evaluates metadata-free screen understanding on relation, clickability, retrieval, and app-classification tasks.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, data-generation\nPublished: 2021-05-25\nAuthors: Jingwen Fu, Xiaoyi Zhang, Yuwang Wang, Wenjun Zeng, Sam Yang, Grayson Hilliard\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2105.11941v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2105-11941/"}
{"id": "arxiv:2105.13231", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0caf694223bc2569cdd01662cadf54cbc96cad370866a231c15437c5d5d0ba", "submitted_at": "2021-05-27T15:20:14Z", "updated_at": "2021-05-27T15:20:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2105.13231v1"}, "authors": ["Daniel Toyama", "Philippe Hamel", "Anita Gergely", "Gheorghe Comanici", "Amelia Glaese", "Zafarali Ahmed", "Tyler Jackson", "Shibl Mourad", "Doina Precup"], "canonical_source_url": "https://arxiv.org/abs/2105.13231", "citation_url": "https://awesomecua.github.io/resources/arxiv-2105-13231/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0caf694223bc2569cdd01662cadf54cbc96cad370866a231c15437c5d5d0ba", "submitted_at": "2021-05-27T15:20:14Z", "updated_at": "2021-05-27T15:20:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2105.13231v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2105.13231"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2105.13231"}], "name": "AndroidEnv: A Reinforcement Learning Platform for Android", "publication": null, "published_at": "2021-05-27", "record_sha256": "84e6c9f39712bac72c3146752340d4446bef22d62394fe8cd831e3cce8d4ff7c", "relations": [], "resource_id": "arxiv:2105.13231", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2105.13231v1", "topics": ["reinforcement-learning", "sandbox"], "updated_at": "2021-05-27"}, "text": "AndroidEnv: A Reinforcement Learning Platform for Android\nResource type: paper\nReview status: curated\nScope: core\n\nPresents an Android environment for training and evaluating reinforcement-learning agents through touchscreen interaction.\nPlatforms: android\nTopics: reinforcement-learning, sandbox\nPublished: 2021-05-27\nAuthors: Daniel Toyama, Philippe Hamel, Anita Gergely, Gheorghe Comanici, Amelia Glaese, Zafarali Ahmed, Tyler Jackson, Shibl Mourad, Doina Precup\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2105.13231v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2105-13231/"}
{"id": "arxiv:2107.13731", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4f582015ccaf5bbe9be4b7ab925b2bbf6bd8451a4e63d89e1fe38363e26b31", "submitted_at": "2021-07-29T03:51:36Z", "updated_at": "2021-08-10T08:55:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2107.13731v2"}, "authors": ["Chongyang Bai", "Xiaoxue Zang", "Ying Xu", "Srinivas Sunkara", "Abhinav Rastogi", "Jindong Chen", "Blaise Aguera y Arcas"], "canonical_source_url": "https://arxiv.org/abs/2107.13731", "citation_url": "https://awesomecua.github.io/resources/arxiv-2107-13731/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4f582015ccaf5bbe9be4b7ab925b2bbf6bd8451a4e63d89e1fe38363e26b31", "submitted_at": "2021-07-29T03:51:36Z", "updated_at": "2021-08-10T08:55:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2107.13731v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2107.13731"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2107.13731"}], "name": "UIBert: Learning Generic Multimodal Representations for UI Understanding", "publication": null, "published_at": "2021-07-29", "record_sha256": "b60252cc522ba35bb0e4a392691a99a4b4289fb8c8cc279f52fe0b30a9966f8a", "relations": [], "resource_id": "arxiv:2107.13731", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2107.13731v2", "topics": ["screen-understanding"], "updated_at": "2021-08-10"}, "text": "UIBert: Learning Generic Multimodal Representations for UI Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nLearns shared image, text, and structural UI representations through self-supervised tasks based on relationships among interface components.\nPlatforms: unspecified\nTopics: screen-understanding\nPublished: 2021-07-29\nAuthors: Chongyang Bai, Xiaoxue Zang, Ying Xu, Srinivas Sunkara, Abhinav Rastogi, Jindong Chen, Blaise Aguera y Arcas\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2107.13731v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2107-13731/"}
{"id": "arxiv:2108.03353", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9cf8d8521e643e1a888e706dff4e5106f178061d9af4ed2c2ba5ae42f194785a", "submitted_at": "2021-08-07T03:01:23Z", "updated_at": "2021-08-07T03:01:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2108.03353v1"}, "authors": ["Bryan Wang", "Gang Li", "Xin Zhou", "Zhourong Chen", "Tovi Grossman", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2108.03353", "citation_url": "https://awesomecua.github.io/resources/arxiv-2108-03353/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9cf8d8521e643e1a888e706dff4e5106f178061d9af4ed2c2ba5ae42f194785a", "submitted_at": "2021-08-07T03:01:23Z", "updated_at": "2021-08-07T03:01:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2108.03353v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2108.03353"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2108.03353"}], "name": "Screen2Words: Automatic Mobile UI Summarization with Multimodal Learning", "publication": null, "published_at": "2021-08-07", "record_sha256": "0f47f3213629f5c152ee4ccbf4ce8ba9a1c5846476726b0d6be2e6fff7ee6f31", "relations": [], "resource_id": "arxiv:2108.03353", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2108.03353v1", "topics": ["screen-understanding"], "updated_at": "2021-08-07"}, "text": "Screen2Words: Automatic Mobile UI Summarization with Multimodal Learning\nResource type: paper\nReview status: curated\nScope: core\n\nStudies mobile-screen summarization using visual, textual, and structural UI information, with a dataset of human-written descriptions.\nPlatforms: unspecified\nTopics: screen-understanding\nPublished: 2021-08-07\nAuthors: Bryan Wang, Gang Li, Xin Zhou, Zhourong Chen, Tovi Grossman, Yang Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2108.03353v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2108-03353/"}
{"id": "arxiv:2112.05692", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e88e6e2659df232174d9fc71a2fac61a021941f096355f8ea3cd4c618c223d3", "submitted_at": "2021-12-10T17:37:26Z", "updated_at": "2021-12-10T17:37:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.05692v1"}, "authors": ["Yang Li", "Gang Li", "Xin Zhou", "Mostafa Dehghani", "Alexey Gritsenko"], "canonical_source_url": "https://arxiv.org/abs/2112.05692", "citation_url": "https://awesomecua.github.io/resources/arxiv-2112-05692/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e88e6e2659df232174d9fc71a2fac61a021941f096355f8ea3cd4c618c223d3", "submitted_at": "2021-12-10T17:37:26Z", "updated_at": "2021-12-10T17:37:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2112.05692v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2112.05692"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2112.05692"}], "name": "VUT: Versatile UI Transformer for Multi-Modal Multi-Task User Interface Modeling", "publication": null, "published_at": "2021-12-10", "record_sha256": "aec1a1fe833e87904337319ee6960e72e93bc8d249807fd85a2a7bcfaa21926e", "relations": [], "resource_id": "arxiv:2112.05692", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2112.05692v1", "topics": ["screen-understanding", "grounding", "efficiency"], "updated_at": "2021-12-10"}, "text": "VUT: Versatile UI Transformer for Multi-Modal Multi-Task User Interface Modeling\nResource type: paper\nReview status: curated\nScope: core\n\nThe Versatile UI Transformer jointly models interface screenshots, structure, and language for five UI tasks. Its encoder supports element detection when structural metadata is absent, while an autoregressive component handles UI question answering and command grounding.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, efficiency\nPublished: 2021-12-10\nAuthors: Yang Li, Gang Li, Xin Zhou, Mostafa Dehghani, Alexey Gritsenko\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2112.05692v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2112-05692/"}
{"id": "arxiv:2201.04100", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d971ee06cafd46fe27a01d40bc16e3604e1d153c38ee92be970328d9935b0708", "submitted_at": "2022-01-11T17:52:40Z", "updated_at": "2022-01-13T17:53:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2201.04100v2"}, "authors": ["Gang Li", "Gilles Baechler", "Manuel Tragut", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2201.04100", "citation_url": "https://awesomecua.github.io/resources/arxiv-2201-04100/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d971ee06cafd46fe27a01d40bc16e3604e1d153c38ee92be970328d9935b0708", "submitted_at": "2022-01-11T17:52:40Z", "updated_at": "2022-01-13T17:53:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2201.04100v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:26Z", "role": "primary", "sha256": "3e89566cdb4521a8b3b0d99c514c3401645eda5b5c0de9720eb726b1ffe0a11f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1511.03745%2C1602.02261%2C1809.04860%2C1811.06521%2C1811.07871%2C1812.03426%2C1901.11528%2C1909.00031%2C1912.03098%2C2003.00431%2C2004.14973%2C2006.16241%2C2007.04428%2C2010.09670%2C2010.12083%2C2012.12350%2C2104.08244%2C2105.11941%2C2106.09422%2C2106.13500%2C2107.05911%2C2108.03298%2C2108.05308%2C2110.10661%2C2201.04100%2C2201.11676%2C2203.10897%2C2204.05239%2C2205.10816%2C2206.08853%2C2206.12839%2C2206.15407%2C2208.06032%2C2208.08080%2C2209.14981%2C2210.01379%2C2210.03735%2C2210.08073%2C2210.10275%2C2210.11435%2C2211.07912%2C2211.09527%2C2212.13326%2C2301.10677%2C2301.13779%2C2302.12173%2C2303.00001%2C2303.11366%2C2303.12027%2C2304.03442&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "30f2d08b6bf51489cf780365753d9e9e945156951014ae76501b5796eaacf83c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2201.04100v2"}], "identifiers": {"arxiv": "2201.04100"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2201.04100"}], "name": "Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale", "publication": null, "published_at": "2022-01-11", "record_sha256": "839ee39b2e596374a6d0ffe292fc5bbca63e9eaed26f793cd15609aa18d1862b", "relations": [], "resource_id": "arxiv:2201.04100", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2201.04100v2", "topics": ["screen-understanding", "data-generation"], "updated_at": "2022-01-13"}, "text": "Learning to Denoise Raw Mobile UI Layouts for Improving Datasets at Scale\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CLAY, a pipeline that repairs noisy mobile UI layouts by removing nodes without valid visual counterparts and assigning semantic element types, supported by human-annotated layouts for evaluation.\nPlatforms: unspecified\nTopics: screen-understanding, data-generation\nPublished: 2022-01-11\nAuthors: Gang Li, Gilles Baechler, Manuel Tragut, Yang Li\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2201.04100v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2201-04100/"}
{"id": "arxiv:2202.02312", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1687369e31b0ad3632381d4a46b1d555c2e2ed1d6504e9c1b63006dd29f3445e", "submitted_at": "2022-02-04T18:51:50Z", "updated_at": "2022-08-15T00:24:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2202.02312v3"}, "authors": ["Andrea Burns", "Deniz Arsan", "Sanjna Agrawal", "Ranjitha Kumar", "Kate Saenko", "Bryan A. Plummer"], "canonical_source_url": "https://arxiv.org/abs/2202.02312", "citation_url": "https://awesomecua.github.io/resources/arxiv-2202-02312/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1687369e31b0ad3632381d4a46b1d555c2e2ed1d6504e9c1b63006dd29f3445e", "submitted_at": "2022-02-04T18:51:50Z", "updated_at": "2022-08-15T00:24:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2202.02312v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2202.02312"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2202.02312"}], "name": "A Dataset for Interactive Vision-Language Navigation with Unknown Command Feasibility", "publication": null, "published_at": "2022-02-04", "record_sha256": "1d5739dd8be5acb7d574b3fe692a651d12ce2eed971c41c1473e835e7eb3698e", "relations": [], "resource_id": "arxiv:2202.02312", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2202.02312v3", "topics": ["mobile-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2022-08-15"}, "text": "A Dataset for Interactive Vision-Language Navigation with Unknown Command Feasibility\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MoTIF mobile-app interactions with feasible and infeasible commands, feasibility explanations, and follow-up questions for ambiguous requests.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop, evaluation\nPublished: 2022-02-04\nAuthors: Andrea Burns, Deniz Arsan, Sanjna Agrawal, Ranjitha Kumar, Kate Saenko, Bryan A. Plummer\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2202.02312v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2202-02312/"}
{"id": "arxiv:2202.08137", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94129847ac85f4a8aeab74907cc68a31c435b078ab950695d9328b6c3896b46d", "submitted_at": "2022-02-16T15:23:46Z", "updated_at": "2022-11-11T13:45:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2202.08137v2"}, "authors": ["Peter C Humphreys", "David Raposo", "Toby Pohlen", "Gregory Thornton", "Rachita Chhaparia", "Alistair Muldal", "Josh Abramson", "Petko Georgiev", "Alex Goldin", "Adam Santoro", "Timothy Lillicrap"], "canonical_source_url": "https://arxiv.org/abs/2202.08137", "citation_url": "https://awesomecua.github.io/resources/arxiv-2202-08137/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94129847ac85f4a8aeab74907cc68a31c435b078ab950695d9328b6c3896b46d", "submitted_at": "2022-02-16T15:23:46Z", "updated_at": "2022-11-11T13:45:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2202.08137v2"}, "publication": {"authors": ["Peter Humphreys", "David Raposo", "Toby Pohlen", "Gregory Thornton", "Rachita Chhaparia", "Alistair Muldal", "Josh Abramson", "Petko Georgiev", "Adam Santoro", "Timothy Lillicrap"], "date_basis": "Explicit publisher metadata: 2022/06/28; precision=day.", "date_parts": [2022, 6, 28], "doi": null, "event_dates": "17--23 Jul 2022", "evidence_urls": ["https://proceedings.mlr.press/v162/humphreys22a.html", "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal ICML PDF lists 10 authors beginning Peter Humphreys and uses Toby Pohlen. The landing lists Peter C Humphreys and Tobias Pohlen. The retained arXiv snapshot lists 11 authors, including Alex Goldin. These are preserved as distinct version-specific author lists; the publication field uses only the formal PDF list. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher issue date is 2022-06-28; the July 17–23 conference range is an event date, not a substitute publication day. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9466-9482", "pdf_url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "A Data-Driven Approach for Learning to Control Computers", "type": "conference-paper", "url": "https://proceedings.mlr.press/v162/humphreys22a.html", "venue": "Proceedings of the 39th International Conference on Machine Learning", "volume": "162"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "b0ac6ec4bf3aad3773c53a0fffe8688a6287f25f28faf3743603ec2c1ae08a8b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v162/humphreys22a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "ba2a9a0a5e46aa5f34cfc720ac08049253917407e2fbd4a8cdbd2c8ca3ec8455", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2202.08137"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2202.08137"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v162/humphreys22a.html"}, {"rel": "publication_pdf", "url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"}], "name": "A data-driven approach for learning to control computers", "publication": {"authors": ["Peter Humphreys", "David Raposo", "Toby Pohlen", "Gregory Thornton", "Rachita Chhaparia", "Alistair Muldal", "Josh Abramson", "Petko Georgiev", "Adam Santoro", "Timothy Lillicrap"], "date_basis": "Explicit publisher metadata: 2022/06/28; precision=day.", "date_parts": [2022, 6, 28], "doi": null, "event_dates": "17--23 Jul 2022", "evidence_urls": ["https://proceedings.mlr.press/v162/humphreys22a.html", "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal ICML PDF lists 10 authors beginning Peter Humphreys and uses Toby Pohlen. The landing lists Peter C Humphreys and Tobias Pohlen. The retained arXiv snapshot lists 11 authors, including Alex Goldin. These are preserved as distinct version-specific author lists; the publication field uses only the formal PDF list. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher issue date is 2022-06-28; the July 17–23 conference range is an event date, not a substitute publication day. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9466-9482", "pdf_url": "https://proceedings.mlr.press/v162/humphreys22a/humphreys22a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "A Data-Driven Approach for Learning to Control Computers", "type": "conference-paper", "url": "https://proceedings.mlr.press/v162/humphreys22a.html", "venue": "Proceedings of the 39th International Conference on Machine Learning", "volume": "162"}, "published_at": "2022-02-16", "record_sha256": "aaf3cdd8e0b4413bbb67edbed75d7c92086e8c12a469e344acb010c7fa3da1fb", "relations": [], "resource_id": "arxiv:2202.08137", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2202.08137v2", "topics": ["reinforcement-learning", "automation"], "updated_at": "2022-11-11"}, "text": "A data-driven approach for learning to control computers\nResource type: paper\nReview status: curated\nScope: core\n\nCombines reinforcement learning with behavioral priors from human interactions to learn computer control through keyboard and mouse actions.\nPlatforms: web\nTopics: reinforcement-learning, automation\nPublished: 2022-02-16\nAuthors: Peter C Humphreys, David Raposo, Toby Pohlen, Gregory Thornton, Rachita Chhaparia, Alistair Muldal, Josh Abramson, Petko Georgiev, Alex Goldin, Adam Santoro, Timothy Lillicrap\nRecorded arXiv version: v2\n\nPublication edition: A Data-Driven Approach for Learning to Control Computers\nPublication authors: Peter Humphreys, David Raposo, Toby Pohlen, Gregory Thornton, Rachita Chhaparia, Alistair Muldal, Josh Abramson, Petko Georgiev, Adam Santoro, Timothy Lillicrap\nPublication edition date (recorded precision): 2022-06-28\nPublication venue: Proceedings of the 39th International Conference on Machine Learning\nPublication publisher: PMLR\nPublication volume: 162\nPublication pages: 9466-9482\nPublication date basis: Explicit publisher metadata: 2022/06/28; precision=day.\nPublication event dates: 17--23 Jul 2022\nPublication notes: The formal ICML PDF lists 10 authors beginning Peter Humphreys and uses Toby Pohlen. The landing lists Peter C Humphreys and Tobias Pohlen. The retained arXiv snapshot lists 11 authors, including Alex Goldin. These are preserved as distinct version-specific author lists; the publication field uses only the formal PDF list. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher issue date is 2022-06-28; the July 17–23 conference range is an event date, not a substitute publication day. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://proceedings.mlr.press/v162/humphreys22a.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2202.08137v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2202-08137/"}
{"id": "arxiv:2205.11029", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbe2a4ece7ef72dfa008740991a5c46d9ab4b9de73f3de02f5dabcee2a5c1aa5", "submitted_at": "2022-05-23T04:05:37Z", "updated_at": "2022-11-24T06:34:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2205.11029v2"}, "authors": ["Liangtai Sun", "Xingyu Chen", "Lu Chen", "Tianle Dai", "Zichen Zhu", "Kai Yu"], "canonical_source_url": "https://arxiv.org/abs/2205.11029", "citation_url": "https://awesomecua.github.io/resources/arxiv-2205-11029/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbe2a4ece7ef72dfa008740991a5c46d9ab4b9de73f3de02f5dabcee2a5c1aa5", "submitted_at": "2022-05-23T04:05:37Z", "updated_at": "2022-11-24T06:34:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2205.11029v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2205.11029"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2205.11029"}], "name": "META-GUI: Towards Multi-modal Conversational Agents on Mobile GUI", "publication": null, "published_at": "2022-05-23", "record_sha256": "07826d35bb0b927b20783543a9340bea78e33b5f0c4df81d258f64b5e621da7a", "relations": [], "resource_id": "arxiv:2205.11029", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2205.11029v2", "topics": ["mobile-navigation", "human-in-the-loop"], "updated_at": "2022-11-24"}, "text": "META-GUI: Towards Multi-modal Conversational Agents on Mobile GUI\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-based task-oriented dialogue and a dataset pairing mobile-interface operations with conversational responses.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop\nPublished: 2022-05-23\nAuthors: Liangtai Sun, Xingyu Chen, Lu Chen, Tianle Dai, Zichen Zhu, Kai Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2205.11029v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2205-11029/"}
{"id": "arxiv:2207.01206", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6de4ab2a89875f1ccd99bbaa075c234d82979e845f27fe9530bdad0539c49570", "submitted_at": "2022-07-04T05:30:22Z", "updated_at": "2023-02-08T01:39:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2207.01206v4"}, "authors": ["Shunyu Yao", "Howard Chen", "John Yang", "Karthik Narasimhan"], "canonical_source_url": "https://arxiv.org/abs/2207.01206", "citation_url": "https://awesomecua.github.io/resources/arxiv-2207-01206/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6de4ab2a89875f1ccd99bbaa075c234d82979e845f27fe9530bdad0539c49570", "submitted_at": "2022-07-04T05:30:22Z", "updated_at": "2023-02-08T01:39:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2207.01206v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2207.01206"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2207.01206"}], "name": "WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents", "publication": null, "published_at": "2022-07-04", "record_sha256": "45fd768c98211666474f1a363c4f821ce818507243f0e092a121d830c88b57fc", "relations": [], "resource_id": "arxiv:2207.01206", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2207.01206v4", "topics": ["evaluation", "web-navigation"], "updated_at": "2023-02-08"}, "text": "WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProvides a simulated shopping environment where language agents search, configure, and select products from natural-language requirements.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2022-07-04\nAuthors: Shunyu Yao, Howard Chen, John Yang, Karthik Narasimhan\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2207.01206v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2207-01206/"}
{"id": "arxiv:2209.08199", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37f7ba5d5b9bbc96fd3a723e49190f296e08d2874dcd545272d3fcef3cc33d24", "submitted_at": "2022-09-16T23:49:00Z", "updated_at": "2025-02-09T21:09:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.08199v4"}, "authors": ["Yu-Chung Hsiao", "Fedir Zubach", "Gilles Baechler", "Srinivas Sunkara", "Victor Carbune", "Jason Lin", "Maria Wang", "Yun Zhu", "Jindong Chen"], "canonical_source_url": "https://arxiv.org/abs/2209.08199", "citation_url": "https://awesomecua.github.io/resources/arxiv-2209-08199/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37f7ba5d5b9bbc96fd3a723e49190f296e08d2874dcd545272d3fcef3cc33d24", "submitted_at": "2022-09-16T23:49:00Z", "updated_at": "2025-02-09T21:09:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.08199v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2209.08199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.08199"}], "name": "ScreenQA: Large-Scale Question-Answer Pairs over Mobile App Screenshots", "publication": null, "published_at": "2022-09-16", "record_sha256": "f47b4a25e626240951e2f49a2c9a6a362ebf5aa7f3cd453c47eb87720f247011", "relations": [], "resource_id": "arxiv:2209.08199", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2209.08199v4", "topics": ["screen-understanding", "evaluation"], "updated_at": "2025-02-09"}, "text": "ScreenQA: Large-Scale Question-Answer Pairs over Mobile App Screenshots\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces screenshot question answering over mobile interfaces, linking answers to screen content and bounding boxes.\nPlatforms: unspecified\nTopics: screen-understanding, evaluation\nPublished: 2022-09-16\nAuthors: Yu-Chung Hsiao, Fedir Zubach, Gilles Baechler, Srinivas Sunkara, Victor Carbune, Jason Lin, Maria Wang, Yun Zhu, Jindong Chen\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2209.08199v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2209-08199/"}
{"id": "arxiv:2209.08655", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65813c1784ba00e354aff4e38796b1bdc8e9b691cba9dd56b8232b5d86555d1b", "submitted_at": "2022-09-18T20:58:39Z", "updated_at": "2023-02-17T21:24:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.08655v2"}, "authors": ["Bryan Wang", "Gang Li", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2209.08655", "citation_url": "https://awesomecua.github.io/resources/arxiv-2209-08655/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65813c1784ba00e354aff4e38796b1bdc8e9b691cba9dd56b8232b5d86555d1b", "submitted_at": "2022-09-18T20:58:39Z", "updated_at": "2023-02-17T21:24:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2209.08655v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2209.08655"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.08655"}], "name": "Enabling Conversational Interaction with Mobile UI using Large Language Models", "publication": null, "published_at": "2022-09-18", "record_sha256": "7ba80b557bd6a41b5b16f32a895d322f2759cd51d63da874ac8ba093384c1e31", "relations": [], "resource_id": "arxiv:2209.08655", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2209.08655v2", "topics": ["mobile-navigation", "screen-understanding", "training-free"], "updated_at": "2023-02-17"}, "text": "Enabling Conversational Interaction with Mobile UI using Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nStudies few-shot prompting of a single language model for four mobile-UI interaction tasks, aiming to support conversational interface use without a separate trained model for each task.\nPlatforms: unspecified\nTopics: mobile-navigation, screen-understanding, training-free\nPublished: 2022-09-18\nAuthors: Bryan Wang, Gang Li, Yang Li\nRecorded arXiv version: v2\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2209.08655v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2209-08655/"}
{"id": "arxiv:2209.14927", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eaa351e34fa6bdcbe26aa6714eacda1f667411090abfdd5575735605dc89e864", "submitted_at": "2022-09-29T16:45:43Z", "updated_at": "2023-02-24T01:41:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.14927v4"}, "authors": ["Gang Li", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2209.14927", "citation_url": "https://awesomecua.github.io/resources/arxiv-2209-14927/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eaa351e34fa6bdcbe26aa6714eacda1f667411090abfdd5575735605dc89e864", "submitted_at": "2022-09-29T16:45:43Z", "updated_at": "2023-02-24T01:41:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2209.14927v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2209.14927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.14927"}], "name": "Spotlight: Mobile UI Understanding using Vision-Language Models with a Focus", "publication": null, "published_at": "2022-09-29", "record_sha256": "2b8b2d9a7d7d219e8cb0f6949ab0bda713c5e7cc1eea8a7c7c5869bfb4e40a54", "relations": [], "resource_id": "arxiv:2209.14927", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2209.14927v4", "topics": ["screen-understanding", "grounding"], "updated_at": "2023-02-24"}, "text": "Spotlight: Mobile UI Understanding using Vision-Language Models with a Focus\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Spotlight, a vision-language model conditioned on a UI screenshot and a selected region of interest. It evaluates screenshot-based mobile UI understanding and explores multitask learning and few-shot prompting.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\nPublished: 2022-09-29\nAuthors: Gang Li, Yang Li\nRecorded arXiv version: v4\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2209.14927v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2209-14927/"}
{"id": "arxiv:2209.15099", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e57e7dd30e6f3ae8b13cbd4a190b39339f2bbf7a551a6a5df5190dc62b5434", "submitted_at": "2022-09-29T21:08:18Z", "updated_at": "2022-09-29T21:08:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.15099v1"}, "authors": ["Tao Li", "Gang Li", "Jingjie Zheng", "Purple Wang", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2209.15099", "citation_url": "https://awesomecua.github.io/resources/arxiv-2209-15099/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e57e7dd30e6f3ae8b13cbd4a190b39339f2bbf7a551a6a5df5190dc62b5434", "submitted_at": "2022-09-29T21:08:18Z", "updated_at": "2022-09-29T21:08:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2209.15099v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2209.15099"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2209.15099"}], "name": "MUG: Interactive Multimodal Grounding on User Interfaces", "publication": null, "published_at": "2022-09-29", "record_sha256": "b12fb7d77ed5ab0f08bba1561fdb8509aae61272b47f3f292dc295c187d80e7e", "relations": [], "resource_id": "arxiv:2209.15099", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2209.15099v1", "topics": ["grounding", "human-in-the-loop"], "updated_at": "2022-09-29"}, "text": "MUG: Interactive Multimodal Grounding on User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nStudies iterative UI grounding in which users refine or correct an agent's interpretation over multiple conversational turns.\nPlatforms: unspecified\nTopics: grounding, human-in-the-loop\nPublished: 2022-09-29\nAuthors: Tao Li, Gang Li, Jingjie Zheng, Purple Wang, Yang Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2209.15099v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2209-15099/"}
{"id": "arxiv:2210.02663", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24d037f1abeb7708852ac5517ce02cc46622e359d9feab407f188c0657cc72f4", "submitted_at": "2022-10-06T03:48:54Z", "updated_at": "2022-10-06T03:48:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.02663v1"}, "authors": ["Srinivas Sunkara", "Maria Wang", "Lijuan Liu", "Gilles Baechler", "Yu-Chung Hsiao", "Jindong", "Chen", "Abhanshu Sharma", "James Stout"], "canonical_source_url": "https://arxiv.org/abs/2210.02663", "citation_url": "https://awesomecua.github.io/resources/arxiv-2210-02663/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "24d037f1abeb7708852ac5517ce02cc46622e359d9feab407f188c0657cc72f4", "submitted_at": "2022-10-06T03:48:54Z", "updated_at": "2022-10-06T03:48:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2210.02663v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2210.02663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.02663"}], "name": "Towards Better Semantic Understanding of Mobile Interfaces", "publication": null, "published_at": "2022-10-06", "record_sha256": "05e2b5b657737bff6cacd46e080476a74bdf59d75d24a7ed5ec1ba3ad0a0d1a4", "relations": [], "resource_id": "arxiv:2210.02663", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2210.02663v1", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2022-10-06"}, "text": "Towards Better Semantic Understanding of Mobile Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nAugments RICO screens and view hierarchies with human labels for icon shape and meaning and for links between UI elements and text labels. It compares image-only and multimodal models on semantic understanding of previously unseen apps.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, data-generation\nPublished: 2022-10-06\nAuthors: Srinivas Sunkara, Maria Wang, Lijuan Liu, Gilles Baechler, Yu-Chung Hsiao, Jindong, Chen, Abhanshu Sharma, James Stout\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2210.02663v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2210-02663/"}
{"id": "arxiv:2210.03347", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ebb243c4690aba5f1a92b0d1e050399560998a7f304ea483dae2f451cf51ec0", "submitted_at": "2022-10-07T06:42:06Z", "updated_at": "2023-06-15T21:34:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03347v2"}, "authors": ["Kenton Lee", "Mandar Joshi", "Iulia Turc", "Hexiang Hu", "Fangyu Liu", "Julian Eisenschlos", "Urvashi Khandelwal", "Peter Shaw", "Ming-Wei Chang", "Kristina Toutanova"], "canonical_source_url": "https://arxiv.org/abs/2210.03347", "citation_url": "https://awesomecua.github.io/resources/arxiv-2210-03347/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ebb243c4690aba5f1a92b0d1e050399560998a7f304ea483dae2f451cf51ec0", "submitted_at": "2022-10-07T06:42:06Z", "updated_at": "2023-06-15T21:34:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03347v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2210.03347"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.03347"}], "name": "Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding", "publication": null, "published_at": "2022-10-07", "record_sha256": "d4a1298eecb532aad3063699eaaefb7c218193ec4f3e3e6b745038c4d20b824a", "relations": [], "resource_id": "arxiv:2210.03347", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2210.03347v2", "topics": ["screen-understanding", "data-generation"], "updated_at": "2023-06-15"}, "text": "Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding\nResource type: paper\nReview status: curated\nScope: adjacent\n\nPix2Struct pretrains an image-to-text model by reconstructing simplified HTML from masked webpage screenshots. Its variable-resolution visual representation is evaluated across several domains, including user-interface understanding.\nPlatforms: web\nTopics: screen-understanding, data-generation\nPublished: 2022-10-07\nAuthors: Kenton Lee, Mandar Joshi, Iulia Turc, Hexiang Hu, Fangyu Liu, Julian Eisenschlos, Urvashi Khandelwal, Peter Shaw, Ming-Wei Chang, Kristina Toutanova\nRecorded arXiv version: v2\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced. Adjacent visual-language pretraining: webpage screenshot parsing and explicit UI evaluation support interface understanding; the paper does not establish an autonomous GUI policy.\n\nPrimary source: https://arxiv.org/abs/2210.03347v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2210-03347/"}
{"id": "arxiv:2210.03945", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b9089164b5c784735d8915596fe902840e01c8215f507caed0e10d4361d13ec", "submitted_at": "2022-10-08T07:27:17Z", "updated_at": "2023-05-19T17:42:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03945v2"}, "authors": ["Izzeddin Gur", "Ofir Nachum", "Yingjie Miao", "Mustafa Safdari", "Austin Huang", "Aakanksha Chowdhery", "Sharan Narang", "Noah Fiedel", "Aleksandra Faust"], "canonical_source_url": "https://arxiv.org/abs/2210.03945", "citation_url": "https://awesomecua.github.io/resources/arxiv-2210-03945/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b9089164b5c784735d8915596fe902840e01c8215f507caed0e10d4361d13ec", "submitted_at": "2022-10-08T07:27:17Z", "updated_at": "2023-05-19T17:42:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2210.03945v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2210.03945"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2210.03945"}], "name": "Understanding HTML with Large Language Models", "publication": null, "published_at": "2022-10-08", "record_sha256": "eb1f2187eb84c6c9f0bb374ff97bbf3638d4e36479456f9a4214f7ea6bfd2140", "relations": [], "resource_id": "arxiv:2210.03945", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2210.03945v2", "topics": ["screen-understanding", "web-navigation", "data-generation"], "updated_at": "2023-05-19"}, "text": "Understanding HTML with Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nStudies language models for HTML element classification, page description and autonomous web navigation. Fine-tuned models are evaluated on MiniWoB, and a distilled, automatically labeled CommonCrawl HTML corpus supports the training approach.\nPlatforms: web\nTopics: screen-understanding, web-navigation, data-generation\nPublished: 2022-10-08\nAuthors: Izzeddin Gur, Ofir Nachum, Yingjie Miao, Mustafa Safdari, Austin Huang, Aakanksha Chowdhery, Sharan Narang, Noah Fiedel, Aleksandra Faust\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2210.03945v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2210-03945/"}
{"id": "arxiv:2211.07615", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52319819402cec76a69c0bb513f3a7000053208ddd9d2a8c675e055422402fbe", "submitted_at": "2022-11-14T18:36:19Z", "updated_at": "2023-05-23T16:08:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2211.07615v2"}, "authors": ["Sagar Gubbi Venkatesh", "Partha Talukdar", "Srini Narayanan"], "canonical_source_url": "https://arxiv.org/abs/2211.07615", "citation_url": "https://awesomecua.github.io/resources/arxiv-2211-07615/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52319819402cec76a69c0bb513f3a7000053208ddd9d2a8c675e055422402fbe", "submitted_at": "2022-11-14T18:36:19Z", "updated_at": "2023-05-23T16:08:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2211.07615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2211.07615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2211.07615"}], "name": "UGIF: UI Grounded Instruction Following", "publication": null, "published_at": "2022-11-14", "record_sha256": "85ce1769f67364b4b6ebc23661ea8d62c02f29b83be3c3eb8e4ffc74d21ec926", "relations": [], "resource_id": "arxiv:2211.07615", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2211.07615v2", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-05-23"}, "text": "UGIF: UI Grounded Instruction Following\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces multilingual smartphone instruction-following data and a pipeline that retrieves help instructions and maps them to UI actions.\nPlatforms: unspecified\nTopics: mobile-navigation, grounding\nPublished: 2022-11-14\nAuthors: Sagar Gubbi Venkatesh, Partha Talukdar, Srini Narayanan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2211.07615v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2211-07615/"}
{"id": "arxiv:2301.10165", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb647a96c7f8263786e85a5061c8c5284cb92ed0b2ebc4c405e1243b2ca3fa2b", "submitted_at": "2023-01-23T09:05:49Z", "updated_at": "2023-01-23T09:05:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.10165v1"}, "authors": ["Pratyay Banerjee", "Shweti Mahajan", "Kushal Arora", "Chitta Baral", "Oriana Riva"], "canonical_source_url": "https://arxiv.org/abs/2301.10165", "citation_url": "https://awesomecua.github.io/resources/arxiv-2301-10165/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb647a96c7f8263786e85a5061c8c5284cb92ed0b2ebc4c405e1243b2ca3fa2b", "submitted_at": "2023-01-23T09:05:49Z", "updated_at": "2023-01-23T09:05:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.10165v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2301.10165"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2301.10165"}], "name": "Lexi: Self-Supervised Learning of the UI Language", "publication": null, "published_at": "2023-01-23", "record_sha256": "be2df608a4d41ee8546b138cf88aad7ac048e7df27fb51ad3e7bc582bdf4a3d2", "relations": [], "resource_id": "arxiv:2301.10165", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2301.10165v1", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2023-01-23"}, "text": "Lexi: Self-Supervised Learning of the UI Language\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Lexi and UICaption to learn UI image-text representations from descriptions of interface functionality. Evaluation covers UI action entailment, instruction-based retrieval, referring-expression grounding, and UI entity recognition.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, data-generation\nPublished: 2023-01-23\nAuthors: Pratyay Banerjee, Shweti Mahajan, Kushal Arora, Chitta Baral, Oriana Riva\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2301.10165v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2301-10165/"}
{"id": "arxiv:2301.13280", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0cc3b91aa85bc8c7d88a878c7b24926943eea0bcc92e25f517915242f2a82a39", "submitted_at": "2023-01-30T20:47:12Z", "updated_at": "2023-01-30T20:47:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.13280v1"}, "authors": ["Jason Wu", "Siyan Wang", "Siman Shen", "Yi-Hao Peng", "Jeffrey Nichols", "Jeffrey P. Bigham"], "canonical_source_url": "https://arxiv.org/abs/2301.13280", "citation_url": "https://awesomecua.github.io/resources/arxiv-2301-13280/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0cc3b91aa85bc8c7d88a878c7b24926943eea0bcc92e25f517915242f2a82a39", "submitted_at": "2023-01-30T20:47:12Z", "updated_at": "2023-01-30T20:47:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2301.13280v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2301.13280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2301.13280"}], "name": "WebUI: A Dataset for Enhancing Visual UI Understanding with Web Semantics", "publication": null, "published_at": "2023-01-30", "record_sha256": "e2c9a1dee02509ea153bd92f6c39220cba7564dd3ae9b4f705469cfff4955722", "relations": [], "resource_id": "arxiv:2301.13280", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2301.13280v1", "topics": ["screen-understanding", "data-generation"], "updated_at": "2023-01-30"}, "text": "WebUI: A Dataset for Enhancing Visual UI Understanding with Web Semantics\nResource type: paper\nReview status: curated\nScope: core\n\nCollects rendered webpages and structural metadata for visual UI modeling, studying how web semantics transfer to interface-understanding tasks.\nPlatforms: web\nTopics: screen-understanding, data-generation\nPublished: 2023-01-30\nAuthors: Jason Wu, Siyan Wang, Siman Shen, Yi-Hao Peng, Jeffrey Nichols, Jeffrey P. Bigham\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2301.13280v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2301-13280/"}
{"id": "arxiv:2303.17491", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f36bd6088175b337b750d91d0b7c04d2c156742e5c4090a74f00a482157e386", "submitted_at": "2023-03-30T16:01:52Z", "updated_at": "2023-11-16T20:15:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2303.17491v3"}, "authors": ["Geunwoo Kim", "Pierre Baldi", "Stephen McAleer"], "canonical_source_url": "https://arxiv.org/abs/2303.17491", "citation_url": "https://awesomecua.github.io/resources/arxiv-2303-17491/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f36bd6088175b337b750d91d0b7c04d2c156742e5c4090a74f00a482157e386", "submitted_at": "2023-03-30T16:01:52Z", "updated_at": "2023-11-16T20:15:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2303.17491v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2303.17491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2303.17491"}], "name": "Language Models can Solve Computer Tasks", "publication": null, "published_at": "2023-03-30", "record_sha256": "45ceb43be3a1e44f022c5e1ac2700ddfdf4b9f80f0bcff4c9e966bb439ef511f", "relations": [], "resource_id": "arxiv:2303.17491", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2303.17491v3", "topics": ["web-navigation", "self-reflection", "training-free"], "updated_at": "2023-11-16"}, "text": "Language Models can Solve Computer Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces recursive criticism and improvement prompting for language-model agents performing computer tasks, evaluated on MiniWoB++ with a small number of demonstrations per task.\nPlatforms: web\nTopics: web-navigation, self-reflection, training-free\nPublished: 2023-03-30\nAuthors: Geunwoo Kim, Pierre Baldi, Stephen McAleer\nRecorded arXiv version: v3\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2303.17491v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2303-17491/"}
{"id": "arxiv:2304.07061", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "420c9005fd0c687c27c758da70ef05a60b3b2d9b9efbf91a5d7661bf893a8b09", "submitted_at": "2023-04-14T11:31:56Z", "updated_at": "2024-01-07T08:23:47Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2304.07061v5"}, "authors": ["Hao Wen", "Hongming Wang", "Jiaxuan Liu", "Yuanchun Li"], "canonical_source_url": "https://arxiv.org/abs/2304.07061", "citation_url": "https://awesomecua.github.io/resources/arxiv-2304-07061/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "420c9005fd0c687c27c758da70ef05a60b3b2d9b9efbf91a5d7661bf893a8b09", "submitted_at": "2023-04-14T11:31:56Z", "updated_at": "2024-01-07T08:23:47Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2304.07061v5"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2304.07061"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2304.07061"}], "name": "DroidBot-GPT: GPT-powered UI Automation for Android", "publication": null, "published_at": "2023-04-14", "record_sha256": "591c8af5acde1c6424745ac72abd37188851c5446a9db68e0d4932e2cce460fe", "relations": [], "resource_id": "arxiv:2304.07061", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2304.07061v5", "topics": ["mobile-navigation", "automation"], "updated_at": "2024-01-07"}, "text": "DroidBot-GPT: GPT-powered UI Automation for Android\nResource type: paper\nReview status: curated\nScope: core\n\nTransforms Android UI states and available actions into language prompts for an LLM to choose and execute task steps.\nPlatforms: android\nTopics: mobile-navigation, automation\nPublished: 2023-04-14\nAuthors: Hao Wen, Hongming Wang, Jiaxuan Liu, Yuanchun Li\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2304.07061v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2304-07061/"}
{"id": "arxiv:2305.08144", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e454f547bdf8e01f47eb63c20f0cacfa8bd17a8c4ad37dfd3a1cceb732de178", "submitted_at": "2023-05-14T12:31:03Z", "updated_at": "2024-06-13T11:51:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.08144v4"}, "authors": ["Danyang Zhang", "Zhennan Shen", "Rui Xie", "Situo Zhang", "Tianbao Xie", "Zihan Zhao", "Siyuan Chen", "Lu Chen", "Hongshen Xu", "Ruisheng Cao", "Kai Yu"], "canonical_source_url": "https://arxiv.org/abs/2305.08144", "citation_url": "https://awesomecua.github.io/resources/arxiv-2305-08144/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e454f547bdf8e01f47eb63c20f0cacfa8bd17a8c4ad37dfd3a1cceb732de178", "submitted_at": "2023-05-14T12:31:03Z", "updated_at": "2024-06-13T11:51:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.08144v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2305.08144"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.08144"}], "name": "Mobile-Env: Building Qualified Evaluation Benchmarks for LLM-GUI Interaction", "publication": null, "published_at": "2023-05-14", "record_sha256": "0e45f0f2314fb83e8db17a8efb9690243bf57b06736f36d44c0d80754127bab1", "relations": [], "resource_id": "arxiv:2305.08144", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2305.08144v4", "topics": ["evaluation", "sandbox"], "updated_at": "2024-06-13"}, "text": "Mobile-Env: Building Qualified Evaluation Benchmarks for LLM-GUI Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nProvides an Android benchmark toolkit with controlled environments, intermediate feedback, and task sets for reproducible GUI-agent evaluation.\nPlatforms: android\nTopics: evaluation, sandbox\nPublished: 2023-05-14\nAuthors: Danyang Zhang, Zhennan Shen, Rui Xie, Situo Zhang, Tianbao Xie, Zihan Zhao, Siyuan Chen, Lu Chen, Hongshen Xu, Ruisheng Cao, Kai Yu\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2305.08144v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2305-08144/"}
{"id": "arxiv:2305.11854", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac0ea5a462890317ea29ade32c879d016181afbe5962e2d28cc175e251df41e6", "submitted_at": "2023-05-19T17:44:34Z", "updated_at": "2024-02-25T16:21:00Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.11854v4"}, "authors": ["Hiroki Furuta", "Kuang-Huei Lee", "Ofir Nachum", "Yutaka Matsuo", "Aleksandra Faust", "Shixiang Shane Gu", "Izzeddin Gur"], "canonical_source_url": "https://arxiv.org/abs/2305.11854", "citation_url": "https://awesomecua.github.io/resources/arxiv-2305-11854/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac0ea5a462890317ea29ade32c879d016181afbe5962e2d28cc175e251df41e6", "submitted_at": "2023-05-19T17:44:34Z", "updated_at": "2024-02-25T16:21:00Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2305.11854v4"}, "publication": {"authors": ["Hiroki Furuta", "Kuang-Huei Lee", "Ofir Nachum", "Yutaka Matsuo", "Aleksandra Faust", "Shixiang Shane Gu", "Izzeddin Gur"], "date_basis": "Explicit publisher metadata: 2024-05-31; precision=day.", "date_parts": [2024, 5, 31], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ICLR landing shortens Shixiang Shane Gu to Shixiang Gu. Publication authors use the full PDF title-page form. Volume 2024 and pages 29528–29558 are explicit publisher citation metadata, not values inferred from the event year or PDF page count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "29528-29558", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2024"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "6669190c7fc2a37f9cd68d0c49431b8f17b626c16bb3ef4093fc83a6ac62acd3", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "da26ea085252edfda4c084058aa81ea023f885ff97c67f2ffae0698bab6f568b", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2305.11854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.11854"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"}], "name": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models", "publication": {"authors": ["Hiroki Furuta", "Kuang-Huei Lee", "Ofir Nachum", "Yutaka Matsuo", "Aleksandra Faust", "Shixiang Shane Gu", "Izzeddin Gur"], "date_basis": "Explicit publisher metadata: 2024-05-31; precision=day.", "date_parts": [2024, 5, 31], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ICLR landing shortens Shixiang Shane Gu to Shixiang Gu. Publication authors use the full PDF title-page form. Volume 2024 and pages 29528–29558 are explicit publisher citation metadata, not values inferred from the event year or PDF page count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "29528-29558", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2024/file/7ef7d8359036afd8c2378d82c21058a4-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2024"}, "published_at": "2023-05-19", "record_sha256": "9a9fac749e919cd5de4744c1a6c0ee19f1c2716cbf4244aca93dacf8eeba8c31", "relations": [], "resource_id": "arxiv:2305.11854", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2305.11854v4", "topics": ["web-navigation", "imitation-learning", "action-prediction"], "updated_at": "2024-02-25"}, "text": "Multimodal Web Navigation with Instruction-Finetuned Foundation Models\nResource type: paper\nReview status: curated\nScope: core\n\nTrains WebGUM offline from demonstrations by jointly tuning an instruction-following language model and a vision encoder over webpage screenshots and HTML. The agent predicts browser actions and is evaluated on MiniWoB, WebShop, and Mind2Web.\nPlatforms: web\nTopics: web-navigation, imitation-learning, action-prediction\nPublished: 2023-05-19\nAuthors: Hiroki Furuta, Kuang-Huei Lee, Ofir Nachum, Yutaka Matsuo, Aleksandra Faust, Shixiang Shane Gu, Izzeddin Gur\nRecorded arXiv version: v4\n\nPublication edition: Multimodal Web Navigation with Instruction-Finetuned Foundation Models\nPublication authors: Hiroki Furuta, Kuang-Huei Lee, Ofir Nachum, Yutaka Matsuo, Aleksandra Faust, Shixiang Shane Gu, Izzeddin Gur\nPublication edition date (recorded precision): 2024-05-31\nPublication venue: International Conference on Learning Representations\nPublication volume: 2024\nPublication pages: 29528-29558\nPublication date basis: Explicit publisher metadata: 2024-05-31; precision=day.\nPublication notes: The ICLR landing shortens Shixiang Shane Gu to Shixiang Gu. Publication authors use the full PDF title-page form. Volume 2024 and pages 29528–29558 are explicit publisher citation metadata, not values inferred from the event year or PDF page count. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2024/hash/7ef7d8359036afd8c2378d82c21058a4-Abstract-Conference.html\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2305.11854v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2305-11854/"}
{"id": "arxiv:2305.14257", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "896d5f5e9248e2e49da254fb383f84e5b6ce4286b6d30a37ed7944fcfa1f4958", "submitted_at": "2023-05-23T17:10:39Z", "updated_at": "2023-10-30T00:55:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.14257v3"}, "authors": ["Abishek Sridhar", "Robert Lo", "Frank F. Xu", "Hao Zhu", "Shuyan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2305.14257", "citation_url": "https://awesomecua.github.io/resources/arxiv-2305-14257/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "896d5f5e9248e2e49da254fb383f84e5b6ce4286b6d30a37ed7944fcfa1f4958", "submitted_at": "2023-05-23T17:10:39Z", "updated_at": "2023-10-30T00:55:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2305.14257v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2305.14257"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.14257"}], "name": "Hierarchical Prompting Assists Large Language Model on Web Navigation", "publication": null, "published_at": "2023-05-23", "record_sha256": "19c9ba1dc59513d49e7a1c2b27bdbc02a496ffe15979895776a28bf485fdc27d", "relations": [], "resource_id": "arxiv:2305.14257", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2305.14257v3", "topics": ["web-navigation", "planning", "efficiency"], "updated_at": "2023-10-30"}, "text": "Hierarchical Prompting Assists Large Language Model on Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates web navigation prompting into a SUMMARIZER that extracts action-relevant observations and an ACTOR that chooses the next action from the condensed context.\nPlatforms: web\nTopics: web-navigation, planning, efficiency\nPublished: 2023-05-23\nAuthors: Abishek Sridhar, Robert Lo, Frank F. Xu, Hao Zhu, Shuyan Zhou\nRecorded arXiv version: v3\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2305.14257v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2305-14257/"}
{"id": "arxiv:2305.16653", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c254a26ebb8e1bcdca70b7581c780aa6b0dd4e5f5ded85e456f7e8fb8b38fcc6", "submitted_at": "2023-05-26T05:52:27Z", "updated_at": "2023-05-26T05:52:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.16653v1"}, "authors": ["Haotian Sun", "Yuchen Zhuang", "Lingkai Kong", "Bo Dai", "Chao Zhang"], "canonical_source_url": "https://arxiv.org/abs/2305.16653", "citation_url": "https://awesomecua.github.io/resources/arxiv-2305-16653/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c254a26ebb8e1bcdca70b7581c780aa6b0dd4e5f5ded85e456f7e8fb8b38fcc6", "submitted_at": "2023-05-26T05:52:27Z", "updated_at": "2023-05-26T05:52:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2305.16653v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2305.16653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.16653"}], "name": "AdaPlanner: Adaptive Planning from Feedback with Language Models", "publication": null, "published_at": "2023-05-26", "record_sha256": "407830e9f8f5790c8336ef18391d969d8ee5d24a68b4aaab88f29d161ab4c10a", "relations": [], "resource_id": "arxiv:2305.16653", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2305.16653v1", "topics": ["planning", "self-reflection", "evaluation"], "updated_at": "2023-05-26"}, "text": "AdaPlanner: Adaptive Planning from Feedback with Language Models\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces AdaPlanner, which revises code-structured plans in response to environmental feedback and reuses successful plans as exemplars. Experiments include MiniWoB++ web tasks alongside ALFWorld.\nPlatforms: web\nTopics: planning, self-reflection, evaluation\nPublished: 2023-05-26\nAuthors: Haotian Sun, Yuchen Zhuang, Lingkai Kong, Bo Dai, Chao Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is a general planning framework; MiniWoB++ supplies the interface-agent evidence, while ALFWorld is a separate environment.\n\nPrimary source: https://arxiv.org/abs/2305.16653v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2305-16653/"}
{"id": "arxiv:2305.19308", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "320ff40872677e9d32efae360dc289ca345076e037e7044188868f765e816008", "submitted_at": "2023-05-30T17:59:30Z", "updated_at": "2023-10-30T06:36:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.19308v2"}, "authors": ["Hongxin Li", "Jingran Su", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2305.19308", "citation_url": "https://awesomecua.github.io/resources/arxiv-2305-19308/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "320ff40872677e9d32efae360dc289ca345076e037e7044188868f765e816008", "submitted_at": "2023-05-30T17:59:30Z", "updated_at": "2023-10-30T06:36:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2305.19308v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "beb8c811bf2d070845aa6fcc7e94ffd43ca4f9f60ce42c72d958598790fbdffc", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md"}], "identifiers": {"arxiv": "2305.19308"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2305.19308"}], "name": "SheetCopilot: Bringing Software Productivity to the Next Level through Large Language Models", "publication": null, "published_at": "2023-05-30", "record_sha256": "99100b0915c3f4ee2b3da68766dfd4c6ee63c14e665f614b12a46c029fda6b65", "relations": [], "resource_id": "arxiv:2305.19308", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2305.19308v2", "topics": ["workflow", "tool-integration", "evaluation"], "updated_at": "2023-10-30"}, "text": "SheetCopilot: Bringing Software Productivity to the Next Level through Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SheetCopilot, an agent that steers spreadsheet software via atomic virtual APIs, along with a 221-task spreadsheet manipulation benchmark.\nPlatforms: desktop, windows\nTopics: workflow, tool-integration, evaluation\nPublished: 2023-05-30\nAuthors: Hongxin Li, Jingran Su, Yuntao Chen, Qing Li, Zhaoxiang Zhang\nRecorded arXiv version: v2\n\nThe contribution was reviewed from the retained paper abstract. The original Windows/Excel implementation is identified in the separately retained project README; the later Ubuntu evaluator is not treated as a paper result. No results were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2305.19308v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2305-19308/"}
{"id": "arxiv:2306.00245", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d06f7480695f45546b62b5b8c3d8c174ca52936e158d46e626689d0918711417", "submitted_at": "2023-05-31T23:39:18Z", "updated_at": "2023-12-06T23:46:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.00245v2"}, "authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "canonical_source_url": "https://arxiv.org/abs/2306.00245", "citation_url": "https://awesomecua.github.io/resources/arxiv-2306-00245/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d06f7480695f45546b62b5b8c3d8c174ca52936e158d46e626689d0918711417", "submitted_at": "2023-05-31T23:39:18Z", "updated_at": "2023-12-06T23:46:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.00245v2"}, "publication": {"authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "date_basis": "Explicit publisher metadata: 2023-12-15; precision=day.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-1490", "event_dates": null, "evidence_urls": ["https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The NeurIPS landing metadata adds the initial N to Kristina Toutanova; the formal paper title page does not. Publication authors use the PDF. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "34354-34370", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "type": "conference-paper", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "938e0a31512bb9a7793b568858f81eb3ceb650a71aae2d91b5292990e890201b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:47+00:00", "role": "primary", "sha256": "3827298f44d2f24f1333b4d46ecb5eb88c8b028d192b11424580655b38e024f9", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2306.00245"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.00245"}, {"rel": "publication", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"}], "name": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "publication": {"authors": ["Peter Shaw", "Mandar Joshi", "James Cohan", "Jonathan Berant", "Panupong Pasupat", "Hexiang Hu", "Urvashi Khandelwal", "Kenton Lee", "Kristina Toutanova"], "date_basis": "Explicit publisher metadata: 2023-12-15; precision=day.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-1490", "event_dates": null, "evidence_urls": ["https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The NeurIPS landing metadata adds the initial N to Kristina Toutanova; the formal paper title page does not. Publication authors use the PDF. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "34354-34370", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6c52a8a4fadc9129c6e1d1745f2dfd0f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces", "type": "conference-paper", "url": "https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}, "published_at": "2023-05-31", "record_sha256": "2eccd714c167cb140e4e0b7b1424ef5fd40621a65d194d148806f699e38d8fe8", "relations": [], "resource_id": "arxiv:2306.00245", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2306.00245v2", "topics": ["grounding", "action-prediction", "web-navigation"], "updated_at": "2023-12-06"}, "text": "From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nStudies instruction-following agents that observe screenshots and act through a generic mouse-and-keyboard action space, evaluating pixel-based pretraining for GUI tasks on MiniWoB++.\nPlatforms: web\nTopics: grounding, action-prediction, web-navigation\nPublished: 2023-05-31\nAuthors: Peter Shaw, Mandar Joshi, James Cohan, Jonathan Berant, Panupong Pasupat, Hexiang Hu, Urvashi Khandelwal, Kenton Lee, Kristina Toutanova\nRecorded arXiv version: v2\n\nPublication edition: From Pixels to UI Actions: Learning to Follow Instructions via Graphical User Interfaces\nPublication authors: Peter Shaw, Mandar Joshi, James Cohan, Jonathan Berant, Panupong Pasupat, Hexiang Hu, Urvashi Khandelwal, Kenton Lee, Kristina Toutanova\nPublication edition date (recorded precision): 2023-12-15\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 36\nPublication pages: 34354-34370\nPublication doi: 10.52202/075280-1490\nPublication date basis: Explicit publisher metadata: 2023-12-15; precision=day.\nPublication notes: The NeurIPS landing metadata adds the initial N to Kristina Toutanova; the formal paper title page does not. Publication authors use the PDF. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://papers.nips.cc/paper_files/paper/2023/hash/6c52a8a4fadc9129c6e1d1745f2dfd0f-Abstract-Conference.html\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2306.00245v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2306-00245/"}
{"id": "arxiv:2306.01242", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "631a39d2929c68bb93a7f95af22c17d2ab208668fd97d9082fe0601a73d123dd", "submitted_at": "2023-06-02T02:42:58Z", "updated_at": "2023-12-04T13:36:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.01242v2"}, "authors": ["Zhizheng Zhang", "Xiaoyi Zhang", "Wenxuan Xie", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2306.01242", "citation_url": "https://awesomecua.github.io/resources/arxiv-2306-01242/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "631a39d2929c68bb93a7f95af22c17d2ab208668fd97d9082fe0601a73d123dd", "submitted_at": "2023-06-02T02:42:58Z", "updated_at": "2023-12-04T13:36:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2306.01242v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2306.01242"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.01242"}], "name": "Responsible Task Automation: Empowering Large Language Models as Responsible Task Automators", "publication": null, "published_at": "2023-06-02", "record_sha256": "f911296ef79fa1261f79cc0ffb4c8f67f7d6906eea30c6f21e858e73b357eac9", "relations": [], "resource_id": "arxiv:2306.01242", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2306.01242v2", "topics": ["safety-security", "evaluation", "automation"], "updated_at": "2023-12-04"}, "text": "Responsible Task Automation: Empowering Large Language Models as Responsible Task Automators\nResource type: paper\nReview status: curated\nScope: adjacent\n\nPresents ResponsibleTA, a coordinator-executor framework that checks command feasibility and task completion while using local memory for privacy-related controls. The paper compares prompted and learned implementations and evaluates them on UI task automation.\nPlatforms: unspecified\nTopics: safety-security, evaluation, automation\nPublished: 2023-06-02\nAuthors: Zhizheng Zhang, Xiaoyi Zhang, Wenxuan Xie, Yan Lu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. UI task automation is stated in the abstract, but its platform and detailed evaluation protocol are not identified.\n\nPrimary source: https://arxiv.org/abs/2306.01242v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2306-01242/"}
{"id": "arxiv:2306.06070", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de055dfa452130f3576839c99c007053c62a0c96c9bfbcdd79ecad1a9f74f830", "submitted_at": "2023-06-09T17:44:31Z", "updated_at": "2023-12-09T05:57:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.06070v3"}, "authors": ["Xiang Deng", "Yu Gu", "Boyuan Zheng", "Shijie Chen", "Samuel Stevens", "Boshi Wang", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2306.06070", "citation_url": "https://awesomecua.github.io/resources/arxiv-2306-06070/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de055dfa452130f3576839c99c007053c62a0c96c9bfbcdd79ecad1a9f74f830", "submitted_at": "2023-06-09T17:44:31Z", "updated_at": "2023-12-09T05:57:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.06070v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2306.06070"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.06070"}], "name": "Mind2Web: Towards a Generalist Agent for the Web", "publication": null, "published_at": "2023-06-09", "record_sha256": "43b14354201afd688aff283e54b1af2423e2617b086722b9f7fe88fe9aafe818", "relations": [], "resource_id": "arxiv:2306.06070", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2306.06070v3", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2023-12-09"}, "text": "Mind2Web: Towards a Generalist Agent for the Web\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a web-task dataset spanning diverse websites and evaluates generalization to unseen sites and domains.\nPlatforms: web\nTopics: data-generation, web-navigation, evaluation\nPublished: 2023-06-09\nAuthors: Xiang Deng, Yu Gu, Boyuan Zheng, Shijie Chen, Samuel Stevens, Boshi Wang, Huan Sun, Yu Su\nRecorded arXiv version: v3\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2306.06070v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2306-06070/"}
{"id": "arxiv:2306.07863", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fdb5fed2b2fa0ecbf5910dd6813d8de7910af286a1e4e00cdebb734a82e78b0", "submitted_at": "2023-06-13T15:49:41Z", "updated_at": "2024-01-19T06:59:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.07863v3"}, "authors": ["Longtao Zheng", "Rundong Wang", "Xinrun Wang", "Bo An"], "canonical_source_url": "https://arxiv.org/abs/2306.07863", "citation_url": "https://awesomecua.github.io/resources/arxiv-2306-07863/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fdb5fed2b2fa0ecbf5910dd6813d8de7910af286a1e4e00cdebb734a82e78b0", "submitted_at": "2023-06-13T15:49:41Z", "updated_at": "2024-01-19T06:59:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2306.07863v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2306.07863"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2306.07863"}], "name": "Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control", "publication": null, "published_at": "2023-06-13", "record_sha256": "30a8bc47209acf1e39eaf4219a733b6a4b367db3dda7fb8941737c9804d24957", "relations": [], "resource_id": "arxiv:2306.07863", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2306.07863v3", "topics": ["memory", "web-navigation", "planning", "efficiency"], "updated_at": "2024-01-19"}, "text": "Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control\nResource type: paper\nReview status: curated\nScope: core\n\nSynapse filters task-irrelevant page state and places complete abstracted state-action trajectories into the prompt. An embedding-based exemplar memory retrieves relevant demonstrations for new tasks, evaluated on MiniWoB++ and Mind2Web.\nPlatforms: web\nTopics: memory, web-navigation, planning, efficiency\nPublished: 2023-06-13\nAuthors: Longtao Zheng, Rundong Wang, Xinrun Wang, Bo An\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2306.07863v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2306-07863/"}
{"id": "arxiv:2307.10088", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "def4616f7cb13fe839e500c4b3171eafbe37b155b5d14301137806d6a5cde939", "submitted_at": "2023-07-19T15:57:24Z", "updated_at": "2023-10-27T14:24:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.10088v2"}, "authors": ["Christopher Rawles", "Alice Li", "Daniel Rodriguez", "Oriana Riva", "Timothy Lillicrap"], "canonical_source_url": "https://arxiv.org/abs/2307.10088", "citation_url": "https://awesomecua.github.io/resources/arxiv-2307-10088/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "def4616f7cb13fe839e500c4b3171eafbe37b155b5d14301137806d6a5cde939", "submitted_at": "2023-07-19T15:57:24Z", "updated_at": "2023-10-27T14:24:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2307.10088v2"}, "publication": {"authors": ["Christopher Rawles", "Alice Li", "Daniel Rodriguez", "Oriana Riva", "Timothy Lillicrap"], "date_basis": "Explicit publisher citation_publication_date: 2023-12-15. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-2609", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "59708-59728", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/bbbb6308b402fe909c39dd29950c32e0-Paper-Datasets_and_Benchmarks.pdf", "publisher": null, "series": null, "title": "AndroidInTheWild: A Large-Scale Dataset For Android Device Control", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "7fa75eae1a17878ef28414f64e1c46cc83539008f2644331b762248ec96efce5", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2307.10088"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.10088"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/bbbb6308b402fe909c39dd29950c32e0-Paper-Datasets_and_Benchmarks.pdf"}], "name": "Android in the Wild: A Large-Scale Dataset for Android Device Control", "publication": {"authors": ["Christopher Rawles", "Alice Li", "Daniel Rodriguez", "Oriana Riva", "Timothy Lillicrap"], "date_basis": "Explicit publisher citation_publication_date: 2023-12-15. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2023, 12, 15], "doi": "10.52202/075280-2609", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "59708-59728", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2023/file/bbbb6308b402fe909c39dd29950c32e0-Paper-Datasets_and_Benchmarks.pdf", "publisher": null, "series": null, "title": "AndroidInTheWild: A Large-Scale Dataset For Android Device Control", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html", "venue": "Advances in Neural Information Processing Systems", "volume": "36"}, "published_at": "2023-07-19", "record_sha256": "66db4203b60d301044b2c4f2806fa51eccf0d72df07f0fc05fc861982142a5c8", "relations": [], "resource_id": "arxiv:2307.10088", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2307.10088v2", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2023-10-27"}, "text": "Android in the Wild: A Large-Scale Dataset for Android Device Control\nResource type: paper\nReview status: curated\nScope: core\n\nCollects Android interaction demonstrations with screens, actions, and instructions across device and platform variations for robustness evaluation.\nPlatforms: android\nTopics: data-generation, mobile-navigation, evaluation\nPublished: 2023-07-19\nAuthors: Christopher Rawles, Alice Li, Daniel Rodriguez, Oriana Riva, Timothy Lillicrap\nRecorded arXiv version: v2\n\nPublication edition: AndroidInTheWild: A Large-Scale Dataset For Android Device Control\nPublication authors: Christopher Rawles, Alice Li, Daniel Rodriguez, Oriana Riva, Timothy Lillicrap\nPublication edition date (recorded precision): 2023-12-15\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 36\nPublication pages: 59708-59728\nPublication doi: 10.52202/075280-2609\nPublication date basis: Explicit publisher citation_publication_date: 2023-12-15. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2023/hash/bbbb6308b402fe909c39dd29950c32e0-Abstract-Datasets_and_Benchmarks.html\n\nAnnotation reviewed against the primary abstract. Original dataset scale and experimental results are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2307.10088v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2307-10088/"}
{"id": "arxiv:2307.12856", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65328d525a6458ef8c401fb0ac4a9b6a5142bad462bd15a212d124c83a56209f", "submitted_at": "2023-07-24T14:56:30Z", "updated_at": "2024-02-25T16:17:43Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.12856v4"}, "authors": ["Izzeddin Gur", "Hiroki Furuta", "Austin Huang", "Mustafa Safdari", "Yutaka Matsuo", "Douglas Eck", "Aleksandra Faust"], "canonical_source_url": "https://arxiv.org/abs/2307.12856", "citation_url": "https://awesomecua.github.io/resources/arxiv-2307-12856/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65328d525a6458ef8c401fb0ac4a9b6a5142bad462bd15a212d124c83a56209f", "submitted_at": "2023-07-24T14:56:30Z", "updated_at": "2024-02-25T16:17:43Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.12856v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2307.12856"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.12856"}], "name": "A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesis", "publication": null, "published_at": "2023-07-24", "record_sha256": "d2a697261031996a1720f07f18b59336a99c1c0a046b263c68650e457da64bc2", "relations": [], "resource_id": "arxiv:2307.12856", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2307.12856v4", "topics": ["web-navigation", "planning"], "updated_at": "2024-02-25"}, "text": "A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nCombines task decomposition, HTML summarization, and generated Python programs to carry out instructions on websites.\nPlatforms: web\nTopics: web-navigation, planning\nPublished: 2023-07-24\nAuthors: Izzeddin Gur, Hiroki Furuta, Austin Huang, Mustafa Safdari, Yutaka Matsuo, Douglas Eck, Aleksandra Faust\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2307.12856v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2307-12856/"}
{"id": "arxiv:2307.13854", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e9eca0165573ba53b663acbb1c7962fd2d64bc6633f10397f9cad229772a363", "submitted_at": "2023-07-25T22:59:32Z", "updated_at": "2024-04-16T15:13:18Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.13854v4"}, "authors": ["Shuyan Zhou", "Frank F. Xu", "Hao Zhu", "Xuhui Zhou", "Robert Lo", "Abishek Sridhar", "Xianyi Cheng", "Tianyue Ou", "Yonatan Bisk", "Daniel Fried", "Uri Alon", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2307.13854", "citation_url": "https://awesomecua.github.io/resources/arxiv-2307-13854/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e9eca0165573ba53b663acbb1c7962fd2d64bc6633f10397f9cad229772a363", "submitted_at": "2023-07-25T22:59:32Z", "updated_at": "2024-04-16T15:13:18Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2307.13854v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}], "identifiers": {"arxiv": "2307.13854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2307.13854"}], "name": "WebArena: A Realistic Web Environment for Building Autonomous Agents", "publication": null, "published_at": "2023-07-25", "record_sha256": "5862033a9c69b1d3c40b9a402162e4e24e5bae2dd5e434bc9838d62af71d02a6", "relations": [], "resource_id": "arxiv:2307.13854", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2307.13854v4", "topics": ["evaluation", "web-navigation", "long-horizon"], "updated_at": "2024-04-16"}, "text": "WebArena: A Realistic Web Environment for Building Autonomous Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces reproducible web environments with task-completion checks for evaluating autonomous web agents.\nPlatforms: web\nTopics: evaluation, web-navigation, long-horizon\nPublished: 2023-07-25\nAuthors: Shuyan Zhou, Frank F. Xu, Hao Zhu, Xuhui Zhou, Robert Lo, Abishek Sridhar, Xianyi Cheng, Tianyue Ou, Yonatan Bisk, Daniel Fried, Uri Alon, Graham Neubig\nRecorded arXiv version: v4\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2307.13854v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2307-13854/"}
{"id": "arxiv:2308.03688", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8c4ba4db4aeed3a7b7ce7f77d610d7e34b0e8db26a36463e2c0bdffcad8f22a", "submitted_at": "2023-08-07T16:08:11Z", "updated_at": "2025-10-04T03:54:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2308.03688v3"}, "authors": ["Xiao Liu", "Hao Yu", "Hanchen Zhang", "Yifan Xu", "Xuanyu Lei", "Hanyu Lai", "Yu Gu", "Hangliang Ding", "Kaiwen Men", "Kejuan Yang", "Shudan Zhang", "Xiang Deng", "Aohan Zeng", "Zhengxiao Du", "Chenhui Zhang", "Sheng Shen", "Tianjun Zhang", "Yu Su", "Huan Sun", "Minlie Huang", "Yuxiao Dong", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2308.03688", "citation_url": "https://awesomecua.github.io/resources/arxiv-2308-03688/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8c4ba4db4aeed3a7b7ce7f77d610d7e34b0e8db26a36463e2c0bdffcad8f22a", "submitted_at": "2023-08-07T16:08:11Z", "updated_at": "2025-10-04T03:54:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2308.03688v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2308.03688"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2308.03688"}], "name": "AgentBench: Evaluating LLMs as Agents", "publication": null, "published_at": "2023-08-07", "record_sha256": "499212b502fdf6a886893b287d9c0e1da99f5b456bf6213fbb25e48705449122", "relations": [], "resource_id": "arxiv:2308.03688", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2308.03688v3", "topics": ["evaluation", "planning"], "updated_at": "2025-10-04"}, "text": "AgentBench: Evaluating LLMs as Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nEvaluates language agents across eight interactive environments and analyzes failures in reasoning, decisions, and instruction following.\nPlatforms: cross-platform\nTopics: evaluation, planning\nPublished: 2023-08-07\nAuthors: Xiao Liu, Hao Yu, Hanchen Zhang, Yifan Xu, Xuanyu Lei, Hanyu Lai, Yu Gu, Hangliang Ding, Kaiwen Men, Kejuan Yang, Shudan Zhang, Xiang Deng, Aohan Zeng, Zhengxiao Du, Chenhui Zhang, Sheng Shen, Tianjun Zhang, Yu Su, Huan Sun, Minlie Huang, Yuxiao Dong, Jie Tang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2308.03688v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2308-03688/"}
{"id": "arxiv:2308.15272", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a103fe10850a39d25bbb2b4a9ea0593d686938243bb912c674fd906e754c1fb", "submitted_at": "2023-08-29T13:02:30Z", "updated_at": "2024-03-09T09:38:51Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.15272v4"}, "authors": ["Hao Wen", "Yuanchun Li", "Guohong Liu", "Shanhui Zhao", "Tao Yu", "Toby Jia-Jun Li", "Shiqi Jiang", "Yunhao Liu", "Yaqin Zhang", "Yunxin Liu"], "canonical_source_url": "https://arxiv.org/abs/2308.15272", "citation_url": "https://awesomecua.github.io/resources/arxiv-2308-15272/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a103fe10850a39d25bbb2b4a9ea0593d686938243bb912c674fd906e754c1fb", "submitted_at": "2023-08-29T13:02:30Z", "updated_at": "2024-03-09T09:38:51Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2308.15272v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2308.15272"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2308.15272"}], "name": "AutoDroid: LLM-powered Task Automation in Android", "publication": null, "published_at": "2023-08-29", "record_sha256": "aa27b20131ff745c2607f7688938f4be3fd52320f94ce01fd2d302e230e81f00", "relations": [], "resource_id": "arxiv:2308.15272", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2308.15272v4", "topics": ["mobile-navigation", "memory"], "updated_at": "2024-03-09"}, "text": "AutoDroid: LLM-powered Task Automation in Android\nResource type: paper\nReview status: curated\nScope: core\n\nCombines language-model reasoning with app analysis, UI representations, and exploration-derived memory for Android task automation.\nPlatforms: android\nTopics: mobile-navigation, memory\nPublished: 2023-08-29\nAuthors: Hao Wen, Yuanchun Li, Guohong Liu, Shanhui Zhao, Tao Yu, Toby Jia-Jun Li, Shiqi Jiang, Yunhao Liu, Yaqin Zhang, Yunxin Liu\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2308.15272v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2308-15272/"}
{"id": "arxiv:2309.08172", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "399c97382cac325a9d60a073a14221dee6ee6254cb0b8aa88a5341a7bb758c97", "submitted_at": "2023-09-15T05:44:08Z", "updated_at": "2024-02-21T17:42:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.08172v2"}, "authors": ["Kaixin Ma", "Hongming Zhang", "Hongwei Wang", "Xiaoman Pan", "Wenhao Yu", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2309.08172", "citation_url": "https://awesomecua.github.io/resources/arxiv-2309-08172/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "399c97382cac325a9d60a073a14221dee6ee6254cb0b8aa88a5341a7bb758c97", "submitted_at": "2023-09-15T05:44:08Z", "updated_at": "2024-02-21T17:42:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2309.08172v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2309.08172"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2309.08172"}], "name": "LASER: LLM Agent with State-Space Exploration for Web Navigation", "publication": null, "published_at": "2023-09-15", "record_sha256": "e01e4864f5b8c8b999d11d7f0b31af15f5d944c2670df207c67d6b992df887c4", "relations": [], "resource_id": "arxiv:2309.08172", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2309.08172v2", "topics": ["planning", "robustness", "web-navigation"], "updated_at": "2024-02-21"}, "text": "LASER: LLM Agent with State-Space Exploration for Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nLASER represents web navigation as exploration over a defined set of states, allowing an agent to revisit earlier states after mistakes. The study evaluates this backtracking approach on WebShop and Amazon tasks.\nPlatforms: web\nTopics: planning, robustness, web-navigation\nPublished: 2023-09-15\nAuthors: Kaixin Ma, Hongming Zhang, Hongwei Wang, Xiaoman Pan, Wenhao Yu, Dong Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2309.08172v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2309-08172/"}
{"id": "arxiv:2309.11436", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d9cff1ece6115e87b7a9032f5720deb49694fbf570bea1f152018c4fa80d981", "submitted_at": "2023-09-20T16:12:32Z", "updated_at": "2024-06-07T04:52:29Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2309.11436v4"}, "authors": ["Zhuosheng Zhang", "Aston Zhang"], "canonical_source_url": "https://arxiv.org/abs/2309.11436", "citation_url": "https://awesomecua.github.io/resources/arxiv-2309-11436/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d9cff1ece6115e87b7a9032f5720deb49694fbf570bea1f152018c4fa80d981", "submitted_at": "2023-09-20T16:12:32Z", "updated_at": "2024-06-07T04:52:29Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2309.11436v4"}, "publication": {"authors": ["Zhuosheng Zhang", "Aston Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.186", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.186/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3132-3149", "pdf_url": "https://aclanthology.org/2024.findings-acl.186.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "You Only Look at Screens: Multimodal Chain-of-Action Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.186/", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "80fb0185db71f714655c22f5e1eb46d890c045b8af09c49d1b5ee2e0b6911478", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-acl.186/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2309.11436"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2309.11436"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-acl.186/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-acl.186.pdf"}], "name": "You Only Look at Screens: Multimodal Chain-of-Action Agents", "publication": {"authors": ["Zhuosheng Zhang", "Aston Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.186", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.186/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3132-3149", "pdf_url": "https://aclanthology.org/2024.findings-acl.186.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "You Only Look at Screens: Multimodal Chain-of-Action Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.186/", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}, "published_at": "2023-09-20", "record_sha256": "3fae95ddcda0faddd23732d70500e0556ffd7410affdb87932a39b27507de462", "relations": [], "resource_id": "arxiv:2309.11436", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2309.11436v4", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-06-07"}, "text": "You Only Look at Screens: Multimodal Chain-of-Action Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Auto-GUI, a screenshot-based agent using previous actions and future action plans to guide device interaction.\nPlatforms: android\nTopics: mobile-navigation, planning\nPublished: 2023-09-20\nAuthors: Zhuosheng Zhang, Aston Zhang\nRecorded arXiv version: v4\n\nPublication edition: You Only Look at Screens: Multimodal Chain-of-Action Agents\nPublication authors: Zhuosheng Zhang, Aston Zhang\nPublication edition date (recorded precision): 2024-08\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 3132-3149\nPublication doi: 10.18653/v1/2024.findings-acl.186\nPublication date basis: Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2024.findings-acl.186/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2309.11436v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2309-11436/"}
{"id": "arxiv:2310.03720", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d843f3c65e5961da773ec0265c2cfbf51d7c8214d9232ecad1638cc0dbb9ef64", "submitted_at": "2023-10-05T17:40:09Z", "updated_at": "2024-08-08T18:00:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.03720v4"}, "authors": ["Paloma Sodhi", "S. R. K. Branavan", "Yoav Artzi", "Ryan McDonald"], "canonical_source_url": "https://arxiv.org/abs/2310.03720", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-03720/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d843f3c65e5961da773ec0265c2cfbf51d7c8214d9232ecad1638cc0dbb9ef64", "submitted_at": "2023-10-05T17:40:09Z", "updated_at": "2024-08-08T18:00:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2310.03720v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2310.03720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.03720"}], "name": "SteP: Stacked LLM Policies for Web Actions", "publication": null, "published_at": "2023-10-05", "record_sha256": "bbd40ab88181ff085462df07b7ba85010f6f52ef23473e205ce8c026d399bcf8", "relations": [], "resource_id": "arxiv:2310.03720", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.03720v4", "topics": ["planning", "web-navigation", "workflow"], "updated_at": "2024-08-08"}, "text": "SteP: Stacked LLM Policies for Web Actions\nResource type: paper\nReview status: curated\nScope: core\n\nSteP dynamically composes specialized web-action policies through a stack that records the active chain of policy calls. It studies this control structure across WebArena, MiniWoB++, and CRM workflows.\nPlatforms: web\nTopics: planning, web-navigation, workflow\nPublished: 2023-10-05\nAuthors: Paloma Sodhi, S. R. K. Branavan, Yoav Artzi, Ryan McDonald\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.03720v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-03720/"}
{"id": "arxiv:2310.04716", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f389de80bb86a759df2cfc7000188286efcd73c99a43ede376c46f190832ef80", "submitted_at": "2023-10-07T07:22:41Z", "updated_at": "2023-10-07T07:22:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04716v1"}, "authors": ["Zhizheng Zhang", "Wenxuan Xie", "Xiaoyi Zhang", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2310.04716", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-04716/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f389de80bb86a759df2cfc7000188286efcd73c99a43ede376c46f190832ef80", "submitted_at": "2023-10-07T07:22:41Z", "updated_at": "2023-10-07T07:22:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04716v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04716"}], "name": "Reinforced UI Instruction Grounding: Towards a Generic UI Task Automation API", "publication": null, "published_at": "2023-10-07", "record_sha256": "993c1bc96b90b70102582996ae41769fab221ba3e5344a65b7cce027eabf0517", "relations": [], "resource_id": "arxiv:2310.04716", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.04716v1", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2023-10-07"}, "text": "Reinforced UI Instruction Grounding: Towards a Generic UI Task Automation API\nResource type: paper\nReview status: curated\nScope: core\n\nTrains a screenshot-based instruction-grounding model to predict target coordinates, using reinforcement learning to improve spatial decoding.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2023-10-07\nAuthors: Zhizheng Zhang, Wenxuan Xie, Xiaoyi Zhang, Yan Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.04716v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-04716/"}
{"id": "arxiv:2310.04869", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a49e879891017071fce4df0ff4b07bd9e8b5a6e14385af565549aadbde1bc1cf", "submitted_at": "2023-10-07T16:32:34Z", "updated_at": "2023-10-07T16:32:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04869v1"}, "authors": ["Yue Jiang", "Eldon Schoop", "Amanda Swearngin", "Jeffrey Nichols"], "canonical_source_url": "https://arxiv.org/abs/2310.04869", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-04869/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a49e879891017071fce4df0ff4b07bd9e8b5a6e14385af565549aadbde1bc1cf", "submitted_at": "2023-10-07T16:32:34Z", "updated_at": "2023-10-07T16:32:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.04869v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.04869"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.04869"}], "name": "ILuvUI: Instruction-tuned LangUage-Vision modeling of UIs from Machine Conversations", "publication": null, "published_at": "2023-10-07", "record_sha256": "df591995ed0504f875381a5e32957afccbd531816677a2debfd0b8a5ed3cd28a", "relations": [], "resource_id": "arxiv:2310.04869", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.04869v1", "topics": ["data-generation", "screen-understanding", "grounding", "planning"], "updated_at": "2023-10-07"}, "text": "ILuvUI: Instruction-tuned LangUage-Vision modeling of UIs from Machine Conversations\nResource type: paper\nReview status: curated\nScope: core\n\nILuvUI generates 335,000 conversational training examples paired with UI screenshots, covering questions, descriptions, and plans. The resulting vision-language model is assessed on element detection, response quality, and examples of multi-step UI navigation.\nPlatforms: unspecified\nTopics: data-generation, screen-understanding, grounding, planning\nPublished: 2023-10-07\nAuthors: Yue Jiang, Eldon Schoop, Amanda Swearngin, Jeffrey Nichols\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.04869v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-04869/"}
{"id": "arxiv:2310.05126", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "39d5cf803efe3abaf70a9b78483bb0e750a8741adec1c4b7f3c05b9754d88560", "submitted_at": "2023-10-08T11:33:09Z", "updated_at": "2023-10-08T11:33:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.05126v1"}, "authors": ["Jiabo Ye", "Anwen Hu", "Haiyang Xu", "Qinghao Ye", "Ming Yan", "Guohai Xu", "Chenliang Li", "Junfeng Tian", "Qi Qian", "Ji Zhang", "Qin Jin", "Liang He", "Xin Alex Lin", "Fei Huang"], "canonical_source_url": "https://arxiv.org/abs/2310.05126", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-05126/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "39d5cf803efe3abaf70a9b78483bb0e750a8741adec1c4b7f3c05b9754d88560", "submitted_at": "2023-10-08T11:33:09Z", "updated_at": "2023-10-08T11:33:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.05126v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.05126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.05126"}], "name": "UReader: Universal OCR-free Visually-situated Language Understanding with Multimodal Large Language Model", "publication": null, "published_at": "2023-10-08", "record_sha256": "3c83ad93378e92d171235a1f52b0d3520ff027d1c0dc1adedcb9b4e999cd162d", "relations": [], "resource_id": "arxiv:2310.05126", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2310.05126v1", "topics": ["screen-understanding", "efficiency"], "updated_at": "2023-10-08"}, "text": "UReader: Universal OCR-free Visually-situated Language Understanding with Multimodal Large Language Model\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces UReader, which combines unified instruction tuning, auxiliary text-reading tasks and shape-adaptive image crops for OCR-free visual language understanding. Its evaluation includes webpage screenshots alongside documents, tables, charts and natural images.\nPlatforms: web\nTopics: screen-understanding, efficiency\nPublished: 2023-10-08\nAuthors: Jiabo Ye, Anwen Hu, Haiyang Xu, Qinghao Ye, Ming Yan, Guohai Xu, Chenliang Li, Junfeng Tian, Qi Qian, Ji Zhang, Qin Jin, Liang He, Xin Alex Lin, Fei Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The interface connection is static webpage-screenshot understanding; browser control is not claimed.\n\nPrimary source: https://arxiv.org/abs/2310.05126v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-05126/"}
{"id": "arxiv:2310.08740", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95092986184a232af62c61fa2aedefb837851dce9e0ff7a1128a67d47edfde40", "submitted_at": "2023-10-12T21:53:37Z", "updated_at": "2023-10-23T17:39:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.08740v3"}, "authors": ["Tao Li", "Gang Li", "Zhiwei Deng", "Bryan Wang", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2310.08740", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-08740/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95092986184a232af62c61fa2aedefb837851dce9e0ff7a1128a67d47edfde40", "submitted_at": "2023-10-12T21:53:37Z", "updated_at": "2023-10-23T17:39:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2310.08740v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.08740"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.08740"}], "name": "A Zero-Shot Language Agent for Computer Control with Structured Reflection", "publication": null, "published_at": "2023-10-12", "record_sha256": "fc77f466525ea3ee5dceb46c6f2300f8dcefb2b080c71f834f0adc026110c506", "relations": [], "resource_id": "arxiv:2310.08740", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.08740v3", "topics": ["self-reflection", "planning", "web-navigation"], "updated_at": "2023-10-23"}, "text": "A Zero-Shot Language Agent for Computer Control with Structured Reflection\nResource type: paper\nReview status: curated\nScope: core\n\nStudies a zero-shot computer-control agent that plans executable actions and organizes its own reflections after mistakes. MiniWoB++ experiments examine learning to complete tasks without supplied expert trajectories.\nPlatforms: web\nTopics: self-reflection, planning, web-navigation\nPublished: 2023-10-12\nAuthors: Tao Li, Gang Li, Zhiwei Deng, Bryan Wang, Yang Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.08740v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-08740/"}
{"id": "arxiv:2310.10634", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15d7945fcf1ee4290761d34bbeef9e24bed82bb748bdb7aaa791457c42892fec", "submitted_at": "2023-10-16T17:54:53Z", "updated_at": "2023-10-16T17:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.10634v1"}, "authors": ["Tianbao Xie", "Fan Zhou", "Zhoujun Cheng", "Peng Shi", "Luoxuan Weng", "Yitao Liu", "Toh Jing Hua", "Junning Zhao", "Qian Liu", "Che Liu", "Leo Z. Liu", "Yiheng Xu", "Hongjin Su", "Dongchan Shin", "Caiming Xiong", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2310.10634", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-10634/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15d7945fcf1ee4290761d34bbeef9e24bed82bb748bdb7aaa791457c42892fec", "submitted_at": "2023-10-16T17:54:53Z", "updated_at": "2023-10-16T17:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.10634v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "90df698b27a336e2fb7f0ac60a3b8d3e3897c00035ab7968c593a5ab539d177a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.10634"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.10634"}], "name": "OpenAgents: An Open Platform for Language Agents in the Wild", "publication": null, "published_at": "2023-10-16", "record_sha256": "3a552a6a14049838cfb0b5736cac2821bbd89b26f9b99c02f87ced96d6ee6966", "relations": [], "resource_id": "arxiv:2310.10634", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2310.10634v1", "topics": ["web-navigation", "tool-integration"], "updated_at": "2023-10-16"}, "text": "OpenAgents: An Open Platform for Language Agents in the Wild\nResource type: paper\nReview status: curated\nScope: adjacent\n\nOpenAgents hosts three language agents, for data analysis, API plugins and autonomous web browsing, behind a web interface for non-expert users, with local deployment for developers.\nPlatforms: web\nTopics: web-navigation, tool-integration\nPublished: 2023-10-16\nAuthors: Tianbao Xie, Fan Zhou, Zhoujun Cheng, Peng Shi, Luoxuan Weng, Yitao Liu, Toh Jing Hua, Junning Zhao, Qian Liu, Che Liu, Leo Z. Liu, Yiheng Xu, Hongjin Su, Dongchan Shin, Caiming Xiong, Tao Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Scope is adjacent: autonomous web browsing is one of three hosted agents, beside data-analysis and API-plugin agents.\n\nPrimary source: https://arxiv.org/abs/2310.10634v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-10634/"}
{"id": "arxiv:2310.11441", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abafe2b974ddd80416b215cd4c48d801e18fe9162b9c097d819d42eaff68a069", "submitted_at": "2023-10-17T17:51:31Z", "updated_at": "2023-11-06T07:39:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.11441v2"}, "authors": ["Jianwei Yang", "Hao Zhang", "Feng Li", "Xueyan Zou", "Chunyuan Li", "Jianfeng Gao"], "canonical_source_url": "https://arxiv.org/abs/2310.11441", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-11441/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abafe2b974ddd80416b215cd4c48d801e18fe9162b9c097d819d42eaff68a069", "submitted_at": "2023-10-17T17:51:31Z", "updated_at": "2023-11-06T07:39:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.11441v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.11441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.11441"}], "name": "Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V", "publication": null, "published_at": "2023-10-17", "record_sha256": "5e274c197a4657f5e29cfc7a6e79265dffe9f672d2a6d812025befd2cb2873ec", "relations": [], "resource_id": "arxiv:2310.11441", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2310.11441v2", "topics": ["grounding"], "updated_at": "2023-11-06"}, "text": "Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V\nResource type: paper\nReview status: curated\nScope: adjacent\n\nOverlays labeled image regions to help multimodal models refer to and reason about visual targets; a general visual-prompting method relevant to GUI grounding.\nPlatforms: unspecified\nTopics: grounding\nPublished: 2023-10-17\nAuthors: Jianwei Yang, Hao Zhang, Feng Li, Xueyan Zou, Chunyuan Li, Jianfeng Gao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.11441v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-11441/"}
{"id": "arxiv:2310.14495", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1648acae2228b8ae2f042d7986585a66bc6a649320d569b12fd093b96c15cb5c", "submitted_at": "2023-10-23T02:00:55Z", "updated_at": "2023-10-23T02:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.14495v1"}, "authors": ["Justin Payan", "Swaroop Mishra", "Mukul Singh", "Carina Negreanu", "Christian Poelitz", "Chitta Baral", "Subhro Roy", "Rasika Chakravarthy", "Benjamin Van Durme", "Elnaz Nouri"], "canonical_source_url": "https://arxiv.org/abs/2310.14495", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-14495/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1648acae2228b8ae2f042d7986585a66bc6a649320d569b12fd093b96c15cb5c", "submitted_at": "2023-10-23T02:00:55Z", "updated_at": "2023-10-23T02:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2310.14495v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:30Z", "role": "primary", "sha256": "4856a9c146f74964be623730f3cee6b15fcaad83f7d83a4956e971b541a63070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2304.08479%2C2305.09699%2C2305.09770%2C2305.11271%2C2305.14343%2C2305.16338%2C2305.16985%2C2306.04529%2C2306.05499%2C2306.07179%2C2306.11134%2C2307.00804%2C2307.02443%2C2308.01552%2C2308.05391%2C2308.05822%2C2308.06595%2C2308.07201%2C2308.09289%2C2309.04561%2C2309.05251%2C2309.07550%2C2309.09958%2C2309.10094%2C2309.13965%2C2310.07433%2C2310.07899%2C2310.09103%2C2310.12670%2C2310.12793%2C2310.12815%2C2310.12945%2C2310.12972%2C2310.13345%2C2310.14495%2C2310.17596%2C2310.19805%2C2311.01011%2C2311.07215%2C2311.09227%2C2311.10538%2C2311.10751%2C2311.11261%2C2311.11855%2C2311.18492%2C2311.18760%2C2312.02976%2C2312.03815%2C2312.07424%2C2312.09238&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:11:47Z", "role": "primary", "sha256": "40fd8a31083963f1cbbcfeb308244494139277e4a6d0e56c2ebe5a7cda821e7c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2310.14495v1"}], "identifiers": {"arxiv": "2310.14495"}, "kind": "paper", "last_verified_at": "2026-10-17T09:11:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.14495"}], "name": "InstructExcel: A Benchmark for Natural Language Instruction in Excel", "publication": null, "published_at": "2023-10-23", "record_sha256": "9ee351c19580b630b1ea3ec87496048f222b5a9e103d6f6b050dde7f28d16239", "relations": [], "resource_id": "arxiv:2310.14495", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2310.14495v1", "topics": ["evaluation", "workflow", "automation"], "updated_at": "2023-10-23"}, "text": "InstructExcel: A Benchmark for Natural Language Instruction in Excel\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces InstructExcel, a benchmark for generating Excel OfficeScripts from natural-language requests, built using scripts recorded from user actions and a collection of existing spreadsheets.\nPlatforms: unspecified\nTopics: evaluation, workflow, automation\nPublished: 2023-10-23\nAuthors: Justin Payan, Swaroop Mishra, Mukul Singh, Carina Negreanu, Christian Poelitz, Chitta Baral, Subhro Roy, Rasika Chakravarthy, Benjamin Van Durme, Elnaz Nouri\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. This is an application-specific automation benchmark grounded in recorded Excel user actions. Its execution interface is OfficeScripts, so it is adjacent rather than a visual GUI navigation benchmark.\n\nPrimary source: https://arxiv.org/abs/2310.14495v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-14495/"}
{"id": "arxiv:2310.16042", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7defced3e3beb77a5a6513dbfdff3f102edd922907b523bd062522e54720fa5e", "submitted_at": "2023-10-24T17:57:03Z", "updated_at": "2023-10-25T03:54:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.16042v2"}, "authors": ["Heyi Tao", "Sethuraman T", "Michal Shlapentokh-Rothman", "Derek Hoiem"], "canonical_source_url": "https://arxiv.org/abs/2310.16042", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-16042/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7defced3e3beb77a5a6513dbfdff3f102edd922907b523bd062522e54720fa5e", "submitted_at": "2023-10-24T17:57:03Z", "updated_at": "2023-10-25T03:54:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.16042v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.16042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.16042"}], "name": "WebWISE: Web Interface Control and Sequential Exploration with Large Language Models", "publication": null, "published_at": "2023-10-24", "record_sha256": "a012f0942caa07dce5913fb60b70193681def8dae854f96022cb0ea94f14e44b", "relations": [], "resource_id": "arxiv:2310.16042", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.16042v2", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2023-10-25"}, "text": "WebWISE: Web Interface Control and Sequential Exploration with Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nWebWISE filters DOM elements into observations and generates short programs for click, scroll, and text-entry steps. It evaluates in-context learning from either one supplied example or an example collected during a successful zero-shot MiniWoB++ trial.\nPlatforms: web\nTopics: web-navigation, tool-integration, planning\nPublished: 2023-10-24\nAuthors: Heyi Tao, Sethuraman T, Michal Shlapentokh-Rothman, Derek Hoiem\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.16042v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-16042/"}
{"id": "arxiv:2310.18331", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c72eab40cdc28981cf83a753d88f5181b5441d82cd568265457fef96d75b36b6", "submitted_at": "2023-10-20T11:10:14Z", "updated_at": "2023-10-31T06:25:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.18331v2"}, "authors": ["Jiarun Liu", "Wentao Hu", "Chunhong Zhang"], "canonical_source_url": "https://arxiv.org/abs/2310.18331", "citation_url": "https://awesomecua.github.io/resources/arxiv-2310-18331/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c72eab40cdc28981cf83a753d88f5181b5441d82cd568265457fef96d75b36b6", "submitted_at": "2023-10-20T11:10:14Z", "updated_at": "2023-10-31T06:25:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2310.18331v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2310.18331"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2310.18331"}], "name": "AllTogether: Investigating the Efficacy of Spliced Prompt for Web Navigation using Large Language Models", "publication": null, "published_at": "2023-10-20", "record_sha256": "7c187f30f9074ef02391801a739da9f9f8ee3ac49512eb244b1014377f0a3454", "relations": [], "resource_id": "arxiv:2310.18331", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2310.18331v2", "topics": ["web-navigation", "interaction-history", "evaluation"], "updated_at": "2023-10-31"}, "text": "AllTogether: Investigating the Efficacy of Spliced Prompt for Web Navigation using Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nAllTogether standardizes prompts for HTML-based web navigation and compares prompting with instruction fine-tuning. Its analysis examines how HTML context, interaction history, and current environmental feedback affect task execution.\nPlatforms: web\nTopics: web-navigation, interaction-history, evaluation\nPublished: 2023-10-20\nAuthors: Jiarun Liu, Wentao Hu, Chunhong Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2310.18331v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2310-18331/"}
{"id": "arxiv:2311.01767", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58dfa75fd82a2cc1d41954f10feb1d987de6469782d11d04c3b8bfffc8a3b745", "submitted_at": "2023-11-03T08:06:35Z", "updated_at": "2023-11-07T10:13:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.01767v2"}, "authors": ["Yiduo Guo", "Zekai Zhang", "Yaobo Liang", "Dongyan Zhao", "Nan Duan"], "canonical_source_url": "https://arxiv.org/abs/2311.01767", "citation_url": "https://awesomecua.github.io/resources/arxiv-2311-01767/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "58dfa75fd82a2cc1d41954f10feb1d987de6469782d11d04c3b8bfffc8a3b745", "submitted_at": "2023-11-03T08:06:35Z", "updated_at": "2023-11-07T10:13:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2311.01767v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2311.01767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.01767"}], "name": "PPTC Benchmark: Evaluating Large Language Models for PowerPoint Task Completion", "publication": null, "published_at": "2023-11-03", "record_sha256": "4bdc5c6bb9ba54e7e2a4296f07e91f9abeb53be0cd82bd5e8916add0dcc89617", "relations": [], "resource_id": "arxiv:2311.01767", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2311.01767v2", "topics": ["evaluation", "tool-integration"], "updated_at": "2023-11-07"}, "text": "PPTC Benchmark: Evaluating Large Language Models for PowerPoint Task Completion\nResource type: paper\nReview status: curated\nScope: adjacent\n\nEvaluates multi-turn PowerPoint creation and editing, scoring the resulting presentation files rather than requiring a fixed API sequence.\nPlatforms: unspecified\nTopics: evaluation, tool-integration\nPublished: 2023-11-03\nAuthors: Yiduo Guo, Zekai Zhang, Yaobo Liang, Dongyan Zhao, Nan Duan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2311.01767v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2311-01767/"}
{"id": "arxiv:2311.07562", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebe6ff09f876b81d033a064301a1c23e3c141d2a1ad0d7a5df91e729377fcd3b", "submitted_at": "2023-11-13T18:53:37Z", "updated_at": "2023-11-13T18:53:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.07562v1"}, "authors": ["An Yan", "Zhengyuan Yang", "Wanrong Zhu", "Kevin Lin", "Linjie Li", "Jianfeng Wang", "Jianwei Yang", "Yiwu Zhong", "Julian McAuley", "Jianfeng Gao", "Zicheng Liu", "Lijuan Wang"], "canonical_source_url": "https://arxiv.org/abs/2311.07562", "citation_url": "https://awesomecua.github.io/resources/arxiv-2311-07562/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebe6ff09f876b81d033a064301a1c23e3c141d2a1ad0d7a5df91e729377fcd3b", "submitted_at": "2023-11-13T18:53:37Z", "updated_at": "2023-11-13T18:53:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.07562v1"}}, "environments": ["ios", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2311.07562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.07562"}], "name": "GPT-4V in Wonderland: Large Multimodal Models for Zero-Shot Smartphone GUI Navigation", "publication": null, "published_at": "2023-11-13", "record_sha256": "84071df1f1c16064a4bf08d0cf66e068ed43d69e259978252d4b05ba2ea112c2", "relations": [], "resource_id": "arxiv:2311.07562", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2311.07562v1", "topics": ["mobile-navigation", "grounding"], "updated_at": "2023-11-13"}, "text": "GPT-4V in Wonderland: Large Multimodal Models for Zero-Shot Smartphone GUI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nStudies MM-Navigator, a multimodal agent that interprets smartphone screenshots and predicts localized actions without task-specific training.\nPlatforms: ios, android\nTopics: mobile-navigation, grounding\nPublished: 2023-11-13\nAuthors: An Yan, Zhengyuan Yang, Wanrong Zhu, Kevin Lin, Linjie Li, Jianfeng Wang, Jianwei Yang, Yiwu Zhong, Julian McAuley, Jianfeng Gao, Zicheng Liu, Lijuan Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2311.07562v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2311-07562/"}
{"id": "arxiv:2311.12983", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29ddf2b7b4d495ee18bbd263cd77f86a71b3014b58a6f140c3bfce7cbae9c922", "submitted_at": "2023-11-21T20:34:47Z", "updated_at": "2023-11-21T20:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.12983v1"}, "authors": ["Grégoire Mialon", "Clémentine Fourrier", "Craig Swift", "Thomas Wolf", "Yann LeCun", "Thomas Scialom"], "canonical_source_url": "https://arxiv.org/abs/2311.12983", "citation_url": "https://awesomecua.github.io/resources/arxiv-2311-12983/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29ddf2b7b4d495ee18bbd263cd77f86a71b3014b58a6f140c3bfce7cbae9c922", "submitted_at": "2023-11-21T20:34:47Z", "updated_at": "2023-11-21T20:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2311.12983v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2311.12983"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.12983"}], "name": "GAIA: a benchmark for General AI Assistants", "publication": null, "published_at": "2023-11-21", "record_sha256": "c21ba8558b86ba3b6cd791f95cb9955ee04fd97ddea34ab744d133d510249cb0", "relations": [], "resource_id": "arxiv:2311.12983", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2311.12983v1", "topics": ["evaluation", "tool-integration"], "updated_at": "2023-11-21"}, "text": "GAIA: a benchmark for General AI Assistants\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces questions that test assistant reasoning, multimodal understanding, browsing, and tool use beyond screen-action execution alone.\nPlatforms: web\nTopics: evaluation, tool-integration\nPublished: 2023-11-21\nAuthors: Grégoire Mialon, Clémentine Fourrier, Craig Swift, Thomas Wolf, Yann LeCun, Thomas Scialom\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2311.12983v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2311-12983/"}
{"id": "arxiv:2311.18751", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b00e0a9c74b7ac8819bdda540a1069c33cbac45f0c7fb853dee995f2feb9a59", "submitted_at": "2023-11-30T17:50:47Z", "updated_at": "2024-12-31T04:24:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.18751v3"}, "authors": ["Hiroki Furuta", "Yutaka Matsuo", "Aleksandra Faust", "Izzeddin Gur"], "canonical_source_url": "https://arxiv.org/abs/2311.18751", "citation_url": "https://awesomecua.github.io/resources/arxiv-2311-18751/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b00e0a9c74b7ac8819bdda540a1069c33cbac45f0c7fb853dee995f2feb9a59", "submitted_at": "2023-11-30T17:50:47Z", "updated_at": "2024-12-31T04:24:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2311.18751v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2311.18751"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2311.18751"}], "name": "Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web", "publication": null, "published_at": "2023-11-30", "record_sha256": "2e60df85fc9eb2545302b448de5246c6c1585d7315a8167c8467567524ff3994", "relations": [], "resource_id": "arxiv:2311.18751", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2311.18751v3", "topics": ["evaluation", "long-horizon", "robustness", "web-navigation"], "updated_at": "2024-12-31"}, "text": "Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web\nResource type: paper\nReview status: curated\nScope: core\n\nCompWoB adds 50 composed web-automation tasks to study generalization from simpler interactions. The paper compares prompted and fine-tuned agents, balances task training data, and probes sensitivity to changes in instruction order.\nPlatforms: web\nTopics: evaluation, long-horizon, robustness, web-navigation\nPublished: 2023-11-30\nAuthors: Hiroki Furuta, Yutaka Matsuo, Aleksandra Faust, Izzeddin Gur\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2311.18751v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2311-18751/"}
{"id": "arxiv:2312.03003", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9b6f10b17513f565a13e035802d8bd681b5b0d468ba000df1427a8d59591e2c", "submitted_at": "2023-12-04T06:13:35Z", "updated_at": "2024-10-16T08:15:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.03003v3"}, "authors": ["Sunjae Lee", "Junyoung Choi", "Jungjae Lee", "Munim Hasan Wasi", "Hojun Choi", "Steven Y. Ko", "Sangeun Oh", "Insik Shin"], "canonical_source_url": "https://arxiv.org/abs/2312.03003", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-03003/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9b6f10b17513f565a13e035802d8bd681b5b0d468ba000df1427a8d59591e2c", "submitted_at": "2023-12-04T06:13:35Z", "updated_at": "2024-10-16T08:15:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.03003v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.03003"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.03003"}], "name": "Explore, Select, Derive, and Recall: Augmenting LLM with Human-like Memory for Mobile Task Automation", "publication": null, "published_at": "2023-12-04", "record_sha256": "191d3f00a4bdb9a4f818ca6cc5f8501505441ce0602b2591300662de84008dfa", "relations": [], "resource_id": "arxiv:2312.03003", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.03003v3", "topics": ["mobile-navigation", "memory", "planning"], "updated_at": "2024-10-16"}, "text": "Explore, Select, Derive, and Recall: Augmenting LLM with Human-like Memory for Mobile Task Automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MobileGPT, which organizes mobile-app experience into reusable subtasks using an explore, select, derive, and recall cycle. Its memory supports learning task procedures and adapting them to new contexts.\nPlatforms: unspecified\nTopics: mobile-navigation, memory, planning\nPublished: 2023-12-04\nAuthors: Sunjae Lee, Junyoung Choi, Jungjae Lee, Munim Hasan Wasi, Hojun Choi, Steven Y. Ko, Sangeun Oh, Insik Shin\nRecorded arXiv version: v3\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.03003v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-03003/"}
{"id": "arxiv:2312.06677", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "598c8c30348d454c638944c932dd55355b5d3ecda6eb6f3910a27fffcab9622a", "submitted_at": "2023-12-04T07:51:58Z", "updated_at": "2023-12-04T07:51:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.06677v1"}, "authors": ["Yanchu Guan", "Dong Wang", "Zhixuan Chu", "Shiyu Wang", "Feiyue Ni", "Ruihua Song", "Longfei Li", "Jinjie Gu", "Chenyi Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2312.06677", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-06677/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "598c8c30348d454c638944c932dd55355b5d3ecda6eb6f3910a27fffcab9622a", "submitted_at": "2023-12-04T07:51:58Z", "updated_at": "2023-12-04T07:51:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.06677v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.06677"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.06677"}], "name": "Intelligent Virtual Assistants with LLM-based Process Automation", "publication": null, "published_at": "2023-12-04", "record_sha256": "bf55a772fdd9799c44dcb458ca5596b24b05a1c25e071a9563434428339388fe", "relations": [], "resource_id": "arxiv:2312.06677", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.06677v1", "topics": ["mobile-navigation", "planning", "automation"], "updated_at": "2023-12-04"}, "text": "Intelligent Virtual Assistants with LLM-based Process Automation\nResource type: paper\nReview status: curated\nScope: core\n\nPresents LLMPA, a mobile process-automation architecture that decomposes requests, identifies interface elements, predicts actions and checks errors. Demonstrations evaluate multistep operations in Alipay.\nPlatforms: unspecified\nTopics: mobile-navigation, planning, automation\nPublished: 2023-12-04\nAuthors: Yanchu Guan, Dong Wang, Zhixuan Chu, Shiyu Wang, Feiyue Ni, Ruihua Song, Longfei Li, Jinjie Gu, Chenyi Zhuang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract names Alipay but does not establish Android or iOS specifically.\n\nPrimary source: https://arxiv.org/abs/2312.06677v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-06677/"}
{"id": "arxiv:2312.08914", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755198d83a3af9fe6d032f1446ec9bf57bf7e6f314d37787ce8a6aaa16bddc47", "submitted_at": "2023-12-14T13:20:57Z", "updated_at": "2024-12-27T06:56:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.08914v3"}, "authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxuan Zhang", "Juanzi Li", "Bin Xu", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2312.08914", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-08914/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755198d83a3af9fe6d032f1446ec9bf57bf7e6f314d37787ce8a6aaa16bddc47", "submitted_at": "2023-12-14T13:20:57Z", "updated_at": "2024-12-27T06:56:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.08914v3"}, "publication": {"authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2024, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, and landing page list eleven authors; the retained arXiv snapshot lists fourteen. Each edition retains its own authors.", "pages": "14281-14290", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf", "publisher": null, "series": null, "title": "CogAgent: A Visual Language Model for GUI Agents", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["desktop", "android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "43009fbc322dbe4f6cd1be01b7f414b4747ddcadaf4b63ae16f8eecc7ca6837b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:11.606232+00:00", "role": "primary", "sha256": "b9bfea178ee72b694eb575c73e3d7d8209a72706c9dda04296a70e17b0c98cf4", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:37:01Z", "role": "primary", "sha256": "8bfa074910574410d7988771270feded31e7e1bd1fd0924d4754850d4a5f1914", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=900&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2312.08914"}, "kind": "paper", "last_verified_at": "2026-10-16T06:37:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.08914"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"}], "name": "CogAgent: A Visual Language Model for GUI Agents", "publication": {"authors": ["Wenyi Hong", "Weihan Wang", "Qingsong Lv", "Jiazheng Xu", "Wenmeng Yu", "Junhui Ji", "Yan Wang", "Zihan Wang", "Yuxiao Dong", "Ming Ding", "Jie Tang"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2024, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, and landing page list eleven authors; the retained arXiv snapshot lists fourteen. Each edition retains its own authors.", "pages": "14281-14290", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2024/papers/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.pdf", "publisher": null, "series": null, "title": "CogAgent: A Visual Language Model for GUI Agents", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}, "published_at": "2023-12-14", "record_sha256": "2009e01fd8045f322b18e887cb3db989207661c52fd68d8ec66d8cac4412c5d5", "relations": [], "resource_id": "arxiv:2312.08914", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.08914v3", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-12-27"}, "text": "CogAgent: A Visual Language Model for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a visual-language model with low- and high-resolution image encoders for screenshot understanding and GUI navigation.\nPlatforms: desktop, android, web\nTopics: screen-understanding, grounding\nPublished: 2023-12-14\nAuthors: Wenyi Hong, Weihan Wang, Qingsong Lv, Jiazheng Xu, Wenmeng Yu, Junhui Ji, Yan Wang, Zihan Wang, Yuxuan Zhang, Juanzi Li, Bin Xu, Yuxiao Dong, Ming Ding, Jie Tang\nRecorded arXiv version: v3\n\nPublication edition: CogAgent: A Visual Language Model for GUI Agents\nPublication authors: Wenyi Hong, Weihan Wang, Qingsong Lv, Jiazheng Xu, Wenmeng Yu, Junhui Ji, Yan Wang, Zihan Wang, Yuxiao Dong, Ming Ding, Jie Tang\nPublication edition date (recorded precision): 2024-06\nPublication venue: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition\nPublication pages: 14281-14290\nPublication date basis: Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, and landing page list eleven authors; the retained arXiv snapshot lists fourteen. Each edition retains its own authors.\nPublication source: https://openaccess.thecvf.com/content/CVPR2024/html/Hong_CogAgent_A_Visual_Language_Model_for_GUI_Agents_CVPR_2024_paper.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.08914v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-08914/"}
{"id": "arxiv:2312.11190", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df1390eab35db18dedaf135ddbbc9fdfcacbc9a70c51ec2609cf331cc05e06bd", "submitted_at": "2023-12-18T13:35:46Z", "updated_at": "2024-07-29T14:52:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.11190v2"}, "authors": ["Yunpeng Song", "Yiheng Bian", "Yongtao Tang", "Guiyu Ma", "Zhongmin Cai"], "canonical_source_url": "https://arxiv.org/abs/2312.11190", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-11190/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df1390eab35db18dedaf135ddbbc9fdfcacbc9a70c51ec2609cf331cc05e06bd", "submitted_at": "2023-12-18T13:35:46Z", "updated_at": "2024-07-29T14:52:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.11190v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2312.11190"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.11190"}], "name": "VisionTasker: Mobile Task Automation Using Vision Based UI Understanding and LLM Task Planning", "publication": null, "published_at": "2023-12-18", "record_sha256": "911d82ce64fb74b41533ab6fec0697d9b191c562dc79c2d033cef8d27c664660", "relations": [], "resource_id": "arxiv:2312.11190", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.11190v2", "topics": ["screen-understanding", "planning", "mobile-navigation"], "updated_at": "2024-07-29"}, "text": "VisionTasker: Mobile Task Automation Using Vision Based UI Understanding and LLM Task Planning\nResource type: paper\nReview status: curated\nScope: core\n\nVisionTasker converts screenshots into textual UI interpretations and presents one interface at a time to an LLM planner for action selection. It evaluates visual representations and execution of 147 tasks on an Android smartphone, including integration with programming by demonstration.\nPlatforms: android\nTopics: screen-understanding, planning, mobile-navigation\nPublished: 2023-12-18\nAuthors: Yunpeng Song, Yiheng Bian, Yongtao Tang, Guiyu Ma, Zhongmin Cai\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.11190v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-11190/"}
{"id": "arxiv:2312.13108", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c771b076044c170728041a4f54662d21493adcc3742d40db08cb2f9bd0614655", "submitted_at": "2023-12-20T15:28:38Z", "updated_at": "2024-01-01T14:26:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.13108v2"}, "authors": ["Difei Gao", "Lei Ji", "Zechen Bai", "Mingyu Ouyang", "Peiran Li", "Dongxing Mao", "Qinchen Wu", "Weichen Zhang", "Peiyi Wang", "Xiangwu Guo", "Hengxu Wang", "Luowei Zhou", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2312.13108", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-13108/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c771b076044c170728041a4f54662d21493adcc3742d40db08cb2f9bd0614655", "submitted_at": "2023-12-20T15:28:38Z", "updated_at": "2024-01-01T14:26:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2312.13108v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.13108"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.13108"}], "name": "ASSISTGUI: Task-Oriented Desktop Graphical User Interface Automation", "publication": null, "published_at": "2023-12-20", "record_sha256": "3f57eecbe68c1c25bb0500c5dc91bd737aa50dd797999264c5666dbd1f64e301", "relations": [], "resource_id": "arxiv:2312.13108", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.13108v2", "topics": ["evaluation", "planning"], "updated_at": "2024-01-01"}, "text": "ASSISTGUI: Task-Oriented Desktop Graphical User Interface Automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Windows application tasks and an actor-critic agent framework for mouse-and-keyboard automation of desktop workflows.\nPlatforms: desktop, windows\nTopics: evaluation, planning\nPublished: 2023-12-20\nAuthors: Difei Gao, Lei Ji, Zechen Bai, Mingyu Ouyang, Peiran Li, Dongxing Mao, Qinchen Wu, Weichen Zhang, Peiyi Wang, Xiangwu Guo, Hengxu Wang, Luowei Zhou, Mike Zheng Shou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.13108v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-13108/"}
{"id": "arxiv:2312.13771", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61e54d83304755ce89c2ab69f7bb93a7fdc4d68cb5182daad9b5210f809d7ae1", "submitted_at": "2023-12-21T11:52:45Z", "updated_at": "2026-07-05T07:51:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.13771v3"}, "authors": ["Chi Zhang", "Zhao Yang", "Jiaxuan Liu", "Yanda Li", "Yucheng Han", "Xin Chen", "Zebiao Huang", "Bin Fu", "Gang Yu"], "canonical_source_url": "https://arxiv.org/abs/2312.13771", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-13771/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61e54d83304755ce89c2ab69f7bb93a7fdc4d68cb5182daad9b5210f809d7ae1", "submitted_at": "2023-12-21T11:52:45Z", "updated_at": "2026-07-05T07:51:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2312.13771v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.13771"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.13771"}], "name": "AppAgent: Multimodal Agents as Smartphone Users", "publication": null, "published_at": "2023-12-21", "record_sha256": "cca8cb306e9d32edcd55161340ee4c58affcb3d7b593c0fac43cd8e32e5cfeaf", "relations": [], "resource_id": "arxiv:2312.13771", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.13771v3", "topics": ["mobile-navigation", "memory"], "updated_at": "2026-07-05"}, "text": "AppAgent: Multimodal Agents as Smartphone Users\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds app knowledge through exploration or human demonstrations and uses it to guide multimodal smartphone interactions.\nPlatforms: android\nTopics: mobile-navigation, memory\nPublished: 2023-12-21\nAuthors: Chi Zhang, Zhao Yang, Jiaxuan Liu, Yanda Li, Yucheng Han, Xin Chen, Zebiao Huang, Bin Fu, Gang Yu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.13771v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-13771/"}
{"id": "arxiv:2312.15820", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47dedddced8f06bf436ab5a101d1ec4ecf4380c710e457dc7a35fc86d2e1f230", "submitted_at": "2023-12-25T22:13:26Z", "updated_at": "2023-12-25T22:13:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.15820v1"}, "authors": ["Qi Chen", "Dileepa Pitawela", "Chongyang Zhao", "Gengze Zhou", "Hsiang-Ting Chen", "Qi Wu"], "canonical_source_url": "https://arxiv.org/abs/2312.15820", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-15820/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47dedddced8f06bf436ab5a101d1ec4ecf4380c710e457dc7a35fc86d2e1f230", "submitted_at": "2023-12-25T22:13:26Z", "updated_at": "2023-12-25T22:13:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.15820v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.15820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.15820"}], "name": "WebVLN: Vision-and-Language Navigation on Websites", "publication": null, "published_at": "2023-12-25", "record_sha256": "eae0381019e6aab06ca79f5d4c3372632d4e98426aa7e38728a3931ad6ce8c50", "relations": [], "resource_id": "arxiv:2312.15820", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.15820v1", "topics": ["web-navigation", "data-generation", "evaluation"], "updated_at": "2023-12-25"}, "text": "WebVLN: Vision-and-Language Navigation on Websites\nResource type: paper\nReview status: curated\nScope: core\n\nWebVLN trains website navigation from question-based instructions using both rendered visual information and underlying HTML. It introduces WebVLN-v1 and a website-aware navigation network to study this digital variant of vision-and-language navigation.\nPlatforms: web\nTopics: web-navigation, data-generation, evaluation\nPublished: 2023-12-25\nAuthors: Qi Chen, Dileepa Pitawela, Chongyang Zhao, Gengze Zhou, Hsiang-Ting Chen, Qi Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.15820v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-15820/"}
{"id": "arxiv:2312.16062", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f592712517da90bfeb807ab445034c3974b406e0ed6b4933fbd19078984b9b8", "submitted_at": "2023-12-26T14:20:36Z", "updated_at": "2023-12-26T14:20:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.16062v1"}, "authors": ["Lihang Pan", "Bowen Wang", "Chun Yu", "Yuxuan Chen", "Xiangyu Zhang", "Yuanchun Shi"], "canonical_source_url": "https://arxiv.org/abs/2312.16062", "citation_url": "https://awesomecua.github.io/resources/arxiv-2312-16062/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f592712517da90bfeb807ab445034c3974b406e0ed6b4933fbd19078984b9b8", "submitted_at": "2023-12-26T14:20:36Z", "updated_at": "2023-12-26T14:20:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2312.16062v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2312.16062"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2312.16062"}], "name": "AutoTask: Executing Arbitrary Voice Commands by Exploring and Learning from Mobile GUI", "publication": null, "published_at": "2023-12-26", "record_sha256": "632d93f7db1d671d201e6b33029a5d8efd2e85233a9903ab7cc6750a2922da8d", "relations": [], "resource_id": "arxiv:2312.16062", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2312.16062v1", "topics": ["mobile-navigation", "memory"], "updated_at": "2023-12-26"}, "text": "AutoTask: Executing Arbitrary Voice Commands by Exploring and Learning from Mobile GUI\nResource type: paper\nReview status: curated\nScope: core\n\nExecutes voice-requested mobile tasks through GUI exploration, backtracking, and accumulated interaction experience.\nPlatforms: android\nTopics: mobile-navigation, memory\nPublished: 2023-12-26\nAuthors: Lihang Pan, Bowen Wang, Chun Yu, Yuxuan Chen, Xiangyu Zhang, Yuanchun Shi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2312.16062v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2312-16062/"}
{"id": "arxiv:2401.01614", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a005fb14b21cf17521f4ea966e555d9963092b709bd9280c6e72a9bf0f5fa93a", "submitted_at": "2024-01-03T08:33:09Z", "updated_at": "2024-03-12T23:14:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.01614v2"}, "authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2401.01614", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-01614/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a005fb14b21cf17521f4ea966e555d9963092b709bd9280c6e72a9bf0f5fa93a", "submitted_at": "2024-01-03T08:33:09Z", "updated_at": "2024-03-12T23:14:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.01614v2"}, "publication": {"authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/zheng24e.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "61349-61385", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/zheng24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "eb144b5dfe51e5a9092d273fcffb85343808d011b03d22e059c9cc65e642ef33", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2401.01614"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.01614"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v235/zheng24e.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf"}], "name": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "publication": {"authors": ["Boyuan Zheng", "Boyu Gou", "Jihyung Kil", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/zheng24e.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "61349-61385", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/zheng24e/zheng24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "GPT-4V(ision) is a Generalist Web Agent, if Grounded", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/zheng24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}, "published_at": "2024-01-03", "record_sha256": "0dea08f3373757c9f5da63a74af40c5039e7b6db6e94e98cf5e07aeb9cb29d6c", "relations": [], "resource_id": "arxiv:2401.01614", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.01614v2", "topics": ["web-navigation", "grounding"], "updated_at": "2024-03-12"}, "text": "GPT-4V(ision) is a Generalist Web Agent, if Grounded\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SeeAct and live-website evaluation, examining how visual planning and element grounding affect web-agent task completion.\nPlatforms: web\nTopics: web-navigation, grounding\nPublished: 2024-01-03\nAuthors: Boyuan Zheng, Boyu Gou, Jihyung Kil, Huan Sun, Yu Su\nRecorded arXiv version: v2\n\nPublication edition: GPT-4V(ision) is a Generalist Web Agent, if Grounded\nPublication authors: Boyuan Zheng, Boyu Gou, Jihyung Kil, Huan Sun, Yu Su\nPublication edition date (recorded precision): 2024-07-08\nPublication venue: Proceedings of the 41st International Conference on Machine Learning\nPublication publisher: PMLR\nPublication volume: 235\nPublication pages: 61349-61385\nPublication date basis: Explicit publisher metadata: 2024/07/08; precision=day.\nPublication event dates: 21--27 Jul 2024\nPublication notes: The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://proceedings.mlr.press/v235/zheng24e.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.01614v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-01614/"}
{"id": "arxiv:2401.04124", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b8aaa2bccafd7f41802e5b7c25bc98f78307f77aeb36b7c75c57ef0f8e583a1", "submitted_at": "2024-01-04T03:44:42Z", "updated_at": "2024-01-17T06:35:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.04124v3"}, "authors": ["Tinghe Ding"], "canonical_source_url": "https://arxiv.org/abs/2401.04124", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-04124/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b8aaa2bccafd7f41802e5b7c25bc98f78307f77aeb36b7c75c57ef0f8e583a1", "submitted_at": "2024-01-04T03:44:42Z", "updated_at": "2024-01-17T06:35:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.04124v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2401.04124"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.04124"}], "name": "MobileAgent: enhancing mobile control via human-machine interaction and SOP integration", "publication": null, "published_at": "2024-01-04", "record_sha256": "4f830cfbffe73362d3f9b57e9c9f3ae7728a2fcfff627f464b31dad8c8b5aec7", "relations": [], "resource_id": "arxiv:2401.04124", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.04124v3", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": "2024-01-17"}, "text": "MobileAgent: enhancing mobile control via human-machine interaction and SOP integration\nResource type: paper\nReview status: curated\nScope: core\n\nCombines operating-procedure context with user interaction to guide mobile task automation and handle personalized or sensitive requests.\nPlatforms: android\nTopics: mobile-navigation, workflow, human-in-the-loop\nPublished: 2024-01-04\nAuthors: Tinghe Ding\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.04124v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-04124/"}
{"id": "arxiv:2401.10935", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c92ec9323eec9362aa07984f9fb1cbfbe98ecae426a9f8adc9d436328a87aed3", "submitted_at": "2024-01-17T08:10:35Z", "updated_at": "2024-02-23T04:36:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.10935v2"}, "authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "canonical_source_url": "https://arxiv.org/abs/2401.10935", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-10935/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c92ec9323eec9362aa07984f9fb1cbfbe98ecae426a9f8adc9d436328a87aed3", "submitted_at": "2024-01-17T08:10:35Z", "updated_at": "2024-02-23T04:36:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.10935v2"}, "publication": {"authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.505/", "https://aclanthology.org/2024.acl-long.505.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ACL landing metadata spells the fifth author Li YanTao. The formal PDF title page spells Yantao Li; the publication author field uses the PDF. The retained arXiv author list is untouched. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9313-9332", "pdf_url": "https://aclanthology.org/2024.acl-long.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.505/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "62d56e417487461c8b34da15167eafd5f14f58092e4269222dc5c1c9048731a9", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.505/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "8e1f6809e35751d762714286c55e79e81bae071608f90ef561d804aa59d7e406", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2024.acl-long.505.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2401.10935"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.10935"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.505/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.505.pdf"}], "name": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "publication": {"authors": ["Kanzhi Cheng", "Qiushi Sun", "Yougang Chu", "Fangzhi Xu", "Yantao Li", "Jianbing Zhang", "Zhiyong Wu"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.505/", "https://aclanthology.org/2024.acl-long.505.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The ACL landing metadata spells the fifth author Li YanTao. The formal PDF title page spells Yantao Li; the publication author field uses the PDF. The retained arXiv author list is untouched. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "9313-9332", "pdf_url": "https://aclanthology.org/2024.acl-long.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.505/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-01-17", "record_sha256": "dadda1bc72de0d6dfab8ca0cb6e0f8571cffafe8a194febc91ca401f34a0af14", "relations": [], "resource_id": "arxiv:2401.10935", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.10935v2", "topics": ["grounding", "data-generation"], "updated_at": "2024-02-23"}, "text": "SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains a screenshot-driven GUI agent with grounding pretraining and introduces ScreenSpot for evaluating interface target localization.\nPlatforms: cross-platform\nTopics: grounding, data-generation\nPublished: 2024-01-17\nAuthors: Kanzhi Cheng, Qiushi Sun, Yougang Chu, Fangzhi Xu, Yantao Li, Jianbing Zhang, Zhiyong Wu\nRecorded arXiv version: v2\n\nPublication edition: SeeClick: Harnessing GUI Grounding for Advanced Visual GUI Agents\nPublication authors: Kanzhi Cheng, Qiushi Sun, Yougang Chu, Fangzhi Xu, Yantao Li, Jianbing Zhang, Zhiyong Wu\nPublication edition date (recorded precision): 2024-08\nPublication venue: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 9313-9332\nPublication doi: 10.18653/v1/2024.acl-long.505\nPublication date basis: Explicit publisher metadata: 2024/8; precision=month.\nPublication notes: The ACL landing metadata spells the fifth author Li YanTao. The formal PDF title page spells Yantao Li; the publication author field uses the PDF. The retained arXiv author list is untouched. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://aclanthology.org/2024.acl-long.505/\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.10935v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-10935/"}
{"id": "arxiv:2401.13178", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22c5a60f9ed7a8488ec31fbf24d392e9674b0cae79ab1b269fd9d57202025801", "submitted_at": "2024-01-24T01:51:00Z", "updated_at": "2024-12-23T20:12:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13178v2"}, "authors": ["Chang Ma", "Junlei Zhang", "Zhihao Zhu", "Cheng Yang", "Yujiu Yang", "Yaohui Jin", "Zhenzhong Lan", "Lingpeng Kong", "Junxian He"], "canonical_source_url": "https://arxiv.org/abs/2401.13178", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-13178/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22c5a60f9ed7a8488ec31fbf24d392e9674b0cae79ab1b269fd9d57202025801", "submitted_at": "2024-01-24T01:51:00Z", "updated_at": "2024-12-23T20:12:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13178v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:30Z", "role": "primary", "sha256": "bbfef23bd53d2febb258a91d5a706f630cd597909bdbea56f16ac97dbe4777c0", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2401.13178v2"}], "identifiers": {"arxiv": "2401.13178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13178"}], "name": "AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents", "publication": null, "published_at": "2024-01-24", "record_sha256": "406812b3f9a2ed3f77a5cafe313728d129ebc4cf9977d87a6b60e1d60fa3c0fc", "relations": [], "resource_id": "arxiv:2401.13178", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2401.13178v2", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2024-12-23"}, "text": "AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces AgentBoard, a multiturn agent evaluation framework with intermediate progress metrics and analysis tools. Its broader task suite includes WebShop and WebArena interactions with stepwise progress measurements.\nPlatforms: web\nTopics: evaluation, observability, web-navigation\nPublished: 2024-01-24\nAuthors: Chang Ma, Junlei Zhang, Zhihao Zhu, Cheng Yang, Yujiu Yang, Yaohui Jin, Zhenzhong Lan, Lingpeng Kong, Junxian He\nRecorded arXiv version: v2\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The full paper documents search/click WebShop actions and 11 browser actions in WebArena, while the overall benchmark also covers non-web domains.\n\nPrimary source: https://arxiv.org/abs/2401.13178v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-13178/"}
{"id": "arxiv:2401.13649", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a705b87b43464120ebb01b28b0f6a9e31fad68f3d463c88354f2005dab940c3", "submitted_at": "2024-01-24T18:35:21Z", "updated_at": "2024-06-06T02:01:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13649v2"}, "authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "canonical_source_url": "https://arxiv.org/abs/2401.13649", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-13649/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a705b87b43464120ebb01b28b0f6a9e31fad68f3d463c88354f2005dab940c3", "submitted_at": "2024-01-24T18:35:21Z", "updated_at": "2024-06-06T02:01:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.13649v2"}, "publication": {"authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.50", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.50/", "https://aclanthology.org/2024.acl-long.50.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal PDF title adds Visually Grounded; the landing metadata uses Visual Web Tasks. PDF authors include Ming Chong Lim and Ruslan Salakhutdinov; the landing abbreviates them to Ming Lim and Russ Salakhutdinov. Publication title/authors use the PDF; the retained arXiv title remains unchanged. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "881-905", "pdf_url": "https://aclanthology.org/2024.acl-long.50.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visually Grounded Web Tasks", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.50/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "fc908d474b12a6c0f3ff3e48129d29c809e9f3a0365a3cf647b70273f266849a", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.50/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "6fc85c38c975c2af170c1d1a28f72ba557906d9019d1d9106a15215868bb491e", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://aclanthology.org/2024.acl-long.50.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2401.13649"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13649"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.50/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.50.pdf"}], "name": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks", "publication": {"authors": ["Jing Yu Koh", "Robert Lo", "Lawrence Jang", "Vikram Duvvur", "Ming Chong Lim", "Po-Yu Huang", "Graham Neubig", "Shuyan Zhou", "Ruslan Salakhutdinov", "Daniel Fried"], "date_basis": "Explicit publisher metadata: 2024/8; precision=month.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.50", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.50/", "https://aclanthology.org/2024.acl-long.50.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The formal PDF title adds Visually Grounded; the landing metadata uses Visual Web Tasks. PDF authors include Ming Chong Lim and Ruslan Salakhutdinov; the landing abbreviates them to Ming Lim and Russ Salakhutdinov. Publication title/authors use the PDF; the retained arXiv title remains unchanged. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "881-905", "pdf_url": "https://aclanthology.org/2024.acl-long.50.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visually Grounded Web Tasks", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.50/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-01-24", "record_sha256": "ed4c9faca0c3850113fe1cc8c02f90f1fec1a08fb9612b907fa83e7d11959229", "relations": [], "resource_id": "arxiv:2401.13649", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.13649v2", "topics": ["evaluation", "grounding", "web-navigation"], "updated_at": "2024-06-06"}, "text": "VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces visually grounded website tasks for evaluating multimodal agents that must interpret both images and text.\nPlatforms: web\nTopics: evaluation, grounding, web-navigation\nPublished: 2024-01-24\nAuthors: Jing Yu Koh, Robert Lo, Lawrence Jang, Vikram Duvvur, Ming Chong Lim, Po-Yu Huang, Graham Neubig, Shuyan Zhou, Ruslan Salakhutdinov, Daniel Fried\nRecorded arXiv version: v2\n\nPublication edition: VisualWebArena: Evaluating Multimodal Agents on Realistic Visually Grounded Web Tasks\nPublication authors: Jing Yu Koh, Robert Lo, Lawrence Jang, Vikram Duvvur, Ming Chong Lim, Po-Yu Huang, Graham Neubig, Shuyan Zhou, Ruslan Salakhutdinov, Daniel Fried\nPublication edition date (recorded precision): 2024-08\nPublication venue: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 881-905\nPublication doi: 10.18653/v1/2024.acl-long.50\nPublication date basis: Explicit publisher metadata: 2024/8; precision=month.\nPublication notes: The formal PDF title adds Visually Grounded; the landing metadata uses Visual Web Tasks. PDF authors include Ming Chong Lim and Ruslan Salakhutdinov; the landing abbreviates them to Ming Lim and Russ Salakhutdinov. Publication title/authors use the PDF; the retained arXiv title remains unchanged. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://aclanthology.org/2024.acl-long.50/\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.13649v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-13649/"}
{"id": "arxiv:2401.13919", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc7121c9606c5987df0941b849c216ad8731bebcc74a061cf3f154379d7b8da", "submitted_at": "2024-01-25T03:33:18Z", "updated_at": "2024-06-06T18:37:34Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2401.13919v4"}, "authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Yong Dai", "Hongming Zhang", "Zhenzhong Lan", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2401.13919", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-13919/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc7121c9606c5987df0941b849c216ad8731bebcc74a061cf3f154379d7b8da", "submitted_at": "2024-01-25T03:33:18Z", "updated_at": "2024-06-06T18:37:34Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2401.13919v4"}, "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Yong Dai", "Hongming Zhang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.371", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.371/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "6864-6890", "pdf_url": "https://aclanthology.org/2024.acl-long.371.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.371/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "9772713e428cc47e54a6a236aef9024e44c97c6794a746ade767b51bdba7679a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.371/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2401.13919"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.13919"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.371/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.371.pdf"}], "name": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models", "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Yong Dai", "Hongming Zhang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.371", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.371/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "6864-6890", "pdf_url": "https://aclanthology.org/2024.acl-long.371.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.371/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-01-25", "record_sha256": "b6baf557abc0b70347d216dcf55d1c3dcbed213050fe136e69a0ac976a49acf6", "relations": [], "resource_id": "arxiv:2401.13919", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.13919v4", "topics": ["web-navigation", "evaluation"], "updated_at": "2024-06-06"}, "text": "WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models\nResource type: paper\nReview status: curated\nScope: core\n\nCombines multimodal web interaction with a real-website task collection and an automated evaluation protocol.\nPlatforms: web\nTopics: web-navigation, evaluation\nPublished: 2024-01-25\nAuthors: Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Yong Dai, Hongming Zhang, Zhenzhong Lan, Dong Yu\nRecorded arXiv version: v4\n\nPublication edition: WebVoyager: Building an End-to-End Web Agent with Large Multimodal Models\nPublication authors: Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Yong Dai, Hongming Zhang, Zhenzhong Lan, Dong Yu\nPublication edition date (recorded precision): 2024-08\nPublication venue: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 6864-6890\nPublication doi: 10.18653/v1/2024.acl-long.371\nPublication date basis: Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2024.acl-long.371/\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.13919v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-13919/"}
{"id": "arxiv:2401.14268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1c95264a931b10a57ee4b364498e2d2795abaf9a3e89096a0f29d2408174693", "submitted_at": "2024-01-25T16:02:56Z", "updated_at": "2024-08-14T00:48:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.14268v3"}, "authors": ["Minh Duc Vu", "Han Wang", "Zhuang Li", "Jieshan Chen", "Shengdong Zhao", "Zhenchang Xing", "Chunyang Chen"], "canonical_source_url": "https://arxiv.org/abs/2401.14268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-14268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1c95264a931b10a57ee4b364498e2d2795abaf9a3e89096a0f29d2408174693", "submitted_at": "2024-01-25T16:02:56Z", "updated_at": "2024-08-14T00:48:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2401.14268v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2401.14268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.14268"}], "name": "GPTVoiceTasker: Advancing Multi-step Mobile Task Efficiency Through Dynamic Interface Exploration and Learning", "publication": null, "published_at": "2024-01-25", "record_sha256": "c340609077b68aaf514d978dd631eb2207571ed22f386fb189860539d280d11a", "relations": [], "resource_id": "arxiv:2401.14268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.14268v3", "topics": ["mobile-navigation", "human-in-the-loop", "memory"], "updated_at": "2024-08-14"}, "text": "GPTVoiceTasker: Advancing Multi-step Mobile Task Efficiency Through Dynamic Interface Exploration and Learning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GPTVoiceTasker, a mobile assistant that interprets user commands, executes device interactions, and learns from prior commands to streamline later task execution. The work includes a user study of task efficiency.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop, memory\nPublished: 2024-01-25\nAuthors: Minh Duc Vu, Han Wang, Zhuang Li, Jieshan Chen, Shengdong Zhao, Zhenchang Xing, Chunyang Chen\nRecorded arXiv version: v3\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.14268v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-14268/"}
{"id": "arxiv:2401.16158", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81ac041a3e5f98bfc23c91f768b32ad7adc3f28d8c9d55a285bb29906c8b5a9c", "submitted_at": "2024-01-29T13:46:37Z", "updated_at": "2024-04-18T06:53:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.16158v2"}, "authors": ["Junyang Wang", "Haiyang Xu", "Jiabo Ye", "Ming Yan", "Weizhou Shen", "Ji Zhang", "Fei Huang", "Jitao Sang"], "canonical_source_url": "https://arxiv.org/abs/2401.16158", "citation_url": "https://awesomecua.github.io/resources/arxiv-2401-16158/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81ac041a3e5f98bfc23c91f768b32ad7adc3f28d8c9d55a285bb29906c8b5a9c", "submitted_at": "2024-01-29T13:46:37Z", "updated_at": "2024-04-18T06:53:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2401.16158v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2401.16158"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2401.16158"}], "name": "Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perception", "publication": null, "published_at": "2024-01-29", "record_sha256": "f2f23b755d18f15961c9839925e88cde3365a36a188f530c06a2022ec5a2e164", "relations": [], "resource_id": "arxiv:2401.16158", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2401.16158v2", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-04-18"}, "text": "Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perception\nResource type: paper\nReview status: curated\nScope: core\n\nUses visual perception tools and task decomposition to operate mobile applications without relying on app interface metadata.\nPlatforms: android\nTopics: mobile-navigation, planning\nPublished: 2024-01-29\nAuthors: Junyang Wang, Haiyang Xu, Jiabo Ye, Ming Yan, Weizhou Shen, Ji Zhang, Fei Huang, Jitao Sang\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2401.16158v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2401-16158/"}
{"id": "arxiv:2402.04476", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3da421fc440f7dcfc53d8ef218f55389e4da4dbac196f54b40402bc877ff0365", "submitted_at": "2024-02-06T23:52:10Z", "updated_at": "2024-03-30T05:18:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.04476v2"}, "authors": ["Jihyung Kil", "Chan Hee Song", "Boyuan Zheng", "Xiang Deng", "Yu Su", "Wei-Lun Chao"], "canonical_source_url": "https://arxiv.org/abs/2402.04476", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-04476/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3da421fc440f7dcfc53d8ef218f55389e4da4dbac196f54b40402bc877ff0365", "submitted_at": "2024-02-06T23:52:10Z", "updated_at": "2024-03-30T05:18:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.04476v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.04476"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.04476"}], "name": "Dual-View Visual Contextualization for Web Navigation", "publication": null, "published_at": "2024-02-06", "record_sha256": "ccd7e71918df1ab361f574be6e684632868766dcb482b9b4a22091c690251474", "relations": [], "resource_id": "arxiv:2402.04476", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.04476v2", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2024-03-30"}, "text": "Dual-View Visual Contextualization for Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nUses screenshot geometry and visual content to enrich HTML element representations for web navigation. Nearby elements contribute textual and visual context, with evaluation on cross-task, cross-website, and cross-domain Mind2Web splits.\nPlatforms: web\nTopics: grounding, screen-understanding, web-navigation\nPublished: 2024-02-06\nAuthors: Jihyung Kil, Chan Hee Song, Boyuan Zheng, Xiang Deng, Yu Su, Wei-Lun Chao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.04476v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-04476/"}
{"id": "arxiv:2402.04615", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b293ab33fe0f3607eccc2b99cbf06fb459792444433a4179bb65e081f33e426c", "submitted_at": "2024-02-07T06:42:33Z", "updated_at": "2024-07-04T07:08:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.04615v3"}, "authors": ["Gilles Baechler", "Srinivas Sunkara", "Maria Wang", "Fedir Zubach", "Hassan Mansoor", "Vincent Etter", "Victor Cărbune", "Jason Lin", "Jindong Chen", "Abhanshu Sharma"], "canonical_source_url": "https://arxiv.org/abs/2402.04615", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-04615/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b293ab33fe0f3607eccc2b99cbf06fb459792444433a4179bb65e081f33e426c", "submitted_at": "2024-02-07T06:42:33Z", "updated_at": "2024-07-04T07:08:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.04615v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.04615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.04615"}], "name": "ScreenAI: A Vision-Language Model for UI and Infographics Understanding", "publication": null, "published_at": "2024-02-07", "record_sha256": "77a222acd1422a5fd17d88dde864934a8df736159bc759a24d407c872356410c", "relations": [], "resource_id": "arxiv:2402.04615", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.04615v3", "topics": ["screen-understanding", "data-generation"], "updated_at": "2024-07-04"}, "text": "ScreenAI: A Vision-Language Model for UI and Infographics Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ScreenAI and screen-annotation training that supports UI question answering, navigation, and summarization data generation.\nPlatforms: cross-platform\nTopics: screen-understanding, data-generation\nPublished: 2024-02-07\nAuthors: Gilles Baechler, Srinivas Sunkara, Maria Wang, Fedir Zubach, Hassan Mansoor, Vincent Etter, Victor Cărbune, Jason Lin, Jindong Chen, Abhanshu Sharma\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.04615v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-04615/"}
{"id": "arxiv:2402.05930", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a7c1a1f260f9287e8999525e1adc5534a70f7b96271d469973fceee1ac23609", "submitted_at": "2024-02-08T18:58:02Z", "updated_at": "2024-09-10T22:22:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.05930v2"}, "authors": ["Xing Han Lù", "Zdeněk Kasner", "Siva Reddy"], "canonical_source_url": "https://arxiv.org/abs/2402.05930", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-05930/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a7c1a1f260f9287e8999525e1adc5534a70f7b96271d469973fceee1ac23609", "submitted_at": "2024-02-08T18:58:02Z", "updated_at": "2024-09-10T22:22:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.05930v2"}, "publication": {"authors": ["Xing Han Lù", "Zdeněk Kasner", "Siva Reddy"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/lu24e.html", "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The PMLR landing renders Xing Han Lu; the paper title page has Xing Han Lù. Publication authors use the PDF. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "33007-33056", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/lu24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "1b30008f833b3b06238d87d0509a10876a08153a78a06c79bc7590c30c471777", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v235/lu24e.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:46+00:00", "role": "primary", "sha256": "c644b1e3f071879aec8ea0e741e31e84193905dd730382ada54f1f24d04298ab", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.05930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.05930"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v235/lu24e.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"}], "name": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue", "publication": {"authors": ["Xing Han Lù", "Zdeněk Kasner", "Siva Reddy"], "date_basis": "Explicit publisher metadata: 2024/07/08; precision=day.", "date_parts": [2024, 7, 8], "doi": null, "event_dates": "21--27 Jul 2024", "evidence_urls": ["https://proceedings.mlr.press/v235/lu24e.html", "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The PMLR landing renders Xing Han Lu; the paper title page has Xing Han Lù. Publication authors use the PDF. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "33007-33056", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/lu24e/lu24e.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue", "type": "conference-paper", "url": "https://proceedings.mlr.press/v235/lu24e.html", "venue": "Proceedings of the 41st International Conference on Machine Learning", "volume": "235"}, "published_at": "2024-02-08", "record_sha256": "47c15df8c6537dd8bc2b26ecf2c8b6c9d7673d52da0b4caa9b57fb1fc441a37e", "relations": [], "resource_id": "arxiv:2402.05930", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.05930v2", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2024-09-10"}, "text": "WebLINX: Real-World Website Navigation with Multi-Turn Dialogue\nResource type: paper\nReview status: curated\nScope: core\n\nCollects expert demonstrations of conversational web navigation and studies retrieval-based element selection for processing large pages.\nPlatforms: web\nTopics: data-generation, web-navigation, evaluation\nPublished: 2024-02-08\nAuthors: Xing Han Lù, Zdeněk Kasner, Siva Reddy\nRecorded arXiv version: v2\n\nPublication edition: WebLINX: Real-World Website Navigation with Multi-Turn Dialogue\nPublication authors: Xing Han Lù, Zdeněk Kasner, Siva Reddy\nPublication edition date (recorded precision): 2024-07-08\nPublication venue: Proceedings of the 41st International Conference on Machine Learning\nPublication publisher: PMLR\nPublication volume: 235\nPublication pages: 33007-33056\nPublication date basis: Explicit publisher metadata: 2024/07/08; precision=day.\nPublication event dates: 21--27 Jul 2024\nPublication notes: The PMLR landing renders Xing Han Lu; the paper title page has Xing Han Lù. Publication authors use the PDF. The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://proceedings.mlr.press/v235/lu24e.html\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.05930v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-05930/"}
{"id": "arxiv:2402.06596", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0504cfffc5d46ecf447a4fdeac561aede9482bb8640fcfd9d7f02cf6eaf47597", "submitted_at": "2024-02-09T18:19:25Z", "updated_at": "2024-02-09T18:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.06596v1"}, "authors": ["Mingzhe Xing", "Rongkai Zhang", "Hui Xue", "Qi Chen", "Fan Yang", "Zhen Xiao"], "canonical_source_url": "https://arxiv.org/abs/2402.06596", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-06596/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0504cfffc5d46ecf447a4fdeac561aede9482bb8640fcfd9d7f02cf6eaf47597", "submitted_at": "2024-02-09T18:19:25Z", "updated_at": "2024-02-09T18:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.06596v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.06596"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.06596"}], "name": "Understanding the Weakness of Large Language Model Agents within a Complex Android Environment", "publication": null, "published_at": "2024-02-09", "record_sha256": "140f01e564f9fb36d18e12ad8ce2e00e9ede0a55f93514515dba4c9b2dae7cbd", "relations": [], "resource_id": "arxiv:2402.06596", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.06596v1", "topics": ["evaluation", "cross-app"], "updated_at": "2024-02-09"}, "text": "Understanding the Weakness of Large Language Model Agents within a Complex Android Environment\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AndroidArena and analyzes agent limitations in cross-application tasks, user constraints, exploration, and reflection.\nPlatforms: android\nTopics: evaluation, cross-app\nPublished: 2024-02-09\nAuthors: Mingzhe Xing, Rongkai Zhang, Hui Xue, Qi Chen, Fan Yang, Zhen Xiao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.06596v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-06596/"}
{"id": "arxiv:2402.07456", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "674b005f670a3593bd215a8ff75eac05c90ee068578fc26afdd998c40f5b7b71", "submitted_at": "2024-02-12T07:29:22Z", "updated_at": "2024-02-15T09:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.07456v2"}, "authors": ["Zhiyong Wu", "Chengcheng Han", "Zichen Ding", "Zhenmin Weng", "Zhoumianze Liu", "Shunyu Yao", "Tao Yu", "Lingpeng Kong"], "canonical_source_url": "https://arxiv.org/abs/2402.07456", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-07456/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "674b005f670a3593bd215a8ff75eac05c90ee068578fc26afdd998c40f5b7b71", "submitted_at": "2024-02-12T07:29:22Z", "updated_at": "2024-02-15T09:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2402.07456v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.07456"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07456"}], "name": "OS-Copilot: Towards Generalist Computer Agents with Self-Improvement", "publication": null, "published_at": "2024-02-12", "record_sha256": "e84dc93d1a89c0568d7a03222b04a02872a2d40203ba029dff4aaf09b162e0f3", "relations": [], "resource_id": "arxiv:2402.07456", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.07456v2", "topics": ["tool-integration", "planning"], "updated_at": "2024-02-15"}, "text": "OS-Copilot: Towards Generalist Computer Agents with Self-Improvement\nResource type: paper\nReview status: curated\nScope: core\n\nProvides an operating-system agent framework connecting applications, files, terminals, and web tools, with accumulated skills for subsequent tasks.\nPlatforms: desktop, web\nTopics: tool-integration, planning\nPublished: 2024-02-12\nAuthors: Zhiyong Wu, Chengcheng Han, Zichen Ding, Zhenmin Weng, Zhoumianze Liu, Shunyu Yao, Tao Yu, Lingpeng Kong\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.07456v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-07456/"}
{"id": "arxiv:2402.07939", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "533acb1c9aa3f1c77adb2c67779da86e3d5de95264ba3ec1f81d803f67da7753", "submitted_at": "2024-02-08T15:40:35Z", "updated_at": "2024-05-23T05:18:58Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2402.07939v5"}, "authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2402.07939", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-07939/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "533acb1c9aa3f1c77adb2c67779da86e3d5de95264ba3ec1f81d803f67da7753", "submitted_at": "2024-02-08T15:40:35Z", "updated_at": "2024-05-23T05:18:58Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2402.07939v5"}, "publication": {"authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.26", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.26/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "597-622", "pdf_url": "https://aclanthology.org/2025.naacl-long.26.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UFO: A UI-Focused Agent for Windows OS Interaction", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.26/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:48Z", "role": "primary", "sha256": "bbe15e312bcbc923811ef73e561fa7db6a7244ee3747bb9e3792e166b9327235", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.07939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07939"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.26/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.26.pdf"}], "name": "UFO: A UI-Focused Agent for Windows OS Interaction", "publication": {"authors": ["Chaoyun Zhang", "Liqun Li", "Shilin He", "Xu Zhang", "Bo Qiao", "Si Qin", "Minghua Ma", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.26", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.26/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "597-622", "pdf_url": "https://aclanthology.org/2025.naacl-long.26.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UFO: A UI-Focused Agent for Windows OS Interaction", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.26/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-02-08", "record_sha256": "142a14ea6d603d198d5a3c340ad9b509bada708fdbbcaedad404f652e17c6d17", "relations": [], "resource_id": "arxiv:2402.07939", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.07939v5", "topics": ["cross-app", "planning", "grounding", "automation"], "updated_at": "2024-05-23"}, "text": "UFO: A UI-Focused Agent for Windows OS Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nUFO uses two agents to inspect Windows application interfaces and their control metadata, then grounds and executes user requests within and across applications. The study evaluates this UI-focused architecture on nine Windows applications.\nPlatforms: windows, desktop\nTopics: cross-app, planning, grounding, automation\nPublished: 2024-02-08\nAuthors: Chaoyun Zhang, Liqun Li, Shilin He, Xu Zhang, Bo Qiao, Si Qin, Minghua Ma, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nRecorded arXiv version: v5\n\nPublication edition: UFO: A UI-Focused Agent for Windows OS Interaction\nPublication authors: Chaoyun Zhang, Liqun Li, Shilin He, Xu Zhang, Bo Qiao, Si Qin, Minghua Ma, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 597-622\nPublication doi: 10.18653/v1/2025.naacl-long.26\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.naacl-long.26/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.07939v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-07939/"}
{"id": "arxiv:2402.07945", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63424a86302625b15a91ff0252ca16c0a6448160bfd748ba32e386c82883584e", "submitted_at": "2024-02-09T02:33:45Z", "updated_at": "2024-02-09T02:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.07945v1"}, "authors": ["Runliang Niu", "Jindong Li", "Shiqi Wang", "Yali Fu", "Xiyu Hu", "Xueyuan Leng", "He Kong", "Yi Chang", "Qi Wang"], "canonical_source_url": "https://arxiv.org/abs/2402.07945", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-07945/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63424a86302625b15a91ff0252ca16c0a6448160bfd748ba32e386c82883584e", "submitted_at": "2024-02-09T02:33:45Z", "updated_at": "2024-02-09T02:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.07945v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.07945"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.07945"}], "name": "ScreenAgent: A Vision Language Model-driven Computer Control Agent", "publication": null, "published_at": "2024-02-09", "record_sha256": "875d5b23d111b850a160ad774ad457edaa9b2070b3b95c28aabe5f1cae0f1132", "relations": [], "resource_id": "arxiv:2402.07945", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.07945v1", "topics": ["automation", "planning"], "updated_at": "2024-02-09"}, "text": "ScreenAgent: A Vision Language Model-driven Computer Control Agent\nResource type: paper\nReview status: curated\nScope: core\n\nCombines a screenshot-and-action dataset with a planning, acting, and reflecting loop for computer control through mouse and keyboard actions.\nPlatforms: desktop\nTopics: automation, planning\nPublished: 2024-02-09\nAuthors: Runliang Niu, Jindong Li, Shiqi Wang, Yali Fu, Xiyu Hu, Xueyuan Leng, He Kong, Yi Chang, Qi Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.07945v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-07945/"}
{"id": "arxiv:2402.10196", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6d81a6b9de6fa32c8f5749d029d0819ddf3b310ae00a4872877ee72580e8879", "submitted_at": "2024-02-15T18:51:32Z", "updated_at": "2024-02-15T18:51:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.10196v1"}, "authors": ["Lingbo Mo", "Zeyi Liao", "Boyuan Zheng", "Yu Su", "Chaowei Xiao", "Huan Sun"], "canonical_source_url": "https://arxiv.org/abs/2402.10196", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-10196/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6d81a6b9de6fa32c8f5749d029d0819ddf3b310ae00a4872877ee72580e8879", "submitted_at": "2024-02-15T18:51:32Z", "updated_at": "2024-02-15T18:51:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.10196v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:03.117157+00:00", "role": "primary", "sha256": "a1d4dc44a6572863df1d8cc68bff592c3a04a174a3dbc87ab5cdcd0d33be92b0", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2402.10196v1"}], "identifiers": {"arxiv": "2402.10196"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.10196"}], "name": "A Trembling House of Cards? Mapping Adversarial Attacks against Language Agents", "publication": null, "published_at": "2024-02-15", "record_sha256": "d6301fcbab362a0765d00a3d7358492a3a4d15d2fe081ac117c21af4e94e1daa", "relations": [], "resource_id": "arxiv:2402.10196", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2402.10196v1", "topics": ["safety-security", "survey", "robustness"], "updated_at": "2024-02-15"}, "text": "A Trembling House of Cards? Mapping Adversarial Attacks against Language Agents\nResource type: survey\nReview status: curated\nScope: adjacent\n\nPresents a conceptual perception–reasoning–action framework and 12 potential attack scenarios for language agents, including hypothetical web-navigation attacks.\nPlatforms: web\nTopics: safety-security, survey, robustness\nPublished: 2024-02-15\nAuthors: Lingbo Mo, Zeyi Liao, Boyuan Zheng, Yu Su, Chaowei Xiao, Huan Sun\nRecorded arXiv version: v1\n\nAdjacent conceptual security work: the paper describes hypothetical scenarios and a future generalist agent, rather than an implemented attack benchmark. Scope was checked against the retained abstract and full-text Sections 2.4 and 3; no results were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.10196v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-10196/"}
{"id": "arxiv:2402.11941", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "98930b10832d8169bdc60620f882b4b8f753440979c8e9a8de167b12d38d7060", "submitted_at": "2024-02-19T08:29:03Z", "updated_at": "2024-06-02T13:25:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.11941v3"}, "authors": ["Xinbei Ma", "Zhuosheng Zhang", "Hai Zhao"], "canonical_source_url": "https://arxiv.org/abs/2402.11941", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-11941/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "98930b10832d8169bdc60620f882b4b8f753440979c8e9a8de167b12d38d7060", "submitted_at": "2024-02-19T08:29:03Z", "updated_at": "2024-06-02T13:25:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.11941v3"}, "publication": {"authors": ["Xinbei Ma", "Zhuosheng Zhang", "Hai Zhao (赵海)"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.539", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.539"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "9097-9110", "pdf_url": "https://aclanthology.org/2024.findings-acl.539.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.539", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:49Z", "role": "primary", "sha256": "022e2ca49699ec713f11eec442b2139fc0ce4d075f0812f88844289749f13d6c", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-acl.539"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.11941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.11941"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-acl.539"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-acl.539.pdf"}], "name": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation", "publication": {"authors": ["Xinbei Ma", "Zhuosheng Zhang", "Hai Zhao (赵海)"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.findings-acl.539", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-acl.539"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "9097-9110", "pdf_url": "https://aclanthology.org/2024.findings-acl.539.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-acl.539", "venue": "Findings of the Association for Computational Linguistics: ACL 2024", "volume": null}, "published_at": "2024-02-19", "record_sha256": "5461999d0922b24499dec63d11872e9392207fbda2f1b9ad032807bfc15b0b27", "relations": [], "resource_id": "arxiv:2402.11941", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.11941v3", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": "2024-06-02"}, "text": "CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nCombines screenshots, layout information, and action history with a factored action predictor for smartphone automation.\nPlatforms: android\nTopics: mobile-navigation, screen-understanding\nPublished: 2024-02-19\nAuthors: Xinbei Ma, Zhuosheng Zhang, Hai Zhao\nRecorded arXiv version: v3\n\nPublication edition: CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation\nPublication authors: Xinbei Ma, Zhuosheng Zhang, Hai Zhao (赵海)\nPublication edition date (recorded precision): 2024-08\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 9097-9110\nPublication doi: 10.18653/v1/2024.findings-acl.539\nPublication date basis: Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2024.findings-acl.539\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.11941v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-11941/"}
{"id": "arxiv:2402.14073", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f3eed6e0021be438af9d056e97f71a838af905b948cfd6499423ab18df11e43", "submitted_at": "2024-02-21T19:01:03Z", "updated_at": "2024-02-21T19:01:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.14073v1"}, "authors": ["Tianyu Gao", "Zirui Wang", "Adithya Bhaskar", "Danqi Chen"], "canonical_source_url": "https://arxiv.org/abs/2402.14073", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-14073/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f3eed6e0021be438af9d056e97f71a838af905b948cfd6499423ab18df11e43", "submitted_at": "2024-02-21T19:01:03Z", "updated_at": "2024-02-21T19:01:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.14073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.14073"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.14073"}], "name": "Improving Language Understanding from Screenshots", "publication": null, "published_at": "2024-02-21", "record_sha256": "b92fcb75da1e903ed5b21e7c6a0d960c15708a6015afd5868fb49bfbbc87bd8b", "relations": [], "resource_id": "arxiv:2402.14073", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2402.14073v1", "topics": ["screen-understanding"], "updated_at": "2024-02-21"}, "text": "Improving Language Understanding from Screenshots\nResource type: paper\nReview status: curated\nScope: adjacent\n\nStudies language understanding from text-rendered screenshots using joint image-patch and text prediction; a perception foundation relevant to screen-reading models.\nPlatforms: unspecified\nTopics: screen-understanding\nPublished: 2024-02-21\nAuthors: Tianyu Gao, Zirui Wang, Adithya Bhaskar, Danqi Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.14073v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-14073/"}
{"id": "arxiv:2402.15057", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e4b233e423e9043b6d615c7b185a1eb0d47f646590e4fb9fd43ca156a1a1b16", "submitted_at": "2024-02-23T02:18:12Z", "updated_at": "2024-02-23T02:18:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.15057v1"}, "authors": ["Yang Deng", "Xuan Zhang", "Wenxuan Zhang", "Yifei Yuan", "See-Kiong Ng", "Tat-Seng Chua"], "canonical_source_url": "https://arxiv.org/abs/2402.15057", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-15057/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e4b233e423e9043b6d615c7b185a1eb0d47f646590e4fb9fd43ca156a1a1b16", "submitted_at": "2024-02-23T02:18:12Z", "updated_at": "2024-02-23T02:18:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.15057v1"}, "publication": {"authors": ["Yang Deng", "Xuan Zhang", "Wenxuan Zhang", "Yifei Yuan", "See Kiong Ng", "Tat-Seng Chua"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.477", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.477/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8795-8812", "pdf_url": "https://aclanthology.org/2024.acl-long.477.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "On the Multi-turn Instruction Following for Conversational Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.477/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:49Z", "role": "primary", "sha256": "ba56d6d0a422b340672036d5c872cd00d54a0f12cfe1326a3be65e8702786381", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.acl-long.477/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2402.15057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.15057"}, {"rel": "publication", "url": "https://aclanthology.org/2024.acl-long.477/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.acl-long.477.pdf"}], "name": "On the Multi-turn Instruction Following for Conversational Web Agents", "publication": {"authors": ["Yang Deng", "Xuan Zhang", "Wenxuan Zhang", "Yifei Yuan", "See Kiong Ng", "Tat-Seng Chua"], "date_basis": "Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 8], "doi": "10.18653/v1/2024.acl-long.477", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.acl-long.477/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8795-8812", "pdf_url": "https://aclanthology.org/2024.acl-long.477.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "On the Multi-turn Instruction Following for Conversational Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.acl-long.477/", "venue": "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-02-23", "record_sha256": "36b77d2d4642e029852d18ee42f7cafff4e380b026ceabdf26571e704b2d7c81", "relations": [], "resource_id": "arxiv:2402.15057", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.15057v1", "topics": ["evaluation", "memory", "self-reflection", "web-navigation"], "updated_at": "2024-02-23"}, "text": "On the Multi-turn Instruction Following for Conversational Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMT-Mind2Web evaluates web navigation through sequences of conversational user instructions. Self-MAP combines memory-augmented planning with reflection to preserve context and dependencies across those interaction turns.\nPlatforms: web\nTopics: evaluation, memory, self-reflection, web-navigation\nPublished: 2024-02-23\nAuthors: Yang Deng, Xuan Zhang, Wenxuan Zhang, Yifei Yuan, See-Kiong Ng, Tat-Seng Chua\nRecorded arXiv version: v1\n\nPublication edition: On the Multi-turn Instruction Following for Conversational Web Agents\nPublication authors: Yang Deng, Xuan Zhang, Wenxuan Zhang, Yifei Yuan, See Kiong Ng, Tat-Seng Chua\nPublication edition date (recorded precision): 2024-08\nPublication venue: Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 8795-8812\nPublication doi: 10.18653/v1/2024.acl-long.477\nPublication date basis: Explicit publisher citation_publication_date: 2024/8. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2024.acl-long.477/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.15057v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-15057/"}
{"id": "arxiv:2402.16965", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cf1458c8c4134b2195394a8acba4919f9b3a5a7c79b1bea017ac4558123f69c", "submitted_at": "2024-02-26T19:01:54Z", "updated_at": "2024-02-26T19:01:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.16965v1"}, "authors": ["Fangzhou Wu", "Shutong Wu", "Yulong Cao", "Chaowei Xiao"], "canonical_source_url": "https://arxiv.org/abs/2402.16965", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-16965/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cf1458c8c4134b2195394a8acba4919f9b3a5a7c79b1bea017ac4558123f69c", "submitted_at": "2024-02-26T19:01:54Z", "updated_at": "2024-02-26T19:01:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2402.16965v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2402.16965"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.16965"}], "name": "WIPI: A New Web Threat for LLM-Driven Web Agents", "publication": null, "published_at": "2024-02-26", "record_sha256": "0b52a04cd80ddb3e64e1d53149134d43ab05564806098f59ca65d4149ff84852", "relations": [], "resource_id": "arxiv:2402.16965", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.16965v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2024-02-26"}, "text": "WIPI: A New Web Threat for LLM-Driven Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies WIPI, a black-box indirect-instruction threat in which webpage content redirects LLM-driven web agents. Evaluation spans plugin-based agents, Web GPTs, and open-source web-agent implementations.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2024-02-26\nAuthors: Fangzhou Wu, Shutong Wu, Yulong Cao, Chaowei Xiao\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.16965v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-16965/"}
{"id": "arxiv:2402.17553", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b548ac2bee529d3a769e7ea14b6bfce2270cdf349d832cc8c146e7d779b81a", "submitted_at": "2024-02-27T14:47:53Z", "updated_at": "2024-07-21T23:16:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.17553v3"}, "authors": ["Raghav Kapoor", "Yash Parag Butala", "Melisa Russak", "Jing Yu Koh", "Kiran Kamble", "Waseem Alshikh", "Ruslan Salakhutdinov"], "canonical_source_url": "https://arxiv.org/abs/2402.17553", "citation_url": "https://awesomecua.github.io/resources/arxiv-2402-17553/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b548ac2bee529d3a769e7ea14b6bfce2270cdf349d832cc8c146e7d779b81a", "submitted_at": "2024-02-27T14:47:53Z", "updated_at": "2024-07-21T23:16:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2402.17553v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}], "identifiers": {"arxiv": "2402.17553"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2402.17553"}], "name": "OmniACT: A Dataset and Benchmark for Enabling Multimodal Generalist Autonomous Agents for Desktop and Web", "publication": null, "published_at": "2024-02-27", "record_sha256": "86615372e856e51871ba89ec7bf4bdc8f333688895a16c25df03549f5915f185", "relations": [], "resource_id": "arxiv:2402.17553", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2402.17553v3", "topics": ["evaluation", "grounding"], "updated_at": "2024-07-21"}, "text": "OmniACT: A Dataset and Benchmark for Enabling Multimodal Generalist Autonomous Agents for Desktop and Web\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a dataset and benchmark for generating executable task scripts from screenshots and visually grounded desktop or web instructions.\nPlatforms: desktop, web\nTopics: evaluation, grounding\nPublished: 2024-02-27\nAuthors: Raghav Kapoor, Yash Parag Butala, Melisa Russak, Jing Yu Koh, Kiran Kamble, Waseem Alshikh, Ruslan Salakhutdinov\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2402.17553v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2402-17553/"}
{"id": "arxiv:2403.02713", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4de9e7d68aa8450cab65ecb300524e85b1d2b2a91d58fecef370deceaba64b1e", "submitted_at": "2024-03-05T07:09:35Z", "updated_at": "2024-07-13T02:12:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.02713v2"}, "authors": ["Jiwen Zhang", "Jihao Wu", "Yihua Teng", "Minghui Liao", "Nuo Xu", "Xiao Xiao", "Zhongyu Wei", "Duyu Tang"], "canonical_source_url": "https://arxiv.org/abs/2403.02713", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-02713/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4de9e7d68aa8450cab65ecb300524e85b1d2b2a91d58fecef370deceaba64b1e", "submitted_at": "2024-03-05T07:09:35Z", "updated_at": "2024-07-13T02:12:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.02713v2"}, "publication": {"authors": ["Jiwen Zhang (张霁雯)", "Jihao Wu", "Teng Yihua", "Minghui Liao", "Nuo Xu", "Xiao Xiao", "Zhongyu Wei (魏忠钰)", "Duyu Tang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.702", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.702/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "12016-12031", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.702.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.702/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:50Z", "role": "primary", "sha256": "87a46c27e83152e7741c2398f790320f937b8fd770fa7ac7d3a68395427e87d2", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.702/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2403.02713"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.02713"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.702/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.702.pdf"}], "name": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents", "publication": {"authors": ["Jiwen Zhang (张霁雯)", "Jihao Wu", "Teng Yihua", "Minghui Liao", "Nuo Xu", "Xiao Xiao", "Zhongyu Wei (魏忠钰)", "Duyu Tang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.702", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.702/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "12016-12031", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.702.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.702/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}, "published_at": "2024-03-05", "record_sha256": "3eb453772aaad5aeae81d46263a4b86ceefcf009731fb0a60e6bd5dab7b8b008", "relations": [], "resource_id": "arxiv:2403.02713", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.02713v2", "topics": ["mobile-navigation", "planning"], "updated_at": "2024-07-13"}, "text": "Android in the Zoo: Chain-of-Action-Thought for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Chain-of-Action-Thought and Android-In-The-Zoo annotations that connect prior actions, screen observations, intended actions, and expected outcomes.\nPlatforms: android\nTopics: mobile-navigation, planning\nPublished: 2024-03-05\nAuthors: Jiwen Zhang, Jihao Wu, Yihua Teng, Minghui Liao, Nuo Xu, Xiao Xiao, Zhongyu Wei, Duyu Tang\nRecorded arXiv version: v2\n\nPublication edition: Android in the Zoo: Chain-of-Action-Thought for GUI Agents\nPublication authors: Jiwen Zhang (张霁雯), Jihao Wu, Teng Yihua, Minghui Liao, Nuo Xu, Xiao Xiao, Zhongyu Wei (魏忠钰), Duyu Tang\nPublication edition date (recorded precision): 2024-11\nPublication venue: Findings of the Association for Computational Linguistics: EMNLP 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 12016-12031\nPublication doi: 10.18653/v1/2024.findings-emnlp.702\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2024.findings-emnlp.702/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.02713v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-02713/"}
{"id": "arxiv:2403.03186", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63c68c596f1c81485e77f98beae696deca9785a5d8725ab4456ee35870e64a99", "submitted_at": "2024-03-05T18:22:29Z", "updated_at": "2024-07-02T17:23:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.03186v3"}, "authors": ["Weihao Tan", "Wentao Zhang", "Xinrun Xu", "Haochong Xia", "Ziluo Ding", "Boyu Li", "Bohan Zhou", "Junpeng Yue", "Jiechuan Jiang", "Yewen Li", "Ruyi An", "Molei Qin", "Chuqiao Zong", "Longtao Zheng", "Yujie Wu", "Xiaoqiang Chai", "Yifei Bi", "Tianbao Xie", "Pengjie Gu", "Xiyun Li", "Ceyao Zhang", "Long Tian", "Chaojie Wang", "Xinrun Wang", "Börje F. Karlsson", "Bo An", "Shuicheng Yan", "Zongqing Lu"], "canonical_source_url": "https://arxiv.org/abs/2403.03186", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-03186/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63c68c596f1c81485e77f98beae696deca9785a5d8725ab4456ee35870e64a99", "submitted_at": "2024-03-05T18:22:29Z", "updated_at": "2024-07-02T17:23:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.03186v3"}, "publication": {"authors": ["Weihao Tan", "Wentao Zhang", "Xinrun Xu", "Haochong Xia", "Ziluo Ding", "Boyu Li", "Bohan Zhou", "Junpeng Yue", "Jiechuan Jiang", "Yewen Li", "Ruyi An", "Molei Qin", "Chuqiao Zong", "Longtao Zheng", "Yujie Wu", "Xiaoqiang Chai", "Yifei Bi", "Tianbao Xie", "Pengjie Gu", "Xiyun Li", "Ceyao Zhang", "Long Tian", "Chaojie Wang", "Xinrun Wang", "Börje F. Karlsson", "Bo An", "Shuicheng Yan", "Zongqing Lu"], "date_basis": "Explicit publisher metadata: 2025/10/06; precision=day.", "date_parts": [2025, 10, 6], "doi": null, "event_dates": "13--19 Jul 2025", "evidence_urls": ["https://proceedings.mlr.press/v267/tan25h.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "58658-58725", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/tan25h/tan25h.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "Cradle: Empowering Foundation Agents towards General Computer Control", "type": "conference-paper", "url": "https://proceedings.mlr.press/v267/tan25h.html", "venue": "Proceedings of the 42nd International Conference on Machine Learning", "volume": "267"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:47+00:00", "role": "primary", "sha256": "c081bdd38f20d6c3aac756da474b8cb8a6d782ece213f67ebc333bf2b99de2fc", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v267/tan25h.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2403.03186"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.03186"}, {"rel": "publication", "url": "https://proceedings.mlr.press/v267/tan25h.html"}, {"rel": "publication_pdf", "url": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/tan25h/tan25h.pdf"}], "name": "Cradle: Empowering Foundation Agents Towards General Computer Control", "publication": {"authors": ["Weihao Tan", "Wentao Zhang", "Xinrun Xu", "Haochong Xia", "Ziluo Ding", "Boyu Li", "Bohan Zhou", "Junpeng Yue", "Jiechuan Jiang", "Yewen Li", "Ruyi An", "Molei Qin", "Chuqiao Zong", "Longtao Zheng", "Yujie Wu", "Xiaoqiang Chai", "Yifei Bi", "Tianbao Xie", "Pengjie Gu", "Xiyun Li", "Ceyao Zhang", "Long Tian", "Chaojie Wang", "Xinrun Wang", "Börje F. Karlsson", "Bo An", "Shuicheng Yan", "Zongqing Lu"], "date_basis": "Explicit publisher metadata: 2025/10/06; precision=day.", "date_parts": [2025, 10, 6], "doi": null, "event_dates": "13--19 Jul 2025", "evidence_urls": ["https://proceedings.mlr.press/v267/tan25h.html"], "identity_basis": "Same work: exact normalized landing title, corresponding author list, and matching specific contribution. Publisher and arXiv versions remain distinct.", "issue": null, "notes": "The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.", "pages": "58658-58725", "pdf_url": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/tan25h/tan25h.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "Cradle: Empowering Foundation Agents towards General Computer Control", "type": "conference-paper", "url": "https://proceedings.mlr.press/v267/tan25h.html", "venue": "Proceedings of the 42nd International Conference on Machine Learning", "volume": "267"}, "published_at": "2024-03-05", "record_sha256": "c4135abf01a07b4ec0644ff2cd92fd41dcce326d2cfb76a487059d1c5bdee0ee", "relations": [], "resource_id": "arxiv:2403.03186", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.03186v3", "topics": ["planning", "screen-understanding"], "updated_at": "2024-07-02"}, "text": "Cradle: Empowering Foundation Agents Towards General Computer Control\nResource type: paper\nReview status: curated\nScope: core\n\nUses screenshot observations, planning modules, and generated low-level keyboard and mouse control for general computer interaction.\nPlatforms: desktop\nTopics: planning, screen-understanding\nPublished: 2024-03-05\nAuthors: Weihao Tan, Wentao Zhang, Xinrun Xu, Haochong Xia, Ziluo Ding, Boyu Li, Bohan Zhou, Junpeng Yue, Jiechuan Jiang, Yewen Li, Ruyi An, Molei Qin, Chuqiao Zong, Longtao Zheng, Yujie Wu, Xiaoqiang Chai, Yifei Bi, Tianbao Xie, Pengjie Gu, Xiyun Li, Ceyao Zhang, Long Tian, Chaojie Wang, Xinrun Wang, Börje F. Karlsson, Bo An, Shuicheng Yan, Zongqing Lu\nRecorded arXiv version: v3\n\nPublication edition: Cradle: Empowering Foundation Agents towards General Computer Control\nPublication authors: Weihao Tan, Wentao Zhang, Xinrun Xu, Haochong Xia, Ziluo Ding, Boyu Li, Bohan Zhou, Junpeng Yue, Jiechuan Jiang, Yewen Li, Ruyi An, Molei Qin, Chuqiao Zong, Longtao Zheng, Yujie Wu, Xiaoqiang Chai, Yifei Bi, Tianbao Xie, Pengjie Gu, Xiyun Li, Ceyao Zhang, Long Tian, Chaojie Wang, Xinrun Wang, Börje F. Karlsson, Bo An, Shuicheng Yan, Zongqing Lu\nPublication edition date (recorded precision): 2025-10-06\nPublication venue: Proceedings of the 42nd International Conference on Machine Learning\nPublication publisher: PMLR\nPublication volume: 267\nPublication pages: 58658-58725\nPublication date basis: Explicit publisher metadata: 2025/10/06; precision=day.\nPublication event dates: 13--19 Jul 2025\nPublication notes: The publisher issue date comes from citation_publication_date; the conference date range is retained separately and does not replace it. Publisher-hosted proceedings record; distinct from the retained arXiv preprint snapshot.\nPublication source: https://proceedings.mlr.press/v267/tan25h.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.03186v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-03186/"}
{"id": "arxiv:2403.03788", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51abf6b165b84c647c9a4870a6441900cac5e9a933199fe8a47d1b7f065eb11e", "submitted_at": "2024-03-06T15:33:32Z", "updated_at": "2024-03-06T15:33:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.03788v1"}, "authors": ["Zekai Zhang", "Yiduo Guo", "Yaobo Liang", "Dongyan Zhao", "Nan Duan"], "canonical_source_url": "https://arxiv.org/abs/2403.03788", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-03788/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51abf6b165b84c647c9a4870a6441900cac5e9a933199fe8a47d1b7f065eb11e", "submitted_at": "2024-03-06T15:33:32Z", "updated_at": "2024-03-06T15:33:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.03788v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2403.03788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.03788"}], "name": "PPTC-R benchmark: Towards Evaluating the Robustness of Large Language Models for PowerPoint Task Completion", "publication": null, "published_at": "2024-03-06", "record_sha256": "faf54e92547372c937f425ca4f7f6ed084a97c3aaec0c03f3961f4d6899d3a03", "relations": [], "resource_id": "arxiv:2403.03788", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2403.03788v1", "topics": ["evaluation", "robustness", "workflow"], "updated_at": "2024-03-06"}, "text": "PPTC-R benchmark: Towards Evaluating the Robustness of Large Language Models for PowerPoint Task Completion\nResource type: paper\nReview status: curated\nScope: adjacent\n\nBenchmarks robustness of API-based PowerPoint completion under sentence-level, semantic, and multilingual instruction perturbations and changes to the available API set.\nPlatforms: unspecified\nTopics: evaluation, robustness, workflow\nPublished: 2024-03-06\nAuthors: Zekai Zhang, Yiduo Guo, Yaobo Liang, Dongyan Zhao, Nan Duan\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The benchmark directly concerns PowerPoint task automation but executes through APIs, so it remains adjacent.\n\nPrimary source: https://arxiv.org/abs/2403.03788v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-03788/"}
{"id": "arxiv:2403.04984", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4bfed18cba0c946a8ae013c819d84d7f7b2c639b2e3471d6bf7890248c4530b", "submitted_at": "2024-03-08T01:52:44Z", "updated_at": "2024-03-08T01:52:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.04984v1"}, "authors": ["Shuhong Xiao", "Yunnong Chen", "Yaxuan Song", "Liuqing Chen", "Lingyun Sun", "Yankun Zhen", "Yanfang Chang"], "canonical_source_url": "https://arxiv.org/abs/2403.04984", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-04984/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4bfed18cba0c946a8ae013c819d84d7f7b2c639b2e3471d6bf7890248c4530b", "submitted_at": "2024-03-08T01:52:44Z", "updated_at": "2024-03-08T01:52:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2403.04984v1"}}, "environments": ["android", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2403.04984"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.04984"}], "name": "UI Semantic Group Detection: Grouping UI Elements with Similar Semantics in Mobile Graphical User Interface", "publication": null, "published_at": "2024-03-08", "record_sha256": "df6d7a3ccde3955ea50fc7e5585774a1796279da00cf619c851d01023ca778df", "relations": [], "resource_id": "arxiv:2403.04984", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.04984v1", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-03-08"}, "text": "UI Semantic Group Detection: Grouping UI Elements with Similar Semantics in Mobile Graphical User Interface\nResource type: paper\nReview status: curated\nScope: core\n\nUISCGD detects semantic groups of adjacent text and non-text UI elements. It learns from Android and iOS screenshots to support interface retrieval, UI-to-code structure and screen-reader accessibility.\nPlatforms: android, ios\nTopics: screen-understanding, grounding\nPublished: 2024-03-08\nAuthors: Shuhong Xiao, Yunnong Chen, Yaxuan Song, Liuqing Chen, Lingyun Sun, Yankun Zhen, Yanfang Chang\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.04984v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-04984/"}
{"id": "arxiv:2403.07718", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8033f9c26a5f95a49000374a6edebb4dead83e100ebcf5226007f06b7573c305", "submitted_at": "2024-03-12T14:58:45Z", "updated_at": "2024-07-23T06:19:28Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2403.07718v5"}, "authors": ["Alexandre Drouin", "Maxime Gasse", "Massimo Caccia", "Issam H. Laradji", "Manuel Del Verme", "Tom Marty", "Léo Boisvert", "Megh Thakkar", "Quentin Cappart", "David Vazquez", "Nicolas Chapados", "Alexandre Lacoste"], "canonical_source_url": "https://arxiv.org/abs/2403.07718", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-07718/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8033f9c26a5f95a49000374a6edebb4dead83e100ebcf5226007f06b7573c305", "submitted_at": "2024-03-12T14:58:45Z", "updated_at": "2024-07-23T06:19:28Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2403.07718v5"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2403.07718"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.07718"}], "name": "WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?", "publication": null, "published_at": "2024-03-12", "record_sha256": "a9ef9ee103623f137ca698722e7eb28dd2c4ecd79a6596249a235cd1c9cabd69", "relations": [], "resource_id": "arxiv:2403.07718", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.07718v5", "topics": ["evaluation", "workflow"], "updated_at": "2024-07-23"}, "text": "WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces enterprise web tasks on ServiceNow and BrowserGym interfaces for evaluating browser-based agents.\nPlatforms: web\nTopics: evaluation, workflow\nPublished: 2024-03-12\nAuthors: Alexandre Drouin, Maxime Gasse, Massimo Caccia, Issam H. Laradji, Manuel Del Verme, Tom Marty, Léo Boisvert, Megh Thakkar, Quentin Cappart, David Vazquez, Nicolas Chapados, Alexandre Lacoste\nRecorded arXiv version: v5\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.07718v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-07718/"}
{"id": "arxiv:2403.10170", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aacb4d2cac31e478aa56ee25b6b39e5ba8fe164f89102e665f8fb25cf3822e7b", "submitted_at": "2024-03-15T10:26:52Z", "updated_at": "2024-08-27T18:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.10170v2"}, "authors": ["Andrés Muñoz", "Daniel Borrajo"], "canonical_source_url": "https://arxiv.org/abs/2403.10170", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-10170/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aacb4d2cac31e478aa56ee25b6b39e5ba8fe164f89102e665f8fb25cf3822e7b", "submitted_at": "2024-03-15T10:26:52Z", "updated_at": "2024-08-27T18:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.10170v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2403.10170"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.10170"}], "name": "Computer User Interface Understanding. A New Dataset and a Learning Framework", "publication": null, "published_at": "2024-03-15", "record_sha256": "b3398bc7607b29bc3596943597f728a2180cdf7760b56c8ffbaf9872e5a29857", "relations": [], "resource_id": "arxiv:2403.10170", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.10170v2", "topics": ["screen-understanding", "video", "data-generation"], "updated_at": "2024-08-27"}, "text": "Computer User Interface Understanding. A New Dataset and a Learning Framework\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces desktop interaction videos and a framework combining synthetic examples with hierarchical contrastive learning for interface classification.\nPlatforms: desktop\nTopics: screen-understanding, video, data-generation\nPublished: 2024-03-15\nAuthors: Andrés Muñoz, Daniel Borrajo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.10170v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-10170/"}
{"id": "arxiv:2403.11905", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d82424b84ebddfeb005ffae38c130c6a9c7dae7adcdc59e2b7d001284b785573", "submitted_at": "2024-03-18T16:06:30Z", "updated_at": "2025-02-22T01:33:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2403.11905v4"}, "authors": ["Kevin Xu", "Yeganeh Kordi", "Tanay Nayak", "Adi Asija", "Yizhong Wang", "Kate Sanders", "Adam Byerly", "Jingyu Zhang", "Benjamin Van Durme", "Daniel Khashabi"], "canonical_source_url": "https://arxiv.org/abs/2403.11905", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-11905/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d82424b84ebddfeb005ffae38c130c6a9c7dae7adcdc59e2b7d001284b785573", "submitted_at": "2024-03-18T16:06:30Z", "updated_at": "2025-02-22T01:33:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2403.11905v4"}, "publication": {"authors": ["Kevin Xu", "Yeganeh Kordi", "Tanay Nayak", "Adi Asija", "Yizhong Wang", "Kate Sanders", "Adam Byerly", "Jingyu Zhang", "Benjamin Van Durme", "Daniel Khashabi"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.188", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.188/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3694-3710", "pdf_url": "https://aclanthology.org/2025.naacl-long.188.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "TurkingBench: A Challenge Benchmark for Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.188/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:50Z", "role": "primary", "sha256": "56537fe44b4655949e8c70cc49a4ed7304a8201a12f75ddb3caa5df92d6f8e16", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.188/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2403.11905"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.11905"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.188/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.188.pdf"}], "name": "Tur[k]ingBench: A Challenge Benchmark for Web Agents", "publication": {"authors": ["Kevin Xu", "Yeganeh Kordi", "Tanay Nayak", "Adi Asija", "Yizhong Wang", "Kate Sanders", "Adam Byerly", "Jingyu Zhang", "Benjamin Van Durme", "Daniel Khashabi"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.188", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.188/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "3694-3710", "pdf_url": "https://aclanthology.org/2025.naacl-long.188.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "TurkingBench: A Challenge Benchmark for Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.188/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-03-18", "record_sha256": "06b1e6d8c02ae5d5c966f1ed196f8ae793256f8bb848ce425a598143c312278e", "relations": [], "resource_id": "arxiv:2403.11905", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.11905v4", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-02-22"}, "text": "Tur[k]ingBench: A Challenge Benchmark for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates web agents on multimodal annotation tasks adapted from HTML interfaces originally built for crowdsourcing workers.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2024-03-18\nAuthors: Kevin Xu, Yeganeh Kordi, Tanay Nayak, Adi Asija, Yizhong Wang, Kate Sanders, Adam Byerly, Jingyu Zhang, Benjamin Van Durme, Daniel Khashabi\nRecorded arXiv version: v4\n\nPublication edition: TurkingBench: A Challenge Benchmark for Web Agents\nPublication authors: Kevin Xu, Yeganeh Kordi, Tanay Nayak, Adi Asija, Yizhong Wang, Kate Sanders, Adam Byerly, Jingyu Zhang, Benjamin Van Durme, Daniel Khashabi\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 3694-3710\nPublication doi: 10.18653/v1/2025.naacl-long.188\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.naacl-long.188/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.11905v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-11905/"}
{"id": "arxiv:2403.17918", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe1432b46c15566182418bf3fa2f918607e8b0e9a383459416d03e28dc7e347c", "submitted_at": "2024-03-26T17:54:15Z", "updated_at": "2025-02-14T08:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.17918v3"}, "authors": ["Longtao Zheng", "Zhiyuan Huang", "Zhenghai Xue", "Xinrun Wang", "Bo An", "Shuicheng Yan"], "canonical_source_url": "https://arxiv.org/abs/2403.17918", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-17918/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe1432b46c15566182418bf3fa2f918607e8b0e9a383459416d03e28dc7e347c", "submitted_at": "2024-03-26T17:54:15Z", "updated_at": "2025-02-14T08:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2403.17918v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2403.17918"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.17918"}], "name": "AgentStudio: A Toolkit for Building General Virtual Agents", "publication": null, "published_at": "2024-03-26", "record_sha256": "3b7b30ed77cf6ee617443c532b85d24f9da3c69a67ea14edb1816a141e1ce2a0", "relations": [], "resource_id": "arxiv:2403.17918", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2403.17918v3", "topics": ["evaluation", "sandbox", "grounding", "video"], "updated_at": "2025-02-14"}, "text": "AgentStudio: A Toolkit for Building General Virtual Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAgentStudio combines an interactive GUI/API environment with tools for creating tasks, annotating interface elements, and labeling actions in videos. Its online suite and GroundUI, IDMBench, and CriticBench datasets examine interaction, grounding, learning from video, and success detection.\nPlatforms: unspecified\nTopics: evaluation, sandbox, grounding, video\nPublished: 2024-03-26\nAuthors: Longtao Zheng, Zhiyuan Huang, Zhenghai Xue, Xinrun Wang, Bo An, Shuicheng Yan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2403.17918v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-17918/"}
{"id": "arxiv:2403.20329", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3573a5cf66737178f164d9809b3e6dcadb7c6b2b4ae7d95283e8a136bf78851", "submitted_at": "2024-03-29T17:59:06Z", "updated_at": "2024-08-19T03:06:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.20329v2"}, "authors": ["Joel Ruben Antony Moniz", "Soundarya Krishnan", "Melis Ozyildirim", "Prathamesh Saraf", "Halim Cagri Ates", "Yuan Zhang", "Hong Yu"], "canonical_source_url": "https://arxiv.org/abs/2403.20329", "citation_url": "https://awesomecua.github.io/resources/arxiv-2403-20329/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3573a5cf66737178f164d9809b3e6dcadb7c6b2b4ae7d95283e8a136bf78851", "submitted_at": "2024-03-29T17:59:06Z", "updated_at": "2024-08-19T03:06:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2403.20329v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2403.20329"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2403.20329"}], "name": "ReALM: Reference Resolution As Language Modeling", "publication": null, "published_at": "2024-03-29", "record_sha256": "2ce513de73fd7a80beade6e6b976a36919bb7d3fbb01aeb952ed0863f0385cae", "relations": [], "resource_id": "arxiv:2403.20329", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2403.20329v2", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-08-19"}, "text": "ReALM: Reference Resolution As Language Modeling\nResource type: paper\nReview status: curated\nScope: adjacent\n\nFrames conversational and on-screen reference resolution as language modeling, allowing textual representations of screen entities to support user-reference interpretation. Evaluation separately examines on-screen references against an existing resolver and language-model baselines.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\nPublished: 2024-03-29\nAuthors: Joel Ruben Antony Moniz, Soundarya Krishnan, Melis Ozyildirim, Prathamesh Saraf, Halim Cagri Ates, Yuan Zhang, Hong Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is reference resolution involving screen entities, not an evaluated end-to-end GUI controller.\n\nPrimary source: https://arxiv.org/abs/2403.20329v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2403-20329/"}
{"id": "arxiv:2404.02475", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b2f357e166bddc2bbb5db4ec102721fc12fd27ba45d953193be0354fa5c7bce", "submitted_at": "2024-04-03T05:32:05Z", "updated_at": "2025-03-18T09:59:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.02475v2"}, "authors": ["Tian Huang", "Chun Yu", "Weinan Shi", "Zijian Peng", "David Yang", "Weiqi Sun", "Yuanchun Shi"], "canonical_source_url": "https://arxiv.org/abs/2404.02475", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-02475/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b2f357e166bddc2bbb5db4ec102721fc12fd27ba45d953193be0354fa5c7bce", "submitted_at": "2024-04-03T05:32:05Z", "updated_at": "2025-03-18T09:59:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.02475v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.02475"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.02475"}], "name": "Prompt2Task: Automating UI Tasks on Smartphones from Textual Prompts", "publication": null, "published_at": "2024-04-03", "record_sha256": "335815dfc9ee3c171ea8a4079895266d7cad29fd877518badd20f0b3a0c36227", "relations": [], "resource_id": "arxiv:2404.02475", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.02475v2", "topics": ["mobile-navigation", "human-in-the-loop", "workflow"], "updated_at": "2025-03-18"}, "text": "Prompt2Task: Automating UI Tasks on Smartphones from Textual Prompts\nResource type: paper\nReview status: curated\nScope: core\n\nPrompt2Task converts textual goals or procedures into smartphone UI automation using agents for intent interpretation, external-information handling, and execution. User feedback and accumulated task knowledge support subsequent task generation and operation.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop, workflow\nPublished: 2024-04-03\nAuthors: Tian Huang, Chun Yu, Weinan Shi, Zijian Peng, David Yang, Weiqi Sun, Yuanchun Shi\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.02475v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-02475/"}
{"id": "arxiv:2404.03648", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "340206da25df688ff73da617fd66a872586d1e36981c431d18331cee9b9669e0", "submitted_at": "2024-04-04T17:58:40Z", "updated_at": "2024-10-12T05:05:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.03648v2"}, "authors": ["Hanyu Lai", "Xiao Liu", "Iat Long Iong", "Shuntian Yao", "Yuxuan Chen", "Pengbo Shen", "Hao Yu", "Hanchen Zhang", "Xiaohan Zhang", "Yuxiao Dong", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2404.03648", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-03648/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "340206da25df688ff73da617fd66a872586d1e36981c431d18331cee9b9669e0", "submitted_at": "2024-04-04T17:58:40Z", "updated_at": "2024-10-12T05:05:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.03648v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.03648"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.03648"}], "name": "AutoWebGLM: A Large Language Model-based Web Navigating Agent", "publication": null, "published_at": "2024-04-04", "record_sha256": "1f557c956db7d53c2ae44f36a162115d7f8cc2cd8afb10a817df7c0372bc4cec", "relations": [], "resource_id": "arxiv:2404.03648", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.03648v2", "topics": ["web-navigation", "reinforcement-learning", "data-generation"], "updated_at": "2024-10-12"}, "text": "AutoWebGLM: A Large Language Model-based Web Navigating Agent\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds a web agent through simplified HTML observations, human-and-AI browsing data, curriculum training, and reinforcement-learning refinement.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, data-generation\nPublished: 2024-04-04\nAuthors: Hanyu Lai, Xiao Liu, Iat Long Iong, Shuntian Yao, Yuxuan Chen, Pengbo Shen, Hao Yu, Hanchen Zhang, Xiaohan Zhang, Yuxiao Dong, Jie Tang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.03648v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-03648/"}
{"id": "arxiv:2404.05719", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67e139ad83ab4804dfdc4c91f3fd344d7d3d279b97efb25bd11bab1532968e46", "submitted_at": "2024-04-08T17:55:44Z", "updated_at": "2024-04-08T17:55:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05719v1"}, "authors": ["Keen You", "Haotian Zhang", "Eldon Schoop", "Floris Weers", "Amanda Swearngin", "Jeffrey Nichols", "Yinfei Yang", "Zhe Gan"], "canonical_source_url": "https://arxiv.org/abs/2404.05719", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-05719/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67e139ad83ab4804dfdc4c91f3fd344d7d3d279b97efb25bd11bab1532968e46", "submitted_at": "2024-04-08T17:55:44Z", "updated_at": "2024-04-08T17:55:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05719v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.05719"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05719"}], "name": "Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs", "publication": null, "published_at": "2024-04-08", "record_sha256": "73409a8516a653efe7a93f6a2205ff687ea734d9f588260c39b3adaabe0a2b52", "relations": [], "resource_id": "arxiv:2404.05719", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.05719v1", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-04-08"}, "text": "Ferret-UI: Grounded Mobile UI Understanding with Multimodal LLMs\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts multimodal modeling to mobile interfaces through detailed screen crops, region annotations, and UI reasoning tasks.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\nPublished: 2024-04-08\nAuthors: Keen You, Haotian Zhang, Eldon Schoop, Floris Weers, Amanda Swearngin, Jeffrey Nichols, Yinfei Yang, Zhe Gan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.05719v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-05719/"}
{"id": "arxiv:2404.05902", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c18dbfedcccf6434b3a50a9aa62552372df8b772ccda6a1aa699a8f052476bc", "submitted_at": "2024-04-08T23:10:47Z", "updated_at": "2024-04-08T23:10:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05902v1"}, "authors": ["Michael Lutz", "Arth Bohra", "Manvel Saroyan", "Artem Harutyunyan", "Giovanni Campagna"], "canonical_source_url": "https://arxiv.org/abs/2404.05902", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-05902/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c18dbfedcccf6434b3a50a9aa62552372df8b772ccda6a1aa699a8f052476bc", "submitted_at": "2024-04-08T23:10:47Z", "updated_at": "2024-04-08T23:10:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05902v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2404.05902"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05902"}], "name": "WILBUR: Adaptive In-Context Learning for Robust and Accurate Web Agents", "publication": null, "published_at": "2024-04-08", "record_sha256": "2779ff006dc02211c4785a378e6dd58bfd0c0c4a4cc8c58a46c89dd05b3cb7dd", "relations": [], "resource_id": "arxiv:2404.05902", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.05902v1", "topics": ["web-navigation", "memory", "robustness", "data-generation"], "updated_at": "2024-04-08"}, "text": "WILBUR: Adaptive In-Context Learning for Robust and Accurate Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWILBUR learns to rank prior task demonstrations for a web agent’s prompt and synthesizes instructions from past runs. A generated task curriculum and a backtracking mechanism support adaptation and recovery, evaluated on WebVoyager.\nPlatforms: web\nTopics: web-navigation, memory, robustness, data-generation\nPublished: 2024-04-08\nAuthors: Michael Lutz, Arth Bohra, Manvel Saroyan, Artem Harutyunyan, Giovanni Campagna\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.05902v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-05902/"}
{"id": "arxiv:2404.05955", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7414bfa289baf86a8f2bb54f3f69f25a6a4cc8516fb0296a1899d21ad3020d4f", "submitted_at": "2024-04-09T02:29:39Z", "updated_at": "2024-04-09T02:29:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05955v1"}, "authors": ["Junpeng Liu", "Yifan Song", "Bill Yuchen Lin", "Wai Lam", "Graham Neubig", "Yuanzhi Li", "Xiang Yue"], "canonical_source_url": "https://arxiv.org/abs/2404.05955", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-05955/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7414bfa289baf86a8f2bb54f3f69f25a6a4cc8516fb0296a1899d21ad3020d4f", "submitted_at": "2024-04-09T02:29:39Z", "updated_at": "2024-04-09T02:29:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.05955v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.05955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.05955"}], "name": "VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?", "publication": null, "published_at": "2024-04-09", "record_sha256": "94c89b9417a97b807fb22518f09ab33591aea5c0f0ccdaf752a9a6a41099cd4e", "relations": [], "resource_id": "arxiv:2404.05955", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.05955v1", "topics": ["evaluation", "screen-understanding", "grounding"], "updated_at": "2024-04-09"}, "text": "VisualWebBench: How Far Have Multimodal LLMs Evolved in Web Page Understanding and Grounding?\nResource type: paper\nReview status: curated\nScope: core\n\nSeparately evaluates web-page reading, understanding, and element grounding across tasks drawn from real websites.\nPlatforms: web\nTopics: evaluation, screen-understanding, grounding\nPublished: 2024-04-09\nAuthors: Junpeng Liu, Yifan Song, Bill Yuchen Lin, Wai Lam, Graham Neubig, Yuanzhi Li, Xiang Yue\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.05955v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-05955/"}
{"id": "arxiv:2404.06474", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abc75ee55c7ac5992656dcec2106e286c670a58130a59b0eef7f5fd74d7496c5", "submitted_at": "2024-04-09T17:25:47Z", "updated_at": "2024-10-07T14:19:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.06474v3"}, "authors": ["Jiayi Pan", "Yichi Zhang", "Nicholas Tomlin", "Yifei Zhou", "Sergey Levine", "Alane Suhr"], "canonical_source_url": "https://arxiv.org/abs/2404.06474", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-06474/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "abc75ee55c7ac5992656dcec2106e286c670a58130a59b0eef7f5fd74d7496c5", "submitted_at": "2024-04-09T17:25:47Z", "updated_at": "2024-10-07T14:19:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.06474v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.06474"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.06474"}], "name": "Autonomous Evaluation and Refinement of Digital Agents", "publication": null, "published_at": "2024-04-09", "record_sha256": "c4d03316767bdc8138969779410408c2815e09a292ece9278e7f4c40059bbe0b", "relations": [], "resource_id": "arxiv:2404.06474", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.06474v3", "topics": ["evaluation", "reward-models"], "updated_at": "2024-10-07"}, "text": "Autonomous Evaluation and Refinement of Digital Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCompares automatic trajectory evaluators for web navigation and device control against oracle judgments. The resulting feedback is used for agent fine-tuning and inference-time guidance, including experiments on WebArena.\nPlatforms: web\nTopics: evaluation, reward-models\nPublished: 2024-04-09\nAuthors: Jiayi Pan, Yichi Zhang, Nicholas Tomlin, Yifei Zhou, Sergey Levine, Alane Suhr\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.06474v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-06474/"}
{"id": "arxiv:2404.07972", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a347bfcb4de03a60410392b52d522e5b7ad5dbabb474d60fc1ec097e74c4008d", "submitted_at": "2024-04-11T17:56:05Z", "updated_at": "2024-05-30T08:55:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.07972v2"}, "authors": ["Tianbao Xie", "Danyang Zhang", "Jixuan Chen", "Xiaochuan Li", "Siheng Zhao", "Ruisheng Cao", "Toh Jing Hua", "Zhoujun Cheng", "Dongchan Shin", "Fangyu Lei", "Yitao Liu", "Yiheng Xu", "Shuyan Zhou", "Silvio Savarese", "Caiming Xiong", "Victor Zhong", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2404.07972", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-07972/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a347bfcb4de03a60410392b52d522e5b7ad5dbabb474d60fc1ec097e74c4008d", "submitted_at": "2024-04-11T17:56:05Z", "updated_at": "2024-05-30T08:55:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.07972v2"}, "publication": {"authors": ["Tianbao Xie", "Danyang Zhang", "Jixuan Chen", "Xiaochuan Li", "Siheng Zhao", "Ruisheng Cao", "Toh J. Hua", "Zhoujun Cheng", "Dongchan Shin", "Fangyu Lei", "Yitao Liu", "Yiheng Xu", "Shuyan Zhou", "Silvio Savarese", "Caiming Xiong", "Victor Zhong", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-1650", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "52040-52094", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/5d413e48f84dc61244b6be550f1cd8f5-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "f62cbd2e5c92d250cc894a05e46f9052cfd38f0cdbf14902af6a96b96bcc9e8d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.07972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.07972"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/5d413e48f84dc61244b6be550f1cd8f5-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "publication": {"authors": ["Tianbao Xie", "Danyang Zhang", "Jixuan Chen", "Xiaochuan Li", "Siheng Zhao", "Ruisheng Cao", "Toh J. Hua", "Zhoujun Cheng", "Dongchan Shin", "Fangyu Lei", "Yitao Liu", "Yiheng Xu", "Shuyan Zhou", "Silvio Savarese", "Caiming Xiong", "Victor Zhong", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-1650", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "52040-52094", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/5d413e48f84dc61244b6be550f1cd8f5-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-04-11", "record_sha256": "12d7508c604d863520bd2f7681aa00c3561ac4d9b286b563c2f7bb1aeeeada36", "relations": [], "resource_id": "arxiv:2404.07972", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.07972v2", "topics": ["evaluation", "cross-app"], "updated_at": "2024-05-30"}, "text": "OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a real-computer evaluation environment with execution-based checks for tasks spanning desktop applications and operating-system state.\nPlatforms: desktop, linux, windows, macos\nTopics: evaluation, cross-app\nPublished: 2024-04-11\nAuthors: Tianbao Xie, Danyang Zhang, Jixuan Chen, Xiaochuan Li, Siheng Zhao, Ruisheng Cao, Toh Jing Hua, Zhoujun Cheng, Dongchan Shin, Fangyu Lei, Yitao Liu, Yiheng Xu, Shuyan Zhou, Silvio Savarese, Caiming Xiong, Victor Zhong, Tao Yu\nRecorded arXiv version: v2\n\nPublication edition: OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments\nPublication authors: Tianbao Xie, Danyang Zhang, Jixuan Chen, Xiaochuan Li, Siheng Zhao, Ruisheng Cao, Toh J. Hua, Zhoujun Cheng, Dongchan Shin, Fangyu Lei, Yitao Liu, Yiheng Xu, Shuyan Zhou, Silvio Savarese, Caiming Xiong, Victor Zhong, Tao Yu\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 52040-52094\nPublication doi: 10.52202/079017-1650\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/5d413e48f84dc61244b6be550f1cd8f5-Abstract-Datasets_and_Benchmarks_Track.html\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.07972v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-07972/"}
{"id": "arxiv:2404.08755", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f255b2c2870d9fbe5a543f2a0e09a30c3154fa0c13e9f46eedcefc11d2be43", "submitted_at": "2024-04-12T18:28:44Z", "updated_at": "2024-04-12T18:28:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.08755v1"}, "authors": ["Nicolai Dorka", "Janusz Marecki", "Ammar Anwar"], "canonical_source_url": "https://arxiv.org/abs/2404.08755", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-08755/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f255b2c2870d9fbe5a543f2a0e09a30c3154fa0c13e9f46eedcefc11d2be43", "submitted_at": "2024-04-12T18:28:44Z", "updated_at": "2024-04-12T18:28:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2404.08755v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.08755"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08755"}], "name": "Training a Vision Language Model as Smartphone Assistant", "publication": null, "published_at": "2024-04-12", "record_sha256": "97ce5ffd5e450e55198985c854965e192d9f2d7cf0442afbf8b5eabda82a7b60", "relations": [], "resource_id": "arxiv:2404.08755", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.08755v1", "topics": ["mobile-navigation", "interaction-history", "action-prediction"], "updated_at": "2024-04-12"}, "text": "Training a Vision Language Model as Smartphone Assistant\nResource type: paper\nReview status: curated\nScope: core\n\nTrains a smartphone assistant from sequences of screenshots paired with earlier actions. The model predicts UI-only interactions such as taps and swipes, with evaluation on Android in the Wild.\nPlatforms: android\nTopics: mobile-navigation, interaction-history, action-prediction\nPublished: 2024-04-12\nAuthors: Nicolai Dorka, Janusz Marecki, Ammar Anwar\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.08755v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-08755/"}
{"id": "arxiv:2404.08860", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3a030a53a942eacfabafc64af67591fd0a85af66291c5d1b48baea14556e8aa", "submitted_at": "2024-04-13T00:20:09Z", "updated_at": "2024-07-09T01:14:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.08860v3"}, "authors": ["Lei Ding", "Jeshwanth Bheemanpally", "Yi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2404.08860", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-08860/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b3a030a53a942eacfabafc64af67591fd0a85af66291c5d1b48baea14556e8aa", "submitted_at": "2024-04-13T00:20:09Z", "updated_at": "2024-07-09T01:14:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.08860v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.08860"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08860"}], "name": "Enhancing Mobile \"How-to\" Queries with Automated Search Results Verification and Reranking", "publication": null, "published_at": "2024-04-13", "record_sha256": "5b606ad9dba97dbfb90e668798d485af0d3c32a8df9b19c2d608ff0e15a46112", "relations": [], "resource_id": "arxiv:2404.08860", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.08860v3", "topics": ["mobile-navigation", "evaluation", "automation"], "updated_at": "2024-07-09"}, "text": "Enhancing Mobile \"How-to\" Queries with Automated Search Results Verification and Reranking\nResource type: paper\nReview status: curated\nScope: core\n\nExecutes instructions from mobile how-to search results inside a controlled Android environment, then uses the observed success of those instructions to verify and rerank technical-support results.\nPlatforms: android\nTopics: mobile-navigation, evaluation, automation\nPublished: 2024-04-13\nAuthors: Lei Ding, Jeshwanth Bheemanpally, Yi Zhang\nRecorded arXiv version: v3\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.08860v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-08860/"}
{"id": "arxiv:2404.08948", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a844cdd06a32b8e89f362b7e924b36e27e72c1f1def81ed0b934f761ffcaba78", "submitted_at": "2024-04-13T09:56:50Z", "updated_at": "2026-06-25T14:05:55Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2404.08948v4"}, "authors": ["Chenhui Cui", "Tao Li", "Junjie Wang", "Chunyang Chen", "Dave Towey", "Rubing Huang"], "canonical_source_url": "https://arxiv.org/abs/2404.08948", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-08948/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a844cdd06a32b8e89f362b7e924b36e27e72c1f1def81ed0b934f761ffcaba78", "submitted_at": "2024-04-13T09:56:50Z", "updated_at": "2026-06-25T14:05:55Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2404.08948v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.08948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.08948"}], "name": "Large Language Models for Mobile GUI Text Input Generation: An Empirical Study", "publication": null, "published_at": "2024-04-13", "record_sha256": "e8ab436bb3c5846c12c6764d2e0a59c67f551b11c6d14414581bb7608c020514", "relations": [], "resource_id": "arxiv:2404.08948", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.08948v4", "topics": ["automation", "evaluation"], "updated_at": "2026-06-25"}, "text": "Large Language Models for Mobile GUI Text Input Generation: An Empirical Study\nResource type: paper\nReview status: curated\nScope: core\n\nStudies language-model generation of mobile text inputs, comparing UI representations, execution feedback, and human assistance for app testing.\nPlatforms: android\nTopics: automation, evaluation\nPublished: 2024-04-13\nAuthors: Chenhui Cui, Tao Li, Junjie Wang, Chunyang Chen, Dave Towey, Rubing Huang\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.08948v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-08948/"}
{"id": "arxiv:2404.09992", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "667b1f5053f3ae6ea712999932d5e9ee2f42e402cc72e713048d0790f5931976", "submitted_at": "2024-04-15T17:59:50Z", "updated_at": "2025-06-29T05:24:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.09992v2"}, "authors": ["Shulin Tian", "Ziniu Zhang", "Liangyu Chen", "Ziwei Liu"], "canonical_source_url": "https://arxiv.org/abs/2404.09992", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-09992/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "667b1f5053f3ae6ea712999932d5e9ee2f42e402cc72e713048d0790f5931976", "submitted_at": "2024-04-15T17:59:50Z", "updated_at": "2025-06-29T05:24:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.09992v2"}, "publication": {"authors": ["Shulin Tian", "Ziniu Zhang", "Liang-Yu Chen", "Ziwei Liu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.703", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.703/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "13682-13697", "pdf_url": "https://aclanthology.org/2025.findings-acl.703.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MMInA: Benchmarking Multihop Multimodal Internet Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.703/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "5e725b77b94b484284adaab9ec70674c9fc13a7fce34b764e59561f9180b915f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.703/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.09992"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.09992"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.703/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.703.pdf"}], "name": "MMInA: Benchmarking Multihop Multimodal Internet Agents", "publication": {"authors": ["Shulin Tian", "Ziniu Zhang", "Liang-Yu Chen", "Ziwei Liu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.703", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.703/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "13682-13697", "pdf_url": "https://aclanthology.org/2025.findings-acl.703.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MMInA: Benchmarking Multihop Multimodal Internet Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.703/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-04-15", "record_sha256": "adfe6dfb4bf1cceaa7451c50993cab54e4e72e8aa2500669915ba13cdcde0745", "relations": [], "resource_id": "arxiv:2404.09992", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.09992v2", "topics": ["evaluation", "long-horizon", "memory"], "updated_at": "2025-06-29"}, "text": "MMInA: Benchmarking Multihop Multimodal Internet Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates multimodal agents on compositional tasks across changing websites and studies replaying prior trajectories as memory support.\nPlatforms: web\nTopics: evaluation, long-horizon, memory\nPublished: 2024-04-15\nAuthors: Shulin Tian, Ziniu Zhang, Liangyu Chen, Ziwei Liu\nRecorded arXiv version: v2\n\nPublication edition: MMInA: Benchmarking Multihop Multimodal Internet Agents\nPublication authors: Shulin Tian, Ziniu Zhang, Liang-Yu Chen, Ziwei Liu\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 13682-13697\nPublication doi: 10.18653/v1/2025.findings-acl.703\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.findings-acl.703/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.09992v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-09992/"}
{"id": "arxiv:2404.10887", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "127d037bb462ae4ed7bd72c46a62e2c4427611b30b67193454e40cc5c83cca3f", "submitted_at": "2024-04-16T20:15:32Z", "updated_at": "2025-01-26T19:07:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.10887v2"}, "authors": ["Moghis Fereidouni", "Adib Mosharrof", "A. B. Siddique"], "canonical_source_url": "https://arxiv.org/abs/2404.10887", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-10887/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "127d037bb462ae4ed7bd72c46a62e2c4427611b30b67193454e40cc5c83cca3f", "submitted_at": "2024-04-16T20:15:32Z", "updated_at": "2025-01-26T19:07:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.10887v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.10887"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.10887"}], "name": "Grounded Language Agent for Product Search via Intelligent Web Interactions", "publication": null, "published_at": "2024-04-16", "record_sha256": "59d605be46d96c1cfea84e75d15702b29a9a809fbb30e641f39ad7dabe7a68d3", "relations": [], "resource_id": "arxiv:2404.10887", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.10887v2", "topics": ["web-navigation", "reinforcement-learning", "imitation-learning"], "updated_at": "2025-01-26"}, "text": "Grounded Language Agent for Product Search via Intelligent Web Interactions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GLAINTEL, a Flan-T5-based product-search web agent studied under unsupervised learning, demonstration-based training, and unsupervised domain adaptation. It compares ways to combine demonstrations with reinforcement learning.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, imitation-learning\nPublished: 2024-04-16\nAuthors: Moghis Fereidouni, Adib Mosharrof, A. B. Siddique\nRecorded arXiv version: v2\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.10887v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-10887/"}
{"id": "arxiv:2404.16048", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e29d14a41d59942c88938b8e83217484a1cb619847c5dff0ec1f516095ee5654", "submitted_at": "2024-04-09T11:59:41Z", "updated_at": "2024-10-27T05:54:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16048v2"}, "authors": ["Rajat Chawla", "Adarsh Jha", "Muskaan Kumar", "Mukunda NS", "Ishaan Bhola"], "canonical_source_url": "https://arxiv.org/abs/2404.16048", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-16048/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e29d14a41d59942c88938b8e83217484a1cb619847c5dff0ec1f516095ee5654", "submitted_at": "2024-04-09T11:59:41Z", "updated_at": "2024-10-27T05:54:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16048v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.16048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16048"}], "name": "GUIDE: Graphical User Interface Data for Execution", "publication": null, "published_at": "2024-04-09", "record_sha256": "28efbfd412d41e3290f4048f9950ff48abe0ddc75008aad77863014ef357b15c", "relations": [], "resource_id": "arxiv:2404.16048", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.16048v2", "topics": ["data-generation", "grounding"], "updated_at": "2024-10-27"}, "text": "GUIDE: Graphical User Interface Data for Execution\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces annotated website task trajectories with screenshots, prior actions, reasoning, and grounded next actions for process automation.\nPlatforms: web\nTopics: data-generation, grounding\nPublished: 2024-04-09\nAuthors: Rajat Chawla, Adarsh Jha, Muskaan Kumar, Mukunda NS, Ishaan Bhola\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.16048v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-16048/"}
{"id": "arxiv:2404.16054", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28c3ad086106ec4199dcc9da2b8aa3b20eac9347d9354b9dfb373e50ea4c3f50", "submitted_at": "2024-04-12T15:39:09Z", "updated_at": "2024-08-02T13:49:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16054v2"}, "authors": ["Li Zhang", "Shihe Wang", "Xianqing Jia", "Zhihan Zheng", "Yunhe Yan", "Longxi Gao", "Yuanchun Li", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2404.16054", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-16054/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28c3ad086106ec4199dcc9da2b8aa3b20eac9347d9354b9dfb373e50ea4c3f50", "submitted_at": "2024-04-12T15:39:09Z", "updated_at": "2024-08-02T13:49:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2404.16054v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.16054"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16054"}], "name": "LlamaTouch: A Faithful and Scalable Testbed for Mobile UI Task Automation", "publication": null, "published_at": "2024-04-12", "record_sha256": "5b2797bc9371d8be404beb6626fa9d183673f6ef93722bb8e7ea05e1c7ae21fa", "relations": [], "resource_id": "arxiv:2404.16054", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.16054v2", "topics": ["evaluation", "mobile-navigation", "state-tracking"], "updated_at": "2024-08-02"}, "text": "LlamaTouch: A Faithful and Scalable Testbed for Mobile UI Task Automation\nResource type: paper\nReview status: curated\nScope: core\n\nLlamaTouch executes mobile tasks on devices and evaluates whether an agent visits manually annotated essential application or system states. Screenshot and hierarchy annotations, combined with exact and fuzzy state matching, support evaluation of 496 tasks across multiple agents.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, state-tracking\nPublished: 2024-04-12\nAuthors: Li Zhang, Shihe Wang, Xianqing Jia, Zhihan Zheng, Yunhe Yan, Longxi Gao, Yuanchun Li, Mengwei Xu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.16054v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-16054/"}
{"id": "arxiv:2404.16660", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45c13064bbfbcf153566b86c8969f4de5b5e87dc133342c595b3b58bde3bb615", "submitted_at": "2024-04-25T14:56:32Z", "updated_at": "2025-07-21T02:55:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.16660v3"}, "authors": ["Juyong Lee", "Taywon Min", "Minyong An", "Dongyoon Hahm", "Haeone Lee", "Changyeon Kim", "Kimin Lee"], "canonical_source_url": "https://arxiv.org/abs/2404.16660", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-16660/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45c13064bbfbcf153566b86c8969f4de5b5e87dc133342c595b3b58bde3bb615", "submitted_at": "2024-04-25T14:56:32Z", "updated_at": "2025-07-21T02:55:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.16660v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.16660"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.16660"}], "name": "Benchmarking Mobile Device Control Agents across Diverse Configurations", "publication": null, "published_at": "2024-04-25", "record_sha256": "95f894a4d56b53c35a17a43bd369df177f9ec57576084d6ef47cb15f120484aa", "relations": [], "resource_id": "arxiv:2404.16660", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.16660v3", "topics": ["evaluation", "robustness"], "updated_at": "2025-07-21"}, "text": "Benchmarking Mobile Device Control Agents across Diverse Configurations\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces B-MoCA, an interactive mobile-control benchmark that varies device settings, layouts, and languages to test generalization.\nPlatforms: android\nTopics: evaluation, robustness\nPublished: 2024-04-25\nAuthors: Juyong Lee, Taywon Min, Minyong An, Dongyoon Hahm, Haeone Lee, Changyeon Kim, Kimin Lee\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.16660v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-16660/"}
{"id": "arxiv:2404.18074", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e531ec2ef9e5d90586b8f3052e16d41da95114b7382047d0a02fd671a0db355", "submitted_at": "2024-04-28T05:33:15Z", "updated_at": "2025-03-23T13:04:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.18074v3"}, "authors": ["Zirui Song", "Yaohang Li", "Meng Fang", "Yanda Li", "Zhenhao Chen", "Zecheng Shi", "Yuan Huang", "Xiuying Chen", "Ling Chen"], "canonical_source_url": "https://arxiv.org/abs/2404.18074", "citation_url": "https://awesomecua.github.io/resources/arxiv-2404-18074/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e531ec2ef9e5d90586b8f3052e16d41da95114b7382047d0a02fd671a0db355", "submitted_at": "2024-04-28T05:33:15Z", "updated_at": "2025-03-23T13:04:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2404.18074v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2404.18074"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2404.18074"}], "name": "MMAC-Copilot: Multi-modal Agent Collaboration Operating Copilot", "publication": null, "published_at": "2024-04-28", "record_sha256": "75f0192c480e3c6ea49f61577ac7c32458feafbf765ecc852b48f4a1729d4922", "relations": [], "resource_id": "arxiv:2404.18074", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2404.18074v3", "topics": ["automation", "evaluation", "planning"], "updated_at": "2025-03-23"}, "text": "MMAC-Copilot: Multi-modal Agent Collaboration Operating Copilot\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MMAC-Copilot, which coordinates agents with different expertise to operate PC applications. VIBench evaluates visual interaction with applications that lack usable APIs, spanning office, recreation and game tasks.\nPlatforms: desktop\nTopics: automation, evaluation, planning\nPublished: 2024-04-28\nAuthors: Zirui Song, Yaohang Li, Meng Fang, Yanda Li, Zhenhao Chen, Zecheng Shi, Yuan Huang, Xiuying Chen, Ling Chen\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2404.18074v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2404-18074/"}
{"id": "arxiv:2405.00516", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3498a7bc289c6537ab69780a56a9fd0a2bb697fb74c077949824c4fc018762c6", "submitted_at": "2024-05-01T13:51:45Z", "updated_at": "2024-05-01T13:51:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.00516v1"}, "authors": ["Lucas-Andreï Thil", "Mirela Popa", "Gerasimos Spanakis"], "canonical_source_url": "https://arxiv.org/abs/2405.00516", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-00516/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3498a7bc289c6537ab69780a56a9fd0a2bb697fb74c077949824c4fc018762c6", "submitted_at": "2024-05-01T13:51:45Z", "updated_at": "2024-05-01T13:51:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.00516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2405.00516"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.00516"}], "name": "Navigating WebAI: Training Agents to Complete Web Tasks with Large Language Models and Reinforcement Learning", "publication": null, "published_at": "2024-05-01", "record_sha256": "5b23d7b8b6e1b4353e5bdf34f9187a0ffd172e0bbc45eca75f37855add6331f2", "relations": [], "resource_id": "arxiv:2405.00516", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.00516v1", "topics": ["web-navigation", "reinforcement-learning", "evaluation"], "updated_at": "2024-05-01"}, "text": "Navigating WebAI: Training Agents to Complete Web Tasks with Large Language Models and Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nCombines supervised and reinforcement learning for MiniWoB web tasks and investigates whether agents understand HTML structure or memorize target elements.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, evaluation\nPublished: 2024-05-01\nAuthors: Lucas-Andreï Thil, Mirela Popa, Gerasimos Spanakis\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.00516v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-00516/"}
{"id": "arxiv:2405.04497", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a358d417d8bef89a01e799d88564e3ede4678ebda4950e87c96cf1e1954b85e1", "submitted_at": "2024-05-07T17:10:31Z", "updated_at": "2024-05-08T05:44:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.04497v2"}, "authors": ["Kihoon Son", "Jinhyeon Kwon", "DaEun Choi", "Tae Soo Kim", "Young-Ho Kim", "Sangdoo Yun", "Juho Kim"], "canonical_source_url": "https://arxiv.org/abs/2405.04497", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-04497/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a358d417d8bef89a01e799d88564e3ede4678ebda4950e87c96cf1e1954b85e1", "submitted_at": "2024-05-07T17:10:31Z", "updated_at": "2024-05-08T05:44:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.04497v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2405.04497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.04497"}], "name": "Unveiling Disparities in Web Task Handling Between Human and Web Agent", "publication": null, "published_at": "2024-05-07", "record_sha256": "63b18dd33e8a3ceb98e7c8089482ea64a022c7457c591815587684222968ed20", "relations": [], "resource_id": "arxiv:2405.04497", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.04497v2", "topics": ["evaluation", "human-in-the-loop", "planning"], "updated_at": "2024-05-08"}, "text": "Unveiling Disparities in Web Task Handling Between Human and Web Agent\nResource type: paper\nReview status: curated\nScope: core\n\nCompares human think-aloud web-task behavior with web-agent planning, action, and reflection. The analysis highlights differences in updating plans, handling ambiguity, and investigating failures.\nPlatforms: web\nTopics: evaluation, human-in-the-loop, planning\nPublished: 2024-05-07\nAuthors: Kihoon Son, Jinhyeon Kwon, DaEun Choi, Tae Soo Kim, Young-Ho Kim, Sangdoo Yun, Juho Kim\nRecorded arXiv version: v2\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.04497v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-04497/"}
{"id": "arxiv:2405.07090", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d200a997596ea919d5210b5ca5760241f9dfa3ad581e090a541871cd6e117522", "submitted_at": "2024-05-11T20:32:16Z", "updated_at": "2024-05-11T20:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.07090v1"}, "authors": ["Sidong Feng", "Suyu Ma", "Han Wang", "David Kong", "Chunyang Chen"], "canonical_source_url": "https://arxiv.org/abs/2405.07090", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-07090/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d200a997596ea919d5210b5ca5760241f9dfa3ad581e090a541871cd6e117522", "submitted_at": "2024-05-11T20:32:16Z", "updated_at": "2024-05-11T20:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.07090v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2405.07090"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.07090"}], "name": "MUD: Towards a Large-Scale and Noise-Filtered UI Dataset for Modern Style UI Modeling", "publication": null, "published_at": "2024-05-11", "record_sha256": "85cc216db741141e1ec182ed01f973014e25b479c583cce89a95aa467edbab57", "relations": [], "resource_id": "arxiv:2405.07090", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.07090v1", "topics": ["data-generation", "screen-understanding"], "updated_at": "2024-05-11"}, "text": "MUD: Towards a Large-Scale and Noise-Filtered UI Dataset for Modern Style UI Modeling\nResource type: paper\nReview status: curated\nScope: core\n\nMUD combines LLM-guided Android app exploration, noise filtering and human annotation to collect modern UI screenshots. The study evaluates the resulting data for element detection and UI retrieval.\nPlatforms: android\nTopics: data-generation, screen-understanding\nPublished: 2024-05-11\nAuthors: Sidong Feng, Suyu Ma, Han Wang, David Kong, Chunyang Chen\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.07090v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-07090/"}
{"id": "arxiv:2405.11120", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81f62326dddc5c512fb05472eb1f387064c0a5aad77453e9d2e85415bb627fa9", "submitted_at": "2024-05-17T23:27:33Z", "updated_at": "2024-05-17T23:27:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.11120v1"}, "authors": ["William E Bishop", "Alice Li", "Christopher Rawles", "Oriana Riva"], "canonical_source_url": "https://arxiv.org/abs/2405.11120", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-11120/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81f62326dddc5c512fb05472eb1f387064c0a5aad77453e9d2e85415bb627fa9", "submitted_at": "2024-05-17T23:27:33Z", "updated_at": "2024-05-17T23:27:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.11120v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2405.11120"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.11120"}], "name": "Latent State Estimation Helps UI Agents to Reason", "publication": null, "published_at": "2024-05-17", "record_sha256": "e39bfc4be3adfb5b869e48d4e5aa16f13f49509d796010ca9b784c6b8070752e", "relations": [], "resource_id": "arxiv:2405.11120", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.11120v1", "topics": ["state-tracking", "planning", "evaluation"], "updated_at": "2024-05-17"}, "text": "Latent State Estimation Helps UI Agents to Reason\nResource type: paper\nReview status: curated\nScope: core\n\nStudies explicit textual estimates of latent UI state, including whether commanded actions occurred and how far a task has progressed. The experiments compare agents that reason with these estimates against agents that omit them across public and internal task benchmarks.\nPlatforms: unspecified\nTopics: state-tracking, planning, evaluation\nPublished: 2024-05-17\nAuthors: William E Bishop, Alice Li, Christopher Rawles, Oriana Riva\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.11120v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-11120/"}
{"id": "arxiv:2405.12842", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca8d7e7dc3a537573ac3990d532589cdca12bf441ee40521cc46044fd10c5263", "submitted_at": "2024-05-21T14:49:12Z", "updated_at": "2024-05-21T14:49:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.12842v1"}, "authors": ["Arushi Jain", "Shubham Paliwal", "Monika Sharma", "Lovekesh Vig", "Gautam Shroff"], "canonical_source_url": "https://arxiv.org/abs/2405.12842", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-12842/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ca8d7e7dc3a537573ac3990d532589cdca12bf441ee40521cc46044fd10c5263", "submitted_at": "2024-05-21T14:49:12Z", "updated_at": "2024-05-21T14:49:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2405.12842v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2405.12842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.12842"}], "name": "SmartFlow: Robotic Process Automation using LLMs", "publication": null, "published_at": "2024-05-21", "record_sha256": "6e53475c326e717a51ad3a5fef9c10c49ffcb9ea1f4ebf007635682246c613c1", "relations": [], "resource_id": "arxiv:2405.12842", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.12842v1", "topics": ["automation", "workflow", "screen-understanding"], "updated_at": "2024-05-21"}, "text": "SmartFlow: Robotic Process Automation using LLMs\nResource type: paper\nReview status: curated\nScope: core\n\nPresents SmartFlow, which converts visible GUI elements into text for language-model action planning and executes the plan through a scripting engine. A collection of enterprise applications evaluates adaptation to different layouts and tasks.\nPlatforms: unspecified\nTopics: automation, workflow, screen-understanding\nPublished: 2024-05-21\nAuthors: Arushi Jain, Shubham Paliwal, Monika Sharma, Lovekesh Vig, Gautam Shroff\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.12842v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-12842/"}
{"id": "arxiv:2405.14573", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a33b8b0a03c0dc7ac3f699c34193a526d20791bfedfdde2a2fb89604959e08c2", "submitted_at": "2024-05-23T13:48:54Z", "updated_at": "2025-04-06T20:37:50Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2405.14573v5"}, "authors": ["Christopher Rawles", "Sarah Clinckemaillie", "Yifan Chang", "Jonathan Waltz", "Gabrielle Lau", "Marybeth Fair", "Alice Li", "William Bishop", "Wei Li", "Folawiyo Campbell-Ajala", "Daniel Toyama", "Robert Berry", "Divya Tyamagundlu", "Timothy Lillicrap", "Oriana Riva"], "canonical_source_url": "https://arxiv.org/abs/2405.14573", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-14573/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a33b8b0a03c0dc7ac3f699c34193a526d20791bfedfdde2a2fb89604959e08c2", "submitted_at": "2024-05-23T13:48:54Z", "updated_at": "2025-04-06T20:37:50Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2405.14573v5"}, "publication": {"authors": ["Chris Rawles", "Sarah Clinckemaillie", "Yifan Chang", "Jonathan Waltz", "Gabrielle Lau", "Marybeth Fair", "Alice Li", "William Bishop", "Wei Li", "Folawiyo Campbell-Ajala", "Daniel Toyama", "Robert Berry", "Divya Tyamagundlu", "Timothy Lillicrap", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "406-441", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/01a83bc2f2732a58e6aa731e659e7101-Paper-Conference.pdf", "publisher": null, "series": null, "title": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "e525e5eeedcb08e61765eb0d4034c43381bd3ecc5cc4bf9cc8788cf50e7a2bca", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2405.14573"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.14573"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/01a83bc2f2732a58e6aa731e659e7101-Paper-Conference.pdf"}], "name": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents", "publication": {"authors": ["Chris Rawles", "Sarah Clinckemaillie", "Yifan Chang", "Jonathan Waltz", "Gabrielle Lau", "Marybeth Fair", "Alice Li", "William Bishop", "Wei Li", "Folawiyo Campbell-Ajala", "Daniel Toyama", "Robert Berry", "Divya Tyamagundlu", "Timothy Lillicrap", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "406-441", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/01a83bc2f2732a58e6aa731e659e7101-Paper-Conference.pdf", "publisher": null, "series": null, "title": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-05-23", "record_sha256": "6ba321909c19105bbb3c84a186a75a22e560d557451c30cac092abe4e3382ebc", "relations": [], "resource_id": "arxiv:2405.14573", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.14573v5", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-04-06"}, "text": "AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces parameterized Android tasks with initialization, success checks, and cleanup for reproducible interactive evaluation.\nPlatforms: android\nTopics: evaluation, mobile-navigation\nPublished: 2024-05-23\nAuthors: Christopher Rawles, Sarah Clinckemaillie, Yifan Chang, Jonathan Waltz, Gabrielle Lau, Marybeth Fair, Alice Li, William Bishop, Wei Li, Folawiyo Campbell-Ajala, Daniel Toyama, Robert Berry, Divya Tyamagundlu, Timothy Lillicrap, Oriana Riva\nRecorded arXiv version: v5\n\nPublication edition: AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents\nPublication authors: Chris Rawles, Sarah Clinckemaillie, Yifan Chang, Jonathan Waltz, Gabrielle Lau, Marybeth Fair, Alice Li, William Bishop, Wei Li, Folawiyo Campbell-Ajala, Daniel Toyama, Robert Berry, Divya Tyamagundlu, Timothy Lillicrap, Oriana Riva\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 406-441\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/01a83bc2f2732a58e6aa731e659e7101-Abstract-Conference.html\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.14573v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-14573/"}
{"id": "arxiv:2405.15341", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7cb6087a48596442e1d191bf1313e9459ad62c02c3104ee481cb72cb00e56b1", "submitted_at": "2024-05-24T08:21:45Z", "updated_at": "2024-07-21T07:34:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.15341v2"}, "authors": ["Abdur Rahman", "Rajat Chawla", "Muskaan Kumar", "Arkajit Datta", "Adarsh Jha", "Mukunda NS", "Ishaan Bhola"], "canonical_source_url": "https://arxiv.org/abs/2405.15341", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-15341/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7cb6087a48596442e1d191bf1313e9459ad62c02c3104ee481cb72cb00e56b1", "submitted_at": "2024-05-24T08:21:45Z", "updated_at": "2024-07-21T07:34:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.15341v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:34Z", "role": "primary", "sha256": "b29f4b79250cc5f0b59686d7690309d3492e649c8ef0e14285e01eacb000b1b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2312.09370%2C2401.11880%2C2401.17461%2C2402.00898%2C2402.01586%2C2402.01735%2C2402.03323%2C2402.04247%2C2402.05202%2C2402.06132%2C2402.06819%2C2402.09052%2C2402.12914%2C2402.13225%2C2402.14606%2C2402.15235%2C2402.15391%2C2402.17834%2C2402.17930%2C2402.19446%2C2403.01248%2C2403.02757%2C2403.03636%2C2403.04692%2C2403.04957%2C2403.08295%2C2403.12968%2C2403.13257%2C2403.14720%2C2403.15049%2C2403.15203%2C2404.01131%2C2404.01226%2C2404.01318%2C2404.09001%2C2404.12195%2C2404.12500%2C2404.12608%2C2404.12920%2C2404.13050%2C2404.13501%2C2405.03710%2C2405.03862%2C2405.06682%2C2405.07606%2C2405.09215%2C2405.14979%2C2405.15341%2C2406.02856%2C2406.09913&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "92875dc9298b7952cc6e33ce1165b806421a6e24f3cd1e9bfd951fdd928659ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2405.15341v2"}], "identifiers": {"arxiv": "2405.15341"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.15341"}], "name": "V-Zen: Efficient GUI Understanding and Precise Grounding With A Novel Multimodal LLM", "publication": null, "published_at": "2024-05-24", "record_sha256": "d83d967672846a37a13734a472bf0316b4dd03fecfd74ae317b5796cc242c99f", "relations": [], "resource_id": "arxiv:2405.15341", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.15341v2", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": "2024-07-21"}, "text": "V-Zen: Efficient GUI Understanding and Precise Grounding With A Novel Multimodal LLM\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes V-Zen, a multimodal GUI model with dual-resolution image encoders for grounding and next-action prediction, fine-tuned using GUIDE interface elements and task sequences.\nPlatforms: unspecified\nTopics: grounding, action-prediction, screen-understanding\nPublished: 2024-05-24\nAuthors: Abdur Rahman, Rajat Chawla, Muskaan Kumar, Arkajit Datta, Adarsh Jha, Mukunda NS, Ishaan Bhola\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. The related GUIDE dataset paper remains a separate record; this annotation describes the V-Zen model contribution.\n\nPrimary source: https://arxiv.org/abs/2405.15341v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-15341/"}
{"id": "arxiv:2405.20309", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77b292052a2ab4e8db61606ffe1da8bd5e63b3ce8c84b2e0e4b7a2fdc2f926ca", "submitted_at": "2024-05-30T17:52:36Z", "updated_at": "2024-10-01T21:28:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.20309v2"}, "authors": ["Ajay Patel", "Markus Hofmarcher", "Claudiu Leoveanu-Condrei", "Marius-Constantin Dinu", "Chris Callison-Burch", "Sepp Hochreiter"], "canonical_source_url": "https://arxiv.org/abs/2405.20309", "citation_url": "https://awesomecua.github.io/resources/arxiv-2405-20309/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77b292052a2ab4e8db61606ffe1da8bd5e63b3ce8c84b2e0e4b7a2fdc2f926ca", "submitted_at": "2024-05-30T17:52:36Z", "updated_at": "2024-10-01T21:28:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2405.20309v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2405.20309"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2405.20309"}], "name": "Large Language Models Can Self-Improve At Web Agent Tasks", "publication": null, "published_at": "2024-05-30", "record_sha256": "ded4bf1dbd18f389fe1a5988f94d42055a5c2284b79d4b6f4d200ff2f6bb6509", "relations": [], "resource_id": "arxiv:2405.20309", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2405.20309v2", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2024-10-01"}, "text": "Large Language Models Can Self-Improve At Web Agent Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nExamines self-improvement by fine-tuning web agents on three mixtures of model-generated training data. WebArena experiments and trajectory-level evaluation measures study task performance, robustness, and behavior beyond aggregate success rates.\nPlatforms: web\nTopics: data-generation, web-navigation, evaluation\nPublished: 2024-05-30\nAuthors: Ajay Patel, Markus Hofmarcher, Claudiu Leoveanu-Condrei, Marius-Constantin Dinu, Chris Callison-Burch, Sepp Hochreiter\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2405.20309v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2405-20309/"}
{"id": "arxiv:2406.01014", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1338f3543a04bc2f9734e5fcabbfd5ebda6ac263461c2305f7c39383e938573", "submitted_at": "2024-06-03T05:50:00Z", "updated_at": "2024-06-03T05:50:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01014v1"}, "authors": ["Junyang Wang", "Haiyang Xu", "Haitao Jia", "Xi Zhang", "Ming Yan", "Weizhou Shen", "Ji Zhang", "Fei Huang", "Jitao Sang"], "canonical_source_url": "https://arxiv.org/abs/2406.01014", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-01014/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a1338f3543a04bc2f9734e5fcabbfd5ebda6ac263461c2305f7c39383e938573", "submitted_at": "2024-06-03T05:50:00Z", "updated_at": "2024-06-03T05:50:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01014v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.01014"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.01014"}], "name": "Mobile-Agent-v2: Mobile Device Operation Assistant with Effective Navigation via Multi-Agent Collaboration", "publication": null, "published_at": "2024-06-03", "record_sha256": "cde495fdc311b839b5fd5675197bebdf91bdfb8a5d7ec79c122fbb70cda97c1f", "relations": [], "resource_id": "arxiv:2406.01014", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.01014v1", "topics": ["planning", "memory", "mobile-navigation"], "updated_at": "2024-06-03"}, "text": "Mobile-Agent-v2: Mobile Device Operation Assistant with Effective Navigation via Multi-Agent Collaboration\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates mobile planning, decision-making, and reflection, with a memory unit for task progress and relevant content.\nPlatforms: android\nTopics: planning, memory, mobile-navigation\nPublished: 2024-06-03\nAuthors: Junyang Wang, Haiyang Xu, Haitao Jia, Xi Zhang, Ming Yan, Weizhou Shen, Ji Zhang, Fei Huang, Jitao Sang\nRecorded arXiv version: v1\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.01014v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-01014/"}
{"id": "arxiv:2406.01623", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9a728edf0b834adb72abfc50e13d5b2b52f532f444b3e8af34effd3d86beb1b", "submitted_at": "2024-06-01T00:32:26Z", "updated_at": "2024-06-01T00:32:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01623v1"}, "authors": ["Eric Li", "Jim Waldo"], "canonical_source_url": "https://arxiv.org/abs/2406.01623", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-01623/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9a728edf0b834adb72abfc50e13d5b2b52f532f444b3e8af34effd3d86beb1b", "submitted_at": "2024-06-01T00:32:26Z", "updated_at": "2024-06-01T00:32:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.01623v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2406.01623"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.01623"}], "name": "WebSuite: Systematically Evaluating Why Web Agents Fail", "publication": null, "published_at": "2024-06-01", "record_sha256": "8f55bd14827ee05f15451eeb4d54dd4d9f453c4fe837b576ad4471606917ecf9", "relations": [], "resource_id": "arxiv:2406.01623", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.01623v1", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2024-06-01"}, "text": "WebSuite: Systematically Evaluating Why Web Agents Fail\nResource type: paper\nReview status: curated\nScope: core\n\nWebSuite organizes browser operations into an action taxonomy and tests both isolated controls and complete workflows. Its diagnostic tasks attribute failures to specific actions, enabling comparison of text-based and multimodal web agents.\nPlatforms: web\nTopics: evaluation, observability, web-navigation\nPublished: 2024-06-01\nAuthors: Eric Li, Jim Waldo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.01623v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-01623/"}
{"id": "arxiv:2406.03679", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb3209d8e2a431987fe19f01cd0818f9fb490e684d7369e7d3002dc388ff271", "submitted_at": "2024-06-06T01:49:29Z", "updated_at": "2024-11-13T16:42:22Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2406.03679v6"}, "authors": ["Wei Li", "William Bishop", "Alice Li", "Chris Rawles", "Folawiyo Campbell-Ajala", "Divya Tyamagundlu", "Oriana Riva"], "canonical_source_url": "https://arxiv.org/abs/2406.03679", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-03679/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb3209d8e2a431987fe19f01cd0818f9fb490e684d7369e7d3002dc388ff271", "submitted_at": "2024-06-06T01:49:29Z", "updated_at": "2024-11-13T16:42:22Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2406.03679v6"}, "publication": {"authors": ["Wei Li", "William Bishop", "Alice Li", "Chris Rawles", "Folawiyo Campbell-Ajala", "Divya Tyamagundlu", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2925", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "92130-92154", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a79f3ef3b445fd4659f44648f7ea8ffd-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "On the Effects of Data Scale on UI Control Agents", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "0f499dc380fcd787579aab846e3d1e39b720b672fd9916b4a016688218ca48dd", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.03679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.03679"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a79f3ef3b445fd4659f44648f7ea8ffd-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "On the Effects of Data Scale on UI Control Agents", "publication": {"authors": ["Wei Li", "William Bishop", "Alice Li", "Chris Rawles", "Folawiyo Campbell-Ajala", "Divya Tyamagundlu", "Oriana Riva"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2925", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "92130-92154", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a79f3ef3b445fd4659f44648f7ea8ffd-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "On the Effects of Data Scale on UI Control Agents", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-06-06", "record_sha256": "1aa23496314c084e9725f24f70858c5bcc0a0ac38f0b02aa5c27d6da524d6c3e", "relations": [], "resource_id": "arxiv:2406.03679", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.03679v6", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2024-11-13"}, "text": "On the Effects of Data Scale on UI Control Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how Android task performance changes with demonstration scale and releases AndroidControl with high- and low-level instructions.\nPlatforms: android\nTopics: data-generation, mobile-navigation, evaluation\nPublished: 2024-06-06\nAuthors: Wei Li, William Bishop, Alice Li, Chris Rawles, Folawiyo Campbell-Ajala, Divya Tyamagundlu, Oriana Riva\nRecorded arXiv version: v6\n\nPublication edition: On the Effects of Data Scale on UI Control Agents\nPublication authors: Wei Li, William Bishop, Alice Li, Chris Rawles, Folawiyo Campbell-Ajala, Divya Tyamagundlu, Oriana Riva\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 92130-92154\nPublication doi: 10.52202/079017-2925\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/a79f3ef3b445fd4659f44648f7ea8ffd-Abstract-Datasets_and_Benchmarks_Track.html\n\nAnnotation reviewed against the primary abstract. Original dataset scale and experimental results are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.03679v6\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-03679/"}
{"id": "arxiv:2406.06947", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f85b2cf7650f41ba9ee6e4ae83cd9041e077a44adcb685e702d644bd5779963", "submitted_at": "2024-06-11T05:21:20Z", "updated_at": "2024-12-26T11:02:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.06947v3"}, "authors": ["Junhee Cho", "Jihoon Kim", "Daseul Bae", "Jinho Choo", "Youngjune Gwon", "Yeong-Dae Kwon"], "canonical_source_url": "https://arxiv.org/abs/2406.06947", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-06947/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f85b2cf7650f41ba9ee6e4ae83cd9041e077a44adcb685e702d644bd5779963", "submitted_at": "2024-06-11T05:21:20Z", "updated_at": "2024-12-26T11:02:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.06947v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.06947"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.06947"}], "name": "CAAP: Context-Aware Action Planning Prompting to Solve Computer Tasks with Front-End UI Only", "publication": null, "published_at": "2024-06-11", "record_sha256": "81e29a3b395e458f5d236f58aab8b78d944015604831e3ee1a974fbcd0f2e282", "relations": [], "resource_id": "arxiv:2406.06947", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.06947v3", "topics": ["planning", "screen-understanding", "web-navigation"], "updated_at": "2024-12-26"}, "text": "CAAP: Context-Aware Action Planning Prompting to Solve Computer Tasks with Front-End UI Only\nResource type: paper\nReview status: curated\nScope: core\n\nCAAP converts screenshots into textual observations and plans keyboard-and-mouse actions using a prompt that examines task context from multiple perspectives. The agent is evaluated on MiniWoB++ and WebShop using only front-end visual observations.\nPlatforms: web\nTopics: planning, screen-understanding, web-navigation\nPublished: 2024-06-11\nAuthors: Junhee Cho, Jihoon Kim, Daseul Bae, Jinho Choo, Youngjune Gwon, Yeong-Dae Kwon\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.06947v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-06947/"}
{"id": "arxiv:2406.07822", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47d0ca0467a104f9b8be037ec73131f12b66a22b0a89fe975bec7680073c8f39", "submitted_at": "2024-06-12T02:43:19Z", "updated_at": "2024-08-08T02:36:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.07822v2"}, "authors": ["Andrea Burns", "Kate Saenko", "Bryan A. Plummer"], "canonical_source_url": "https://arxiv.org/abs/2406.07822", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-07822/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47d0ca0467a104f9b8be037ec73131f12b66a22b0a89fe975bec7680073c8f39", "submitted_at": "2024-06-12T02:43:19Z", "updated_at": "2024-08-08T02:36:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.07822v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.07822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.07822"}], "name": "Tell Me What's Next: Textual Foresight for Generic UI Representations", "publication": null, "published_at": "2024-06-12", "record_sha256": "ae93bf1c40aa90494d370cf47b29c813a666ded8a86b50343be55db850b58845", "relations": [], "resource_id": "arxiv:2406.07822", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.07822v2", "topics": ["screen-understanding", "world-models", "data-generation"], "updated_at": "2024-08-08"}, "text": "Tell Me What's Next: Textual Foresight for Generic UI Representations\nResource type: paper\nReview status: curated\nScope: core\n\nTextual Foresight learns UI representations by describing the future screen expected after an action on the current interface. The OpenApp dataset supports pretraining and downstream evaluation of representations that connect individual elements with whole-screen changes.\nPlatforms: unspecified\nTopics: screen-understanding, world-models, data-generation\nPublished: 2024-06-12\nAuthors: Andrea Burns, Kate Saenko, Bryan A. Plummer\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.07822v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-07822/"}
{"id": "arxiv:2406.08184", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "637843b1b8634e608bed71fc0f3df205ae218d1b0b5550ce480046c67bb14b0f", "submitted_at": "2024-06-12T13:14:50Z", "updated_at": "2024-06-12T13:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08184v1"}, "authors": ["Luyuan Wang", "Yongyu Deng", "Yiwei Zha", "Guodong Mao", "Qinmin Wang", "Tianchen Min", "Wei Chen", "Shoufa Chen"], "canonical_source_url": "https://arxiv.org/abs/2406.08184", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-08184/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "637843b1b8634e608bed71fc0f3df205ae218d1b0b5550ce480046c67bb14b0f", "submitted_at": "2024-06-12T13:14:50Z", "updated_at": "2024-06-12T13:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08184v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.08184"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08184"}], "name": "MobileAgentBench: An Efficient and User-Friendly Benchmark for Mobile LLM Agents", "publication": null, "published_at": "2024-06-12", "record_sha256": "9c4554631fa51ee99cb113c279e16a1d9baff8906bcd07c6366ce52aefee71d9", "relations": [], "resource_id": "arxiv:2406.08184", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.08184v1", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2024-06-12"}, "text": "MobileAgentBench: An Efficient and User-Friendly Benchmark for Mobile LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a mobile-agent benchmark with tasks across open-source applications, organized by difficulty to reduce manual evaluation effort.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation\nPublished: 2024-06-12\nAuthors: Luyuan Wang, Yongyu Deng, Yiwei Zha, Guodong Mao, Qinmin Wang, Tianchen Min, Wei Chen, Shoufa Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.08184v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-08184/"}
{"id": "arxiv:2406.08340", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0061e9004afe11bf188164ea37de1acd15e1b95ae2fcad22044264a89e4636b3", "submitted_at": "2024-06-12T15:48:39Z", "updated_at": "2024-06-12T15:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08340v1"}, "authors": ["Shengcheng Yu", "Chunrong Fang", "Mingzhe Du", "Zimin Ding", "Zhenyu Chen", "Zhendong Su"], "canonical_source_url": "https://arxiv.org/abs/2406.08340", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-08340/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0061e9004afe11bf188164ea37de1acd15e1b95ae2fcad22044264a89e4636b3", "submitted_at": "2024-06-12T15:48:39Z", "updated_at": "2024-06-12T15:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.08340v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.08340"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08340"}], "name": "Practical, Automated Scenario-based Mobile App Testing", "publication": null, "published_at": "2024-06-12", "record_sha256": "34d03c2243e1cddea9982b1aeb4a9cab37100323980cf428bbc7ba2f20290bf6", "relations": [], "resource_id": "arxiv:2406.08340", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.08340v1", "topics": ["automation", "cross-app", "screen-understanding", "evaluation"], "updated_at": "2024-06-12"}, "text": "Practical, Automated Scenario-based Mobile App Testing\nResource type: paper\nReview status: curated\nScope: core\n\nScenTest builds scenario-specific event knowledge graphs from crowdsourced test reports containing GUI and action information. It uses the current app state and testing context to generate scenario tests across applications.\nPlatforms: unspecified\nTopics: automation, cross-app, screen-understanding, evaluation\nPublished: 2024-06-12\nAuthors: Shengcheng Yu, Chunrong Fang, Mingzhe Du, Zimin Ding, Zhenyu Chen, Zhendong Su\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.08340v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-08340/"}
{"id": "arxiv:2406.08451", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b45039151353cc0fede71e2e9ba109ec53d0c1f495037a82d8fb742c9e97bbb", "submitted_at": "2024-06-12T17:44:26Z", "updated_at": "2025-08-01T03:37:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.08451v2"}, "authors": ["Quanfeng Lu", "Wenqi Shao", "Zitao Liu", "Lingxiao Du", "Fanqing Meng", "Boxuan Li", "Botong Chen", "Siyuan Huang", "Kaipeng Zhang", "Ping Luo"], "canonical_source_url": "https://arxiv.org/abs/2406.08451", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-08451/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b45039151353cc0fede71e2e9ba109ec53d0c1f495037a82d8fb742c9e97bbb", "submitted_at": "2024-06-12T17:44:26Z", "updated_at": "2025-08-01T03:37:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.08451v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.08451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.08451"}], "name": "GUIOdyssey: A Comprehensive Dataset for Cross-App GUI Navigation on Mobile Devices", "publication": null, "published_at": "2024-06-12", "record_sha256": "48c88d6b05a4665bedd53a1f208b98790389601a4ddd87d4fbab69fae18bb493", "relations": [], "resource_id": "arxiv:2406.08451", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.08451v2", "topics": ["cross-app", "data-generation"], "updated_at": "2025-08-01"}, "text": "GUIOdyssey: A Comprehensive Dataset for Cross-App GUI Navigation on Mobile Devices\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces annotated cross-application mobile trajectories and studies history-aware navigation over long task sequences.\nPlatforms: android\nTopics: cross-app, data-generation\nPublished: 2024-06-12\nAuthors: Quanfeng Lu, Wenqi Shao, Zitao Liu, Lingxiao Du, Fanqing Meng, Boxuan Li, Botong Chen, Siyuan Huang, Kaipeng Zhang, Ping Luo\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.08451v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-08451/"}
{"id": "arxiv:2406.09187", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50a8ea492d3f1c2ff2bd00ebfb18e19077bacc96b44c751d985bfe0f7bef0247", "submitted_at": "2024-06-13T14:49:26Z", "updated_at": "2025-05-29T03:09:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.09187v3"}, "authors": ["Zhen Xiang", "Linzhi Zheng", "Yanjie Li", "Junyuan Hong", "Qinbin Li", "Han Xie", "Jiawei Zhang", "Zidi Xiong", "Chulin Xie", "Carl Yang", "Dawn Song", "Bo Li"], "canonical_source_url": "https://arxiv.org/abs/2406.09187", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-09187/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50a8ea492d3f1c2ff2bd00ebfb18e19077bacc96b44c751d985bfe0f7bef0247", "submitted_at": "2024-06-13T14:49:26Z", "updated_at": "2025-05-29T03:09:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.09187v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2406.09187"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.09187"}], "name": "GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning", "publication": null, "published_at": "2024-06-13", "record_sha256": "cef1230fcac8463f88ca918b93b01f463ff7a5c68722583efe67aacd1cb11b33", "relations": [], "resource_id": "arxiv:2406.09187", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2406.09187v3", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2025-05-29"}, "text": "GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces GuardAgent, which translates safety requests into executable guardrail code using plans and retrieved demonstrations. Mind2Web-SC supplies a web-agent safety-policy evaluation alongside a separate healthcare access-control benchmark.\nPlatforms: web\nTopics: safety-security, evaluation, tool-integration\nPublished: 2024-06-13\nAuthors: Zhen Xiang, Linzhi Zheng, Yanjie Li, Junyuan Hong, Qinbin Li, Han Xie, Jiawei Zhang, Zidi Xiong, Chulin Xie, Carl Yang, Dawn Song, Bo Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The web-agent evidence is Mind2Web-SC; the healthcare benchmark does not independently establish GUI control.\n\nPrimary source: https://arxiv.org/abs/2406.09187v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-09187/"}
{"id": "arxiv:2406.10227", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66b5fb1e839490df80e9750fdc049f7fe3d2d4686c4090d78ea01bc542fe55b0", "submitted_at": "2024-06-14T17:59:08Z", "updated_at": "2024-06-14T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.10227v1"}, "authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Qinchen WU", "Mingyi Yan", "Zhengyuan Yang", "Lijuan Wang", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2406.10227", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-10227/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66b5fb1e839490df80e9750fdc049f7fe3d2d4686c4090d78ea01bc542fe55b0", "submitted_at": "2024-06-14T17:59:08Z", "updated_at": "2024-06-14T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.10227v1"}, "publication": {"authors": ["Kevin Q. Lin", "Linjie Li", "Difei Gao", "Qinchen Wu", "Mingyi Yan", "Zhengyuan Yang", "Lijuan Wang", "Mike Z. Shou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2214", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "69329-69360", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/804e757b7d7043c26701c3a313032101-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "5731b34872eb84e83338aa851d7f3922e304162869753bf6100ffdc1183c96cf", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.10227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.10227"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/804e757b7d7043c26701c3a313032101-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos", "publication": {"authors": ["Kevin Q. Lin", "Linjie Li", "Difei Gao", "Qinchen Wu", "Mingyi Yan", "Zhengyuan Yang", "Lijuan Wang", "Mike Z. Shou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2214", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "69329-69360", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/804e757b7d7043c26701c3a313032101-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-06-14", "record_sha256": "f7b5c55082df7ee5542f378c6fa42b7c382322bf38dd285695994d89dad74608", "relations": [], "resource_id": "arxiv:2406.10227", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.10227v1", "topics": ["evaluation", "video", "planning"], "updated_at": "2024-06-14"}, "text": "VideoGUI: A Benchmark for GUI Automation from Instructional Videos\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates GUI automation from instructional videos at separate levels of high-level planning, action narration, and primitive execution.\nPlatforms: desktop, web\nTopics: evaluation, video, planning\nPublished: 2024-06-14\nAuthors: Kevin Qinghong Lin, Linjie Li, Difei Gao, Qinchen WU, Mingyi Yan, Zhengyuan Yang, Lijuan Wang, Mike Zheng Shou\nRecorded arXiv version: v1\n\nPublication edition: VideoGUI: A Benchmark for GUI Automation from Instructional Videos\nPublication authors: Kevin Q. Lin, Linjie Li, Difei Gao, Qinchen Wu, Mingyi Yan, Zhengyuan Yang, Lijuan Wang, Mike Z. Shou\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 69329-69360\nPublication doi: 10.52202/079017-2214\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/804e757b7d7043c26701c3a313032101-Abstract-Datasets_and_Benchmarks_Track.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.10227v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-10227/"}
{"id": "arxiv:2406.10819", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9997c3ab43f0e3e69ad4dcbe15257ea4392d36fa3c82f90c6b7730667a067f99", "submitted_at": "2024-06-16T06:56:53Z", "updated_at": "2025-03-24T11:46:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.10819v2"}, "authors": ["Dongping Chen", "Yue Huang", "Siyuan Wu", "Jingyu Tang", "Liuyi Chen", "Yilin Bai", "Zhigang He", "Chenlong Wang", "Huichi Zhou", "Yiqiang Li", "Tianshuo Zhou", "Yue Yu", "Chujie Gao", "Qihui Zhang", "Yi Gui", "Zhen Li", "Yao Wan", "Pan Zhou", "Jianfeng Gao", "Lichao Sun"], "canonical_source_url": "https://arxiv.org/abs/2406.10819", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-10819/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9997c3ab43f0e3e69ad4dcbe15257ea4392d36fa3c82f90c6b7730667a067f99", "submitted_at": "2024-06-16T06:56:53Z", "updated_at": "2025-03-24T11:46:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.10819v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.10819"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.10819"}], "name": "GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding", "publication": null, "published_at": "2024-06-16", "record_sha256": "f5eac3e7dfb44279a6a441e6a74be7754da5e66772dd7536a2cb646ca9a336d6", "relations": [], "resource_id": "arxiv:2406.10819", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.10819v2", "topics": ["evaluation", "video", "screen-understanding"], "updated_at": "2025-03-24"}, "text": "GUI-World: A Video Benchmark and Dataset for Multimodal GUI-oriented Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces video-based GUI questions and annotations for studying temporal interface understanding across varied interaction settings.\nPlatforms: cross-platform\nTopics: evaluation, video, screen-understanding\nPublished: 2024-06-16\nAuthors: Dongping Chen, Yue Huang, Siyuan Wu, Jingyu Tang, Liuyi Chen, Yilin Bai, Zhigang He, Chenlong Wang, Huichi Zhou, Yiqiang Li, Tianshuo Zhou, Yue Yu, Chujie Gao, Qihui Zhang, Yi Gui, Zhen Li, Yao Wan, Pan Zhou, Jianfeng Gao, Lichao Sun\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.10819v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-10819/"}
{"id": "arxiv:2406.11317", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e41704dbe59f1fb9b1e743a3c13394105de112213f666c2420fb235d666ef3", "submitted_at": "2024-06-17T08:30:55Z", "updated_at": "2025-05-30T07:30:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.11317v2"}, "authors": ["Wentong Chen", "Junbo Cui", "Jinyi Hu", "Yujia Qin", "Junjie Fang", "Yue Zhao", "Chongyi Wang", "Jun Liu", "Guirong Chen", "Yupeng Huo", "Yuan Yao", "Yankai Lin", "Zhiyuan Liu", "Maosong Sun"], "canonical_source_url": "https://arxiv.org/abs/2406.11317", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-11317/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e41704dbe59f1fb9b1e743a3c13394105de112213f666c2420fb235d666ef3", "submitted_at": "2024-06-17T08:30:55Z", "updated_at": "2025-05-30T07:30:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.11317v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2406.11317"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.11317"}], "name": "GUICourse: From General Vision Language Models to Versatile GUI Agents", "publication": null, "published_at": "2024-06-17", "record_sha256": "235c93f399bdf30451041e4294a85342f1d380c1d7c216a11058a25fd39d4102", "relations": [], "resource_id": "arxiv:2406.11317", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.11317v2", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-05-30"}, "text": "GUICourse: From General Vision Language Models to Versatile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUIEnv, GUIAct, and GUIChat datasets to develop OCR, grounding, component knowledge, and interaction abilities in visual-language models.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\nPublished: 2024-06-17\nAuthors: Wentong Chen, Junbo Cui, Jinyi Hu, Yujia Qin, Junjie Fang, Yue Zhao, Chongyi Wang, Jun Liu, Guirong Chen, Yupeng Huo, Yuan Yao, Yankai Lin, Zhiyuan Liu, Maosong Sun\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.11317v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-11317/"}
{"id": "arxiv:2406.11896", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c5d4e75d6ba88574f6f033bb2c1cf44cf04149e94e017cd8a1a227e7a1c8238", "submitted_at": "2024-06-14T17:49:55Z", "updated_at": "2024-06-14T17:49:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.11896v1"}, "authors": ["Hao Bai", "Yifei Zhou", "Mert Cemri", "Jiayi Pan", "Alane Suhr", "Sergey Levine", "Aviral Kumar"], "canonical_source_url": "https://arxiv.org/abs/2406.11896", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-11896/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c5d4e75d6ba88574f6f033bb2c1cf44cf04149e94e017cd8a1a227e7a1c8238", "submitted_at": "2024-06-14T17:49:55Z", "updated_at": "2024-06-14T17:49:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.11896v1"}, "publication": {"authors": ["Hao Bai", "Yifei Zhou", "Mert Cemri", "Jiayi Pan", "Alane Suhr", "Sergey Levine", "Aviral Kumar"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0397", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "12461-12495", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/1704ddd0bb89f159dfe609b32c889995-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:03Z", "role": "primary", "sha256": "0fd68a0f088591eb695e205286823c7be4059fba8a14e62030ca1b5b11707e21", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {"arxiv": "2406.11896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.11896"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/1704ddd0bb89f159dfe609b32c889995-Paper-Conference.pdf"}], "name": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning", "publication": {"authors": ["Hao Bai", "Yifei Zhou", "Mert Cemri", "Jiayi Pan", "Alane Suhr", "Sergey Levine", "Aviral Kumar"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0397", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "12461-12495", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/1704ddd0bb89f159dfe609b32c889995-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-06-14", "record_sha256": "bf358cc4588b92d0cc8582c9b2a1fcc7f1869083da1ad882dd5cf8a24cbcb0af", "relations": [], "resource_id": "arxiv:2406.11896", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.11896v1", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2024-06-14"}, "text": "DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nCombines offline initialization and online reinforcement learning in Android environments, with automated evaluation and curriculum-based sampling.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation\nPublished: 2024-06-14\nAuthors: Hao Bai, Yifei Zhou, Mert Cemri, Jiayi Pan, Alane Suhr, Sergey Levine, Aviral Kumar\nRecorded arXiv version: v1\n\nPublication edition: DigiRL: Training In-The-Wild Device-Control Agents with Autonomous Reinforcement Learning\nPublication authors: Hao Bai, Yifei Zhou, Mert Cemri, Jiayi Pan, Alane Suhr, Sergey Levine, Aviral Kumar\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 12461-12495\nPublication doi: 10.52202/079017-0397\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/1704ddd0bb89f159dfe609b32c889995-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.11896v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-11896/"}
{"id": "arxiv:2406.12373", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5d0fbea7a4813a1da85c9ecb237c2115cc183b4c9e6fd0cbfd63764b0d418e3", "submitted_at": "2024-06-18T07:58:33Z", "updated_at": "2024-07-16T06:19:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12373v3"}, "authors": ["Yichen Pan", "Dehan Kong", "Sida Zhou", "Cheng Cui", "Yifei Leng", "Bing Jiang", "Hangyu Liu", "Yanyi Shang", "Shuyan Zhou", "Tongshuang Wu", "Zhengyang Wu"], "canonical_source_url": "https://arxiv.org/abs/2406.12373", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-12373/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5d0fbea7a4813a1da85c9ecb237c2115cc183b4c9e6fd0cbfd63764b0d418e3", "submitted_at": "2024-06-18T07:58:33Z", "updated_at": "2024-07-16T06:19:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12373v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2406.12373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.12373"}], "name": "WebCanvas: Benchmarking Web Agents in Online Environments", "publication": null, "published_at": "2024-06-18", "record_sha256": "d2b80d2079fba88f46d5604e92ebbf2b9b50ea5172cd56fd7e44f7aa72575806", "relations": [], "resource_id": "arxiv:2406.12373", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.12373v3", "topics": ["evaluation", "web-navigation"], "updated_at": "2024-07-16"}, "text": "WebCanvas: Benchmarking Web Agents in Online Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebCanvas and Mind2Web-Live, evaluating online task progress through meaningful intermediate states on changing websites.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2024-06-18\nAuthors: Yichen Pan, Dehan Kong, Sida Zhou, Cheng Cui, Yifei Leng, Bing Jiang, Hangyu Liu, Yanyi Shang, Shuyan Zhou, Tongshuang Wu, Zhengyang Wu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.12373v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-12373/"}
{"id": "arxiv:2406.12814", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc1c1ac93f8e328d33accf7cdaa1d65487c4f7abec15234b9e596a501947550", "submitted_at": "2024-06-18T17:32:48Z", "updated_at": "2025-02-04T20:02:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12814v3"}, "authors": ["Chen Henry Wu", "Rishi Shah", "Jing Yu Koh", "Ruslan Salakhutdinov", "Daniel Fried", "Aditi Raghunathan"], "canonical_source_url": "https://arxiv.org/abs/2406.12814", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-12814/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc1c1ac93f8e328d33accf7cdaa1d65487c4f7abec15234b9e596a501947550", "submitted_at": "2024-06-18T17:32:48Z", "updated_at": "2025-02-04T20:02:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.12814v3"}, "publication": {"authors": ["Chen Wu", "Rishi Shah", "Jing Yu Koh", "Russ Salakhutdinov", "Daniel Fried", "Aditi Raghunathan"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "28362-28383", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/460a1d8eac34125dad453b28d6d64446-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Dissecting Adversarial Robustness of Multimodal LM Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "fa23fda1b80dac019741737fc6f611fccf9234a4d73ffbb8e24e54d8caf4dac3", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.12814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.12814"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/460a1d8eac34125dad453b28d6d64446-Paper-Conference.pdf"}], "name": "Dissecting Adversarial Robustness of Multimodal LM Agents", "publication": {"authors": ["Chen Wu", "Rishi Shah", "Jing Yu Koh", "Russ Salakhutdinov", "Daniel Fried", "Aditi Raghunathan"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "28362-28383", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/460a1d8eac34125dad453b28d6d64446-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Dissecting Adversarial Robustness of Multimodal LM Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-06-18", "record_sha256": "8435a688d793d3f873830f97713182b6529132d08380edf05bfec5bebb21cac4", "relations": [], "resource_id": "arxiv:2406.12814", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.12814v3", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-02-04"}, "text": "Dissecting Adversarial Robustness of Multimodal LM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces adversarial tasks on VisualWebArena and an evaluation framework that tracks how manipulated information passes between agent components. The analysis examines vulnerabilities introduced by reflection and tree-search evaluators as well as direct action generation.\nPlatforms: web\nTopics: safety-security, robustness, evaluation\nPublished: 2024-06-18\nAuthors: Chen Henry Wu, Rishi Shah, Jing Yu Koh, Ruslan Salakhutdinov, Daniel Fried, Aditi Raghunathan\nRecorded arXiv version: v3\n\nPublication edition: Dissecting Adversarial Robustness of Multimodal LM Agents\nPublication authors: Chen Wu, Rishi Shah, Jing Yu Koh, Russ Salakhutdinov, Daniel Fried, Aditi Raghunathan\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 28362-28383\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/460a1d8eac34125dad453b28d6d64446-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.12814v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-12814/"}
{"id": "arxiv:2406.13264", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e7d00030ea8c77b9ed6f950d62f2bd417d3870f24b2f4589d8068c9484e328", "submitted_at": "2024-06-19T06:50:15Z", "updated_at": "2024-10-11T00:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.13264v2"}, "authors": ["Michael Wornow", "Avanika Narayan", "Ben Viggiano", "Ishan S. Khare", "Tathagat Verma", "Tibor Thompson", "Miguel Angel Fuentes Hernandez", "Sudharsan Sundar", "Chloe Trujillo", "Krrish Chawla", "Rongfei Lu", "Justin Shen", "Divya Nagaraj", "Joshua Martinez", "Vardhan Agrawal", "Althea Hudson", "Nigam H. Shah", "Christopher Re"], "canonical_source_url": "https://arxiv.org/abs/2406.13264", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-13264/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e7d00030ea8c77b9ed6f950d62f2bd417d3870f24b2f4589d8068c9484e328", "submitted_at": "2024-06-19T06:50:15Z", "updated_at": "2024-10-11T00:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.13264v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:37Z", "role": "primary", "sha256": "a75a630c4d4403233df3cb296d7e6bb60ca2f30705b5fc7ed4bd53b188c394ca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.10290%2C2406.10707%2C2406.11342%2C2406.11775%2C2406.12045%2C2406.12275%2C2406.12934%2C2406.13264%2C2406.13352%2C2406.14884%2C2406.14991%2C2406.18820%2C2406.20041%2C2406.20092%2C2407.01502%2C2407.03502%2C2407.04363%2C2407.04856%2C2407.06048%2C2407.08898%2C2407.09025%2C2407.10839%2C2407.12821%2C2407.13696%2C2407.17789%2C2407.20143%2C2407.20727%2C2408.00545%2C2408.02248%2C2408.04168%2C2408.07720%2C2408.09559%2C2408.10215%2C2408.10605%2C2408.11791%2C2408.16119%2C2409.00608%2C2409.03402%2C2409.04999%2C2409.05655%2C2409.07703%2C2409.08273%2C2409.08831%2C2409.10033%2C2409.17652%2C2409.19715%2C2410.01039%2C2410.01606%2C2410.05340%2C2410.05656&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:30Z", "role": "primary", "sha256": "95c08b3a28f77a248d6f990345ef4b2b16cbf196492920bfa1ce0ec7e398345e", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://arxiv.org/abs/2406.13264v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:16:02Z", "role": "primary", "sha256": "07de7eac76877364a2d9e9d31f6c65f5f2cace2f97f7821862a1441f83d9f95c", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2406.13264v2"}], "identifiers": {"arxiv": "2406.13264"}, "kind": "paper", "last_verified_at": "2026-10-17T09:15:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.13264"}], "name": "WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks", "publication": null, "published_at": "2024-06-19", "record_sha256": "db03d09d8e060f3c4f26195fac19c176a61a528102c51716f6e48e713a3a77d4", "relations": [], "resource_id": "arxiv:2406.13264", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.13264v2", "topics": ["evaluation", "workflow", "video"], "updated_at": "2024-10-11"}, "text": "WONDERBREAD: A Benchmark for Evaluating Multimodal Foundation Models on Business Process Management Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WONDERBREAD, a benchmark pairing human WebArena workflow demonstrations with screen recordings, action logs, and written procedures. Six evaluation tasks cover documentation, knowledge transfer, and process improvement beyond end-to-end execution.\nPlatforms: web\nTopics: evaluation, workflow, video\nPublished: 2024-06-19\nAuthors: Michael Wornow, Avanika Narayan, Ben Viggiano, Ishan S. Khare, Tathagat Verma, Tibor Thompson, Miguel Angel Fuentes Hernandez, Sudharsan Sundar, Chloe Trujillo, Krrish Chawla, Rongfei Lu, Justin Shen, Divya Nagaraj, Joshua Martinez, Vardhan Agrawal, Althea Hudson, Nigam H. Shah, Christopher Re\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. The demonstrations were rehearsed, so their completion rate is not a zero-shot human performance estimate.\n\nPrimary source: https://arxiv.org/abs/2406.13264v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-13264/"}
{"id": "arxiv:2406.13719", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "369c32dba4bab0cfc05b41db16c73d5672605b7f81caabb2dbd72f581b3290ec", "submitted_at": "2024-06-19T17:22:11Z", "updated_at": "2024-06-19T17:22:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.13719v1"}, "authors": ["Qinchen Wu", "Difei Gao", "Kevin Qinghong Lin", "Zhuoyu Wu", "Xiangwu Guo", "Peiran Li", "Weichen Zhang", "Hengxu Wang", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2406.13719", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-13719/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "369c32dba4bab0cfc05b41db16c73d5672605b7f81caabb2dbd72f581b3290ec", "submitted_at": "2024-06-19T17:22:11Z", "updated_at": "2024-06-19T17:22:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2406.13719v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.13719"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.13719"}], "name": "GUI Action Narrator: Where and When Did That Action Take Place?", "publication": null, "published_at": "2024-06-19", "record_sha256": "3516d6ab3cb6397e341b83d58493fb3236bb21d80dddf16cc916aeb2ba4fae48", "relations": [], "resource_id": "arxiv:2406.13719", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.13719v1", "topics": ["video", "screen-understanding"], "updated_at": "2024-06-19"}, "text": "GUI Action Narrator: Where and When Did That Action Take Place?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Act2Cap and a cursor-guided method for describing GUI actions from video, with temporal and spatial localization challenges.\nPlatforms: cross-platform\nTopics: video, screen-understanding\nPublished: 2024-06-19\nAuthors: Qinchen Wu, Difei Gao, Kevin Qinghong Lin, Zhuoyu Wu, Xiangwu Guo, Peiran Li, Weichen Zhang, Hengxu Wang, Mike Zheng Shou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.13719v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-13719/"}
{"id": "arxiv:2406.14056", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea674b3aec3aee51c6f7fd4e99ad2b88394e7df43a4f3a02654d805e9a60343a", "submitted_at": "2024-06-20T07:24:43Z", "updated_at": "2024-11-04T09:31:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14056v3"}, "authors": ["Ziyang Meng", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "canonical_source_url": "https://arxiv.org/abs/2406.14056", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-14056/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea674b3aec3aee51c6f7fd4e99ad2b88394e7df43a4f3a02654d805e9a60343a", "submitted_at": "2024-06-20T07:24:43Z", "updated_at": "2024-11-04T09:31:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14056v3"}, "publication": {"authors": ["Meng Ziyang", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.68", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.68/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1261-1279", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.68.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VGA: Vision GUI Assistant - Minimizing Hallucinations through Image-Centric Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.68/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "7202817c9f68a7c61afa104997fffe2a1a1112977561309ff5811f722b9411e8", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.14056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14056"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.68/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.68.pdf"}], "name": "VGA: Vision GUI Assistant -- Minimizing Hallucinations through Image-Centric Fine-Tuning", "publication": {"authors": ["Meng Ziyang", "Yu Dai", "Zezheng Gong", "Shaoxiong Guo", "Minglong Tang", "Tongquan Wei"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.68", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.68/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1261-1279", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.68.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "VGA: Vision GUI Assistant - Minimizing Hallucinations through Image-Centric Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.68/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}, "published_at": "2024-06-20", "record_sha256": "6f39fd128da6dc28f0b27111d78d73ab6e9b854b3e55383726da6836e4d21420", "relations": [], "resource_id": "arxiv:2406.14056", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.14056v3", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-11-04"}, "text": "VGA: Vision GUI Assistant -- Minimizing Hallucinations through Image-Centric Fine-Tuning\nResource type: paper\nReview status: curated\nScope: core\n\nUses image-dependent question-answer data and staged fine-tuning to improve GUI comprehension and reduce unsupported visual answers.\nPlatforms: cross-platform\nTopics: screen-understanding, grounding\nPublished: 2024-06-20\nAuthors: Ziyang Meng, Yu Dai, Zezheng Gong, Shaoxiong Guo, Minglong Tang, Tongquan Wei\nRecorded arXiv version: v3\n\nPublication edition: VGA: Vision GUI Assistant - Minimizing Hallucinations through Image-Centric Fine-Tuning\nPublication authors: Meng Ziyang, Yu Dai, Zezheng Gong, Shaoxiong Guo, Minglong Tang, Tongquan Wei\nPublication edition date (recorded precision): 2024-11\nPublication venue: Findings of the Association for Computational Linguistics: EMNLP 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 1261-1279\nPublication doi: 10.18653/v1/2024.findings-emnlp.68\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2024.findings-emnlp.68/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.14056v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-14056/"}
{"id": "arxiv:2406.14250", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4d5a46d2141ca59fcf221bbe618601cb7cce29c1b067ae60712e53e907095bd", "submitted_at": "2024-06-20T12:22:05Z", "updated_at": "2024-07-01T08:37:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14250v3"}, "authors": ["Ke Wang", "Tianyu Xia", "Zhangxuan Gu", "Yi Zhao", "Shuheng Shen", "Changhua Meng", "Weiqiang Wang", "Ke Xu"], "canonical_source_url": "https://arxiv.org/abs/2406.14250", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-14250/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4d5a46d2141ca59fcf221bbe618601cb7cce29c1b067ae60712e53e907095bd", "submitted_at": "2024-06-20T12:22:05Z", "updated_at": "2024-07-01T08:37:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14250v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.14250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14250"}], "name": "E-ANT: A Large-Scale Dataset for Efficient Automatic GUI NavigaTion", "publication": null, "published_at": "2024-06-20", "record_sha256": "b2849614c57db9b916b667b78df69a622415f6a827400457b094ac5eacd90fe0", "relations": [], "resource_id": "arxiv:2406.14250", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.14250v3", "topics": ["mobile-navigation", "evaluation"], "updated_at": "2024-07-01"}, "text": "E-ANT: A Large-Scale Dataset for Efficient Automatic GUI NavigaTion\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Chinese-language GUI navigation traces and annotated screenshots collected from human interactions with mini-applications.\nPlatforms: unspecified\nTopics: mobile-navigation, evaluation\nPublished: 2024-06-20\nAuthors: Ke Wang, Tianyu Xia, Zhangxuan Gu, Yi Zhao, Shuheng Shen, Changhua Meng, Weiqiang Wang, Ke Xu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.14250v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-14250/"}
{"id": "arxiv:2406.14314", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0bc798b69e59d1ec0ca8ecb0fab7a7ee6594f4d656ca4212b19be7e3085c0c5", "submitted_at": "2024-06-20T13:46:10Z", "updated_at": "2025-03-03T15:47:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14314v3"}, "authors": ["Omri Berkovitch", "Sapir Caduri", "Noam Kahlon", "Anatoly Efros", "Avi Caciularu", "Ido Dagan"], "canonical_source_url": "https://arxiv.org/abs/2406.14314", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-14314/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0bc798b69e59d1ec0ca8ecb0fab7a7ee6594f4d656ca4212b19be7e3085c0c5", "submitted_at": "2024-06-20T13:46:10Z", "updated_at": "2025-03-03T15:47:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2406.14314v3"}}, "environments": ["android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.14314"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.14314"}], "name": "Identifying User Goals from UI Trajectories", "publication": null, "published_at": "2024-06-20", "record_sha256": "f5ac421ac57b335bde39428c5ec20af99d3aa139ed97b50d7e1aad1552fab25d", "relations": [], "resource_id": "arxiv:2406.14314", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.14314v3", "topics": ["evaluation", "interaction-history", "screen-understanding"], "updated_at": "2025-03-03"}, "text": "Identifying User Goals from UI Trajectories\nResource type: paper\nReview status: curated\nScope: core\n\nDefines goal identification from observed UI trajectories and an evaluation method for comparing intent descriptions within the same interface context. Experiments reuse Android and web automation data to compare human and model inference of users’ detailed goals.\nPlatforms: android, web, cross-platform\nTopics: evaluation, interaction-history, screen-understanding\nPublished: 2024-06-20\nAuthors: Omri Berkovitch, Sapir Caduri, Noam Kahlon, Anatoly Efros, Avi Caciularu, Ido Dagan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.14314v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-14314/"}
{"id": "arxiv:2406.19263", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c86d68b5c3486b099f64792d5d3290d79e25b7c82c6eb2402869b893bb22fea", "submitted_at": "2024-06-27T15:34:16Z", "updated_at": "2024-10-25T18:16:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.19263v2"}, "authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2406.19263", "citation_url": "https://awesomecua.github.io/resources/arxiv-2406-19263/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c86d68b5c3486b099f64792d5d3290d79e25b7c82c6eb2402869b893bb22fea", "submitted_at": "2024-06-27T15:34:16Z", "updated_at": "2024-10-25T18:16:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2406.19263v2"}, "publication": {"authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.533", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.533/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "9503-9522", "pdf_url": "https://aclanthology.org/2024.emnlp-main.533.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.533/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "1d105799f5db853e801090e8ed62273cd72dd3cb9b4060f901747aace8d2406d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2406.19263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2406.19263"}, {"rel": "publication", "url": "https://aclanthology.org/2024.emnlp-main.533/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.emnlp-main.533.pdf"}], "name": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "publication": {"authors": ["Yue Fan", "Lei Ding", "Ching-Chen Kuo", "Shan Jiang", "Yang Zhao", "Xinze Guan", "Jie Yang", "Yi Zhang", "Xin Eric Wang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.533", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.533/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "9503-9522", "pdf_url": "https://aclanthology.org/2024.emnlp-main.533.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.533/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}, "published_at": "2024-06-27", "record_sha256": "123bea303eb3b13601e548ff5e78fa499e3cc6b14c594b24f59f88db98991b41", "relations": [], "resource_id": "arxiv:2406.19263", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2406.19263v2", "topics": ["screen-understanding", "grounding"], "updated_at": "2024-10-25"}, "text": "Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Screen Point-and-Read and a hierarchical layout approach that interprets screen regions indicated by a user-selected point.\nPlatforms: cross-platform\nTopics: screen-understanding, grounding\nPublished: 2024-06-27\nAuthors: Yue Fan, Lei Ding, Ching-Chen Kuo, Shan Jiang, Yang Zhao, Xinze Guan, Jie Yang, Yi Zhang, Xin Eric Wang\nRecorded arXiv version: v2\n\nPublication edition: Read Anywhere Pointed: Layout-aware GUI Screen Reading with Tree-of-Lens Grounding\nPublication authors: Yue Fan, Lei Ding, Ching-Chen Kuo, Shan Jiang, Yang Zhao, Xinze Guan, Jie Yang, Yi Zhang, Xin Eric Wang\nPublication edition date (recorded precision): 2024-11\nPublication venue: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing\nPublication publisher: Association for Computational Linguistics\nPublication pages: 9503-9522\nPublication doi: 10.18653/v1/2024.emnlp-main.533\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2024.emnlp-main.533/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2406.19263v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2406-19263/"}
{"id": "arxiv:2407.00993", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "90ce9900644c21004e0ba9e749e10d94238d2447d0e84bd7205ff55dcbd243f4", "submitted_at": "2024-07-01T06:10:01Z", "updated_at": "2024-07-01T06:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.00993v1"}, "authors": ["Shihan Deng", "Weikai Xu", "Hongda Sun", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Rui Yan", "Shuo Shang"], "canonical_source_url": "https://arxiv.org/abs/2407.00993", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-00993/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "90ce9900644c21004e0ba9e749e10d94238d2447d0e84bd7205ff55dcbd243f4", "submitted_at": "2024-07-01T06:10:01Z", "updated_at": "2024-07-01T06:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.00993v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.00993"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.00993"}], "name": "Mobile-Bench: An Evaluation Benchmark for LLM-based Mobile Agents", "publication": null, "published_at": "2024-07-01", "record_sha256": "d18f4d19ca38fb3bc56522a869d5328f781677d7a3965337b8dc14b2334768dd", "relations": [], "resource_id": "arxiv:2407.00993", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.00993v1", "topics": ["evaluation", "cross-app", "tool-integration"], "updated_at": "2024-07-01"}, "text": "Mobile-Bench: An Evaluation Benchmark for LLM-based Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates mobile agents using combined UI and API actions, tasks with varying app scope, and checkpoints for intermediate progress.\nPlatforms: unspecified\nTopics: evaluation, cross-app, tool-integration\nPublished: 2024-07-01\nAuthors: Shihan Deng, Weikai Xu, Hongda Sun, Wei Liu, Tao Tan, Jianfeng Liu, Ang Li, Jian Luan, Bin Wang, Rui Yan, Shuo Shang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.00993v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-00993/"}
{"id": "arxiv:2407.01476", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec7aaf2a34e38f72da652bb2d8350c951229c154e2bc368a927a79287b36c439", "submitted_at": "2024-07-01T17:07:55Z", "updated_at": "2026-02-08T15:06:40Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01476v4"}, "authors": ["Jing Yu Koh", "Stephen McAleer", "Daniel Fried", "Ruslan Salakhutdinov"], "canonical_source_url": "https://arxiv.org/abs/2407.01476", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-01476/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec7aaf2a34e38f72da652bb2d8350c951229c154e2bc368a927a79287b36c439", "submitted_at": "2024-07-01T17:07:55Z", "updated_at": "2026-02-08T15:06:40Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01476v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.01476"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01476"}], "name": "Tree Search for Language Model Agents", "publication": null, "published_at": "2024-07-01", "record_sha256": "3227fc889f15e9d2843cb55ab7787d4fb2be86295ce689a61b3b29140c36e313", "relations": [], "resource_id": "arxiv:2407.01476", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.01476v4", "topics": ["planning", "web-navigation", "evaluation"], "updated_at": "2026-02-08"}, "text": "Tree Search for Language Model Agents\nResource type: paper\nReview status: curated\nScope: core\n\nApplies best-first search to agent actions within interactive web environments, allowing exploration and multistep planning at inference time. WebArena and VisualWebArena experiments examine search behavior and test-time computation.\nPlatforms: web\nTopics: planning, web-navigation, evaluation\nPublished: 2024-07-01\nAuthors: Jing Yu Koh, Stephen McAleer, Daniel Fried, Ruslan Salakhutdinov\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.01476v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-01476/"}
{"id": "arxiv:2407.01511", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12e5464ea0fc62bd566f19d301b887826b7ff317cfc27c86fa0cda8d3de271d9", "submitted_at": "2024-07-01T17:55:04Z", "updated_at": "2025-07-20T13:42:07Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01511v4"}, "authors": ["Tianqi Xu", "Linyao Chen", "Dai-Jie Wu", "Yanjun Chen", "Zecheng Zhang", "Xiang Yao", "Zhiqiang Xie", "Yongchao Chen", "Shilong Liu", "Bochen Qian", "Anjie Yang", "Zhaoxuan Jin", "Jianbo Deng", "Philip Torr", "Bernard Ghanem", "Guohao Li"], "canonical_source_url": "https://arxiv.org/abs/2407.01511", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-01511/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12e5464ea0fc62bd566f19d301b887826b7ff317cfc27c86fa0cda8d3de271d9", "submitted_at": "2024-07-01T17:55:04Z", "updated_at": "2025-07-20T13:42:07Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2407.01511v4"}, "publication": {"authors": ["Tianqi Xu", "Linyao Chen", "Dai-Jie Wu", "Yanjun Chen", "Zecheng Zhang", "Xiang Yao", "Zhiqiang Xie (谢志强)", "Yongchao Chen", "Shilong Liu", "Bochen Qian", "Anjie Yang", "Zhaoxuan Jin", "Jianbo Deng", "Philip Torr", "Bernard Ghanem", "Guohao Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1113", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1113/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "21607-21647", "pdf_url": "https://aclanthology.org/2025.findings-acl.1113.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1113/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "a8cf178f2faa088fd94b77d03785b8c6e2908db34b411287984066b4e5d1ef3f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.1113/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.01511"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01511"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.1113/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.1113.pdf"}], "name": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents", "publication": {"authors": ["Tianqi Xu", "Linyao Chen", "Dai-Jie Wu", "Yanjun Chen", "Zecheng Zhang", "Xiang Yao", "Zhiqiang Xie (谢志强)", "Yongchao Chen", "Shilong Liu", "Bochen Qian", "Anjie Yang", "Zhaoxuan Jin", "Jianbo Deng", "Philip Torr", "Bernard Ghanem", "Guohao Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1113", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1113/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "21607-21647", "pdf_url": "https://aclanthology.org/2025.findings-acl.1113.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1113/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-07-01", "record_sha256": "bdc3477e9d4967fec5b31ae5f885d3602d5653b78c38db9208ccc18858628e1a", "relations": [], "resource_id": "arxiv:2407.01511", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.01511v4", "topics": ["evaluation", "cross-app"], "updated_at": "2025-07-20"}, "text": "CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CRAB, a framework for cross-environment agent tasks with graph-based evaluation and a benchmark spanning desktop and phone interactions.\nPlatforms: desktop, cross-platform\nTopics: evaluation, cross-app\nPublished: 2024-07-01\nAuthors: Tianqi Xu, Linyao Chen, Dai-Jie Wu, Yanjun Chen, Zecheng Zhang, Xiang Yao, Zhiqiang Xie, Yongchao Chen, Shilong Liu, Bochen Qian, Anjie Yang, Zhaoxuan Jin, Jianbo Deng, Philip Torr, Bernard Ghanem, Guohao Li\nRecorded arXiv version: v4\n\nPublication edition: CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents\nPublication authors: Tianqi Xu, Linyao Chen, Dai-Jie Wu, Yanjun Chen, Zecheng Zhang, Xiang Yao, Zhiqiang Xie (谢志强), Yongchao Chen, Shilong Liu, Bochen Qian, Anjie Yang, Zhaoxuan Jin, Jianbo Deng, Philip Torr, Bernard Ghanem, Guohao Li\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 21607-21647\nPublication doi: 10.18653/v1/2025.findings-acl.1113\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2025.findings-acl.1113/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.01511v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-01511/"}
{"id": "arxiv:2407.01558", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "01d3b3a013f3b10249b8197a8e18aa695bdc20febaa34f07128791f36afdcbac", "submitted_at": "2024-05-05T19:10:19Z", "updated_at": "2025-07-18T11:35:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.01558v3"}, "authors": ["El Hassane Ettifouri", "Jessica López Espejel", "Laura Minkova", "Tassnim Dardouri", "Walid Dahhane"], "canonical_source_url": "https://arxiv.org/abs/2407.01558", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-01558/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "01d3b3a013f3b10249b8197a8e18aa695bdc20febaa34f07128791f36afdcbac", "submitted_at": "2024-05-05T19:10:19Z", "updated_at": "2025-07-18T11:35:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.01558v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.01558"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.01558"}], "name": "Visual Grounding Methods for Efficient Interaction with Desktop Graphical User Interfaces", "publication": null, "published_at": "2024-05-05", "record_sha256": "8d59619b57d6c76c66e2e100dfee2c34a03177db4f1903785d001ef62dc803c3", "relations": [], "resource_id": "arxiv:2407.01558", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.01558v3", "topics": ["grounding", "evaluation"], "updated_at": "2025-07-18"}, "text": "Visual Grounding Methods for Efficient Interaction with Desktop Graphical User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nCompares modular OCR-based and end-to-end methods for locating desktop GUI targets from natural-language instructions.\nPlatforms: desktop\nTopics: grounding, evaluation\nPublished: 2024-05-05\nAuthors: El Hassane Ettifouri, Jessica López Espejel, Laura Minkova, Tassnim Dardouri, Walid Dahhane\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.01558v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-01558/"}
{"id": "arxiv:2407.03037", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b706ac6089a7d19836268437f3fea58d10b9d170ac596ac0d5a12f0d3a9668", "submitted_at": "2024-07-03T11:58:09Z", "updated_at": "2024-12-04T08:00:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.03037v2"}, "authors": ["Zhe Liu", "Cheng Li", "Chunyang Chen", "Junjie Wang", "Mengzhuo Chen", "Boyu Wu", "Yawen Wang", "Jun Hu", "Qing Wang"], "canonical_source_url": "https://arxiv.org/abs/2407.03037", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-03037/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b706ac6089a7d19836268437f3fea58d10b9d170ac596ac0d5a12f0d3a9668", "submitted_at": "2024-07-03T11:58:09Z", "updated_at": "2024-12-04T08:00:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.03037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.03037"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.03037"}], "name": "Seeing is Believing: Vision-driven Non-crash Functional Bug Detection for Mobile Apps", "publication": null, "published_at": "2024-07-03", "record_sha256": "5d3869a41fe177cb0eb43ec9a0937aa43385cf05d6bff73906e05632345816ed", "relations": [], "resource_id": "arxiv:2407.03037", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.03037v2", "topics": ["automation", "evaluation", "screen-understanding", "state-tracking"], "updated_at": "2024-12-04"}, "text": "Seeing is Believing: Vision-driven Non-crash Functional Bug Detection for Mobile Apps\nResource type: paper\nReview status: curated\nScope: core\n\nTrident combines exploration, monitoring, and detection agents to identify functional bugs that do not crash a mobile app. It reasons over screenshot sequences and interface transitions to construct test oracles, evaluated on 590 non-crash bugs and additional live apps.\nPlatforms: unspecified\nTopics: automation, evaluation, screen-understanding, state-tracking\nPublished: 2024-07-03\nAuthors: Zhe Liu, Cheng Li, Chunyang Chen, Junjie Wang, Mengzhuo Chen, Boyu Wu, Yawen Wang, Jun Hu, Qing Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.03037v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-03037/"}
{"id": "arxiv:2407.03913", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa74755c53e20b3a9ff4b72e03a557f455ecceaefc96a50587aefd78ac3c8176", "submitted_at": "2024-07-04T13:12:19Z", "updated_at": "2024-07-04T13:12:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.03913v1"}, "authors": ["Jiayi Zhang", "Chuang Zhao", "Yihan Zhao", "Zhaoyang Yu", "Ming He", "Jianping Fan"], "canonical_source_url": "https://arxiv.org/abs/2407.03913", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-03913/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa74755c53e20b3a9ff4b72e03a557f455ecceaefc96a50587aefd78ac3c8176", "submitted_at": "2024-07-04T13:12:19Z", "updated_at": "2024-07-04T13:12:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.03913v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.03913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.03913"}], "name": "MobileExperts: A Dynamic Tool-Enabled Agent Team in Mobile Devices", "publication": null, "published_at": "2024-07-04", "record_sha256": "4fa00ca134ec115041aaceeb6e509003782845cc7e2254f87609c126dc54851f", "relations": [], "resource_id": "arxiv:2407.03913", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.03913v1", "topics": ["mobile-navigation", "planning", "tool-integration"], "updated_at": "2024-07-04"}, "text": "MobileExperts: A Dynamic Tool-Enabled Agent Team in Mobile Devices\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MobileExperts, which assembles mobile-agent teams, lets agents explore and formulate tools, and coordinates them through two levels of planning. A benchmark organizes tasks by hierarchical intelligence levels.\nPlatforms: unspecified\nTopics: mobile-navigation, planning, tool-integration\nPublished: 2024-07-04\nAuthors: Jiayi Zhang, Chuang Zhao, Yihan Zhao, Zhaoyang Yu, Ming He, Jianping Fan\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.03913v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-03913/"}
{"id": "arxiv:2407.04346", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59708ce4edeba4e913ae890f458b51907dd2f3b0b10bb18c86469b36375c84e4", "submitted_at": "2024-07-05T08:37:10Z", "updated_at": "2024-12-06T07:43:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.04346v3"}, "authors": ["Songqin Nong", "Jiali Zhu", "Rui Wu", "Jiongchao Jin", "Shuo Shan", "Xiutian Huang", "Wenhao Xu"], "canonical_source_url": "https://arxiv.org/abs/2407.04346", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-04346/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "59708ce4edeba4e913ae890f458b51907dd2f3b0b10bb18c86469b36375c84e4", "submitted_at": "2024-07-05T08:37:10Z", "updated_at": "2024-12-06T07:43:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.04346v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2407.04346"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.04346"}], "name": "MobileFlow: A Multimodal LLM For Mobile GUI Agent", "publication": null, "published_at": "2024-07-05", "record_sha256": "d081fe11fcb11a459e2b0a3ccf3adf88b0bc7e185d8ef0a2fdfc5113e8f48cd3", "relations": [], "resource_id": "arxiv:2407.04346", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.04346v3", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": "2024-12-06"}, "text": "MobileFlow: A Multimodal LLM For Mobile GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a mobile GUI model using hybrid visual encoders, mixture-of-experts expansion, and alignment training for multilingual interfaces.\nPlatforms: unspecified\nTopics: mobile-navigation, screen-understanding\nPublished: 2024-07-05\nAuthors: Songqin Nong, Jiali Zhu, Rui Wu, Jiongchao Jin, Shuo Shan, Xiutian Huang, Wenhao Xu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.04346v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-04346/"}
{"id": "arxiv:2407.05291", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b56a6208e4fa3523cae3fbae673affcd5fe3d6a9c19df482d27fe71d43ffef5", "submitted_at": "2024-07-07T07:15:49Z", "updated_at": "2025-02-05T21:50:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.05291v2"}, "authors": ["Léo Boisvert", "Megh Thakkar", "Maxime Gasse", "Massimo Caccia", "Thibault Le Sellier De Chezelles", "Quentin Cappart", "Nicolas Chapados", "Alexandre Lacoste", "Alexandre Drouin"], "canonical_source_url": "https://arxiv.org/abs/2407.05291", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-05291/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b56a6208e4fa3523cae3fbae673affcd5fe3d6a9c19df482d27fe71d43ffef5", "submitted_at": "2024-07-07T07:15:49Z", "updated_at": "2025-02-05T21:50:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.05291v2"}, "publication": {"authors": ["Léo Boisvert", "Megh Thakkar", "Maxime Gasse", "Massimo Caccia", "Thibault L. De Chezelles", "Quentin Cappart", "Nicolas Chapados", "Alexandre Lacoste", "Alexandre Drouin"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0195", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "5996-6051", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/0b82662b6c32e887bb252a74d8cb2d5e-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "4fe01789f93df840699d604de87c068557fc02aa78d22d53c6b3ae7c9db69cc4", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.05291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.05291"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/0b82662b6c32e887bb252a74d8cb2d5e-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks", "publication": {"authors": ["Léo Boisvert", "Megh Thakkar", "Maxime Gasse", "Massimo Caccia", "Thibault L. De Chezelles", "Quentin Cappart", "Nicolas Chapados", "Alexandre Lacoste", "Alexandre Drouin"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-0195", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "5996-6051", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/0b82662b6c32e887bb252a74d8cb2d5e-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-07-07", "record_sha256": "304ed495adbbcce7e4c290947447371c562e67e940529998edcfcfb40cda3ae5", "relations": [], "resource_id": "arxiv:2407.05291", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.05291v2", "topics": ["evaluation", "planning", "workflow"], "updated_at": "2025-02-05"}, "text": "WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nDefines compositional enterprise workflows to evaluate planning and reasoning, with generation of ground-truth interaction traces.\nPlatforms: web\nTopics: evaluation, planning, workflow\nPublished: 2024-07-07\nAuthors: Léo Boisvert, Megh Thakkar, Maxime Gasse, Massimo Caccia, Thibault Le Sellier De Chezelles, Quentin Cappart, Nicolas Chapados, Alexandre Lacoste, Alexandre Drouin\nRecorded arXiv version: v2\n\nPublication edition: WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks\nPublication authors: Léo Boisvert, Megh Thakkar, Maxime Gasse, Massimo Caccia, Thibault L. De Chezelles, Quentin Cappart, Nicolas Chapados, Alexandre Lacoste, Alexandre Drouin\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 5996-6051\nPublication doi: 10.52202/079017-0195\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/0b82662b6c32e887bb252a74d8cb2d5e-Abstract-Datasets_and_Benchmarks_Track.html\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.05291v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-05291/"}
{"id": "arxiv:2407.08850", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390a7415ad5f665de43f7d294e0a575785ef7e2810dead3363632174b2986af4", "submitted_at": "2024-07-11T20:18:19Z", "updated_at": "2024-08-13T23:41:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.08850v3"}, "authors": ["Peitong Duan", "Chin-yi Chen", "Gang Li", "Bjoern Hartmann", "Yang Li"], "canonical_source_url": "https://arxiv.org/abs/2407.08850", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-08850/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390a7415ad5f665de43f7d294e0a575785ef7e2810dead3363632174b2986af4", "submitted_at": "2024-07-11T20:18:19Z", "updated_at": "2024-08-13T23:41:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.08850v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.08850"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.08850"}], "name": "UICrit: Enhancing Automated Design Evaluation with a UICritique Dataset", "publication": null, "published_at": "2024-07-11", "record_sha256": "0e91e8ae96bcd98ad69cf08b3b21c75bf56e84e9260f015182e7615b43ff392f", "relations": [], "resource_id": "arxiv:2407.08850", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2407.08850v3", "topics": ["screen-understanding", "evaluation"], "updated_at": "2024-08-13"}, "text": "UICrit: Enhancing Automated Design Evaluation with a UICritique Dataset\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces designer-authored mobile UI critiques and quality ratings, studying how targeted examples improve automated design feedback.\nPlatforms: unspecified\nTopics: screen-understanding, evaluation\nPublished: 2024-07-11\nAuthors: Peitong Duan, Chin-yi Chen, Gang Li, Bjoern Hartmann, Yang Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.08850v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-08850/"}
{"id": "arxiv:2407.09018", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a594fa7e150c9f5c22ed4d88f0551305f801f5780851ecf844802f2d2cd07c48", "submitted_at": "2024-07-12T06:14:46Z", "updated_at": "2024-07-12T06:14:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.09018v1"}, "authors": ["Yongxiang Hu", "Xuan Wang", "Yingchuan Wang", "Yu Zhang", "Shiyu Guo", "Chaoyi Chen", "Xin Wang", "Yangfan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2407.09018", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-09018/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a594fa7e150c9f5c22ed4d88f0551305f801f5780851ecf844802f2d2cd07c48", "submitted_at": "2024-07-12T06:14:46Z", "updated_at": "2024-07-12T06:14:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.09018v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.09018"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.09018"}], "name": "AUITestAgent: Automatic Requirements Oriented GUI Function Testing", "publication": null, "published_at": "2024-07-12", "record_sha256": "7a727c76703fff83f2e148dfa105f476bd0834b199ef096112b01ecab92f5241", "relations": [], "resource_id": "arxiv:2407.09018", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.09018v1", "topics": ["automation", "evaluation"], "updated_at": "2024-07-12"}, "text": "AUITestAgent: Automatic Requirements Oriented GUI Function Testing\nResource type: paper\nReview status: curated\nScope: core\n\nAutomates mobile GUI tests from natural-language requirements, combining action generation with verification based on interaction traces.\nPlatforms: unspecified\nTopics: automation, evaluation\nPublished: 2024-07-12\nAuthors: Yongxiang Hu, Xuan Wang, Yingchuan Wang, Yu Zhang, Shiyu Guo, Chaoyi Chen, Xin Wang, Yangfan Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.09018v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-09018/"}
{"id": "arxiv:2407.09295", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88576325385d3c361f0666311eae8858ea0d4a0098d73221f789550bebf1e5db", "submitted_at": "2024-07-12T14:30:05Z", "updated_at": "2025-03-16T07:13:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.09295v3"}, "authors": ["Yulong Yang", "Xinshan Yang", "Shuaidong Li", "Chenhao Lin", "Zhengyu Zhao", "Chao Shen", "Tianwei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2407.09295", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-09295/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88576325385d3c361f0666311eae8858ea0d4a0098d73221f789550bebf1e5db", "submitted_at": "2024-07-12T14:30:05Z", "updated_at": "2025-03-16T07:13:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.09295v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2407.09295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.09295"}], "name": "Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents", "publication": null, "published_at": "2024-07-12", "record_sha256": "9c025651f292ad69efaee299ee8b55a2bcc283de5768fdac150d71fb2eb6b8ff", "relations": [], "resource_id": "arxiv:2407.09295", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.09295v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-16"}, "text": "Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDevelops a threat taxonomy and evaluation framework for attacks against multimodal mobile GUI agents, with case studies and dataset experiments.\nPlatforms: unspecified\nTopics: safety-security, evaluation\nPublished: 2024-07-12\nAuthors: Yulong Yang, Xinshan Yang, Shuaidong Li, Chenhao Lin, Zhengyu Zhao, Chao Shen, Tianwei Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.09295v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-09295/"}
{"id": "arxiv:2407.10956", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7922a723142ac62018064fe5fc9357fb77c08e316ac4d6d6ba728f1b337df690", "submitted_at": "2024-07-15T17:54:37Z", "updated_at": "2024-07-15T17:54:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.10956v1"}, "authors": ["Ruisheng Cao", "Fangyu Lei", "Haoyuan Wu", "Jixuan Chen", "Yeqiao Fu", "Hongcheng Gao", "Xinzhuang Xiong", "Hanchong Zhang", "Yuchen Mao", "Wenjing Hu", "Tianbao Xie", "Hongshen Xu", "Danyang Zhang", "Sida Wang", "Ruoxi Sun", "Pengcheng Yin", "Caiming Xiong", "Ansong Ni", "Qian Liu", "Victor Zhong", "Lu Chen", "Kai Yu", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2407.10956", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-10956/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7922a723142ac62018064fe5fc9357fb77c08e316ac4d6d6ba728f1b337df690", "submitted_at": "2024-07-15T17:54:37Z", "updated_at": "2024-07-15T17:54:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.10956v1"}, "publication": {"authors": ["Ruisheng Cao", "Fangyu Lei", "Haoyuan Wu", "Jixuan Chen", "Yeqiao Fu", "Hongcheng Gao", "Xinzhuang Xiong", "Hanchong Zhang", "Yuchen Mao", "Wenjing Hu", "Tianbao Xie", "Hongshen Xu", "Danyang Zhang", "Sida Wang", "Ruoxi Sun", "Pengcheng Yin", "Caiming Xiong", "Ansong Ni", "Qian Liu", "Victor Zhong", "Lu Chen", "Kai Yu", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-3421", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "107703-107744", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/c2f71567cd53464161cab3336e8fc865-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "872286a0210616be9737de716cc6069a1e9cbee1df3182d45c718582e185f62f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.10956"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.10956"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/c2f71567cd53464161cab3336e8fc865-Paper-Datasets_and_Benchmarks_Track.pdf"}], "name": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?", "publication": {"authors": ["Ruisheng Cao", "Fangyu Lei", "Haoyuan Wu", "Jixuan Chen", "Yeqiao Fu", "Hongcheng Gao", "Xinzhuang Xiong", "Hanchong Zhang", "Yuchen Mao", "Wenjing Hu", "Tianbao Xie", "Hongshen Xu", "Danyang Zhang", "Sida Wang", "Ruoxi Sun", "Pengcheng Yin", "Caiming Xiong", "Ansong Ni", "Qian Liu", "Victor Zhong", "Lu Chen", "Kai Yu", "Tao Yu"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-3421", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "107703-107744", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/c2f71567cd53464161cab3336e8fc865-Paper-Datasets_and_Benchmarks_Track.pdf", "publisher": null, "series": null, "title": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-07-15", "record_sha256": "0a3aec0ae9ff8da0b580158fb8b348a47bad85cef36cd3fa835d7624e99d286a", "relations": [], "resource_id": "arxiv:2407.10956", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.10956v1", "topics": ["evaluation", "science", "workflow"], "updated_at": "2024-07-15"}, "text": "Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?\nResource type: paper\nReview status: curated\nScope: core\n\nBenchmarks multimodal agents on data-science and engineering workflows involving enterprise applications, code, and GUI operations in real computer environments.\nPlatforms: desktop, web\nTopics: evaluation, science, workflow\nPublished: 2024-07-15\nAuthors: Ruisheng Cao, Fangyu Lei, Haoyuan Wu, Jixuan Chen, Yeqiao Fu, Hongcheng Gao, Xinzhuang Xiong, Hanchong Zhang, Yuchen Mao, Wenjing Hu, Tianbao Xie, Hongshen Xu, Danyang Zhang, Sida Wang, Ruoxi Sun, Pengcheng Yin, Caiming Xiong, Ansong Ni, Qian Liu, Victor Zhong, Lu Chen, Kai Yu, Tao Yu\nRecorded arXiv version: v1\n\nPublication edition: Spider2-V: How Far Are Multimodal Agents From Automating Data Science and Engineering Workflows?\nPublication authors: Ruisheng Cao, Fangyu Lei, Haoyuan Wu, Jixuan Chen, Yeqiao Fu, Hongcheng Gao, Xinzhuang Xiong, Hanchong Zhang, Yuchen Mao, Wenjing Hu, Tianbao Xie, Hongshen Xu, Danyang Zhang, Sida Wang, Ruoxi Sun, Pengcheng Yin, Caiming Xiong, Ansong Ni, Qian Liu, Victor Zhong, Lu Chen, Kai Yu, Tao Yu\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 107703-107744\nPublication doi: 10.52202/079017-3421\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/c2f71567cd53464161cab3336e8fc865-Abstract-Datasets_and_Benchmarks_Track.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.10956v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-10956/"}
{"id": "arxiv:2407.13032", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18cfd29fcb66d703d90f747c162db1b3fc6a43f400b4eacbddf081bb83ed7947", "submitted_at": "2024-07-17T21:44:28Z", "updated_at": "2024-07-17T21:44:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.13032v1"}, "authors": ["Tamer Abuelsaad", "Deepak Akkil", "Prasenjit Dey", "Ashish Jagmohan", "Aditya Vempaty", "Ravi Kokku"], "canonical_source_url": "https://arxiv.org/abs/2407.13032", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-13032/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18cfd29fcb66d703d90f747c162db1b3fc6a43f400b4eacbddf081bb83ed7947", "submitted_at": "2024-07-17T21:44:28Z", "updated_at": "2024-07-17T21:44:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.13032v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.13032"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.13032"}], "name": "Agent-E: From Autonomous Web Navigation to Foundational Design Principles in Agentic Systems", "publication": null, "published_at": "2024-07-17", "record_sha256": "ad5f870525949724c4b408ae823a588d51d07721d69176140873bb74c4c0769b", "relations": [], "resource_id": "arxiv:2407.13032", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.13032v1", "topics": ["web-navigation", "planning", "efficiency", "state-tracking"], "updated_at": "2024-07-17"}, "text": "Agent-E: From Autonomous Web Navigation to Foundational Design Principles in Agentic Systems\nResource type: paper\nReview status: curated\nScope: core\n\nAgent-E combines hierarchical web-task control with DOM distillation and observations of interface changes. WebVoyager experiments inform a broader discussion of primitive skills, observation reduction, and self-improvement in agent architecture.\nPlatforms: web\nTopics: web-navigation, planning, efficiency, state-tracking\nPublished: 2024-07-17\nAuthors: Tamer Abuelsaad, Deepak Akkil, Prasenjit Dey, Ashish Jagmohan, Aditya Vempaty, Ravi Kokku\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.13032v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-13032/"}
{"id": "arxiv:2407.15673", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4caa87e0b88d9375a23ca412ed19b3806f67712ca4ac222a39746e08ee80b07d", "submitted_at": "2024-07-22T14:41:32Z", "updated_at": "2024-08-18T09:02:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.15673v3"}, "authors": ["Segev Shlomov", "Avi Yaeli", "Sami Marreed", "Sivan Schwartz", "Netanel Eder", "Offer Akrabi", "Sergey Zeltyn"], "canonical_source_url": "https://arxiv.org/abs/2407.15673", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-15673/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4caa87e0b88d9375a23ca412ed19b3806f67712ca4ac222a39746e08ee80b07d", "submitted_at": "2024-07-22T14:41:32Z", "updated_at": "2024-08-18T09:02:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2407.15673v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.15673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.15673"}], "name": "IDA: Breaking Barriers in No-code UI Automation Through Large Language Models and Human-Centric Design", "publication": null, "published_at": "2024-07-22", "record_sha256": "568f683760a55720aa9e7aecb645f6f8cbd9601df5139ca3ffc176929c44b5b8", "relations": [], "resource_id": "arxiv:2407.15673", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.15673v3", "topics": ["human-in-the-loop", "automation", "workflow", "industry"], "updated_at": "2024-08-18"}, "text": "IDA: Breaking Barriers in No-code UI Automation Through Large Language Models and Human-Centric Design\nResource type: paper\nReview status: curated\nScope: core\n\nIDA helps business users create web UI automation through guided demonstrations, a semantic programming model, and a teaching-oriented interface. A prototype user study examines automation creation with enterprise applications by users without programming backgrounds.\nPlatforms: web\nTopics: human-in-the-loop, automation, workflow, industry\nPublished: 2024-07-22\nAuthors: Segev Shlomov, Avi Yaeli, Sami Marreed, Sivan Schwartz, Netanel Eder, Offer Akrabi, Sergey Zeltyn\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.15673v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-15673/"}
{"id": "arxiv:2407.15711", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f028fb7f4d64fd4c29886d49b5d6f0fabddba309b669e019edf13586deb9c90", "submitted_at": "2024-07-22T15:18:45Z", "updated_at": "2024-10-21T15:45:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.15711v2"}, "authors": ["Ori Yoran", "Samuel Joseph Amouyal", "Chaitanya Malaviya", "Ben Bogin", "Ofir Press", "Jonathan Berant"], "canonical_source_url": "https://arxiv.org/abs/2407.15711", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-15711/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f028fb7f4d64fd4c29886d49b5d6f0fabddba309b669e019edf13586deb9c90", "submitted_at": "2024-07-22T15:18:45Z", "updated_at": "2024-10-21T15:45:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.15711v2"}, "publication": {"authors": ["Ori Yoran", "Samuel Joseph Amouyal", "Chaitanya Malaviya", "Ben Bogin", "Ofir Press", "Jonathan Berant"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.505/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8938-8968", "pdf_url": "https://aclanthology.org/2024.emnlp-main.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.505/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:51Z", "role": "primary", "sha256": "a18f0ea465d83796bcd83b95f48c3c13cbdc0aa778ea0dd454aa61f8082920b4", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.emnlp-main.505/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2407.15711"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.15711"}, {"rel": "publication", "url": "https://aclanthology.org/2024.emnlp-main.505/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.emnlp-main.505.pdf"}], "name": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?", "publication": {"authors": ["Ori Yoran", "Samuel Joseph Amouyal", "Chaitanya Malaviya", "Ben Bogin", "Ofir Press", "Jonathan Berant"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.emnlp-main.505", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.emnlp-main.505/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8938-8968", "pdf_url": "https://aclanthology.org/2024.emnlp-main.505.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?", "type": "conference-paper", "url": "https://aclanthology.org/2024.emnlp-main.505/", "venue": "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing", "volume": null}, "published_at": "2024-07-22", "record_sha256": "9b77d37d05f8150ce5bcc26f1c13b9a90f75c6d93a656a13970aad172c4a8860", "relations": [], "resource_id": "arxiv:2407.15711", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.15711v2", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2024-10-21"}, "text": "AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces realistic open-web tasks and the SeePlanAct agent, analyzing the limitations of browsing and retrieval in time-consuming assistant work.\nPlatforms: web\nTopics: evaluation, web-navigation, planning\nPublished: 2024-07-22\nAuthors: Ori Yoran, Samuel Joseph Amouyal, Chaitanya Malaviya, Ben Bogin, Ofir Press, Jonathan Berant\nRecorded arXiv version: v2\n\nPublication edition: AssistantBench: Can Web Agents Solve Realistic and Time-Consuming Tasks?\nPublication authors: Ori Yoran, Samuel Joseph Amouyal, Chaitanya Malaviya, Ben Bogin, Ofir Press, Jonathan Berant\nPublication edition date (recorded precision): 2024-11\nPublication venue: Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing\nPublication publisher: Association for Computational Linguistics\nPublication pages: 8938-8968\nPublication doi: 10.18653/v1/2024.emnlp-main.505\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2024.emnlp-main.505/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.15711v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-15711/"}
{"id": "arxiv:2407.17490", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c80665716bf665fe4688a27a4c7dcf26fcfe93cb7f79a6f28784a795e0bd3cb", "submitted_at": "2024-07-03T17:59:58Z", "updated_at": "2025-05-29T02:43:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.17490v2"}, "authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2407.17490", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-17490/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c80665716bf665fe4688a27a4c7dcf26fcfe93cb7f79a6f28784a795e0bd3cb", "submitted_at": "2024-07-03T17:59:58Z", "updated_at": "2025-05-29T02:43:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2407.17490v2"}, "publication": {"authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Guozhi Wang", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.110", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.110/", "https://aclanthology.org/2025.findings-acl.110.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, lists nine authors including Guozhi Wang; the retained arXiv snapshot lists eight. Each edition retains its own authors.", "pages": "2138-2156", "pdf_url": "https://aclanthology.org/2025.findings-acl.110.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.110/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "9cb5bd38791ee3df95d04815ca650bf0f020518b009c04b4a54857c122f161c8", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000709+00:00", "role": "primary", "sha256": "344342e1c229e46269d26d879b32a4d2947932607b968f7e18382d231ad05ee2", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2407.17490"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.17490"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.110/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.110.pdf"}], "name": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "publication": {"authors": ["Yuxiang Chai", "Siyuan Huang", "Yazhe Niu", "Han Xiao", "Liang Liu", "Guozhi Wang", "Dingyu Zhang", "Shuai Ren", "Hongsheng Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.110", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.110/", "https://aclanthology.org/2025.findings-acl.110.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, lists nine authors including Guozhi Wang; the retained arXiv snapshot lists eight. Each edition retains its own authors.", "pages": "2138-2156", "pdf_url": "https://aclanthology.org/2025.findings-acl.110.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.110/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-07-03", "record_sha256": "976fff8000c47ab673fb6db718abe9b8f557b7d82d02725812ab91dc9fbc5103", "relations": [], "resource_id": "arxiv:2407.17490", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.17490v2", "topics": ["grounding", "screen-understanding", "mobile-navigation"], "updated_at": "2025-05-29"}, "text": "AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Android screenshots with element locations, functionality descriptions, and instructions linked to sequences of GUI actions.\nPlatforms: android\nTopics: grounding, screen-understanding, mobile-navigation\nPublished: 2024-07-03\nAuthors: Yuxiang Chai, Siyuan Huang, Yazhe Niu, Han Xiao, Liang Liu, Dingyu Zhang, Shuai Ren, Hongsheng Li\nRecorded arXiv version: v2\n\nPublication edition: AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents\nPublication authors: Yuxiang Chai, Siyuan Huang, Yazhe Niu, Han Xiao, Liang Liu, Guozhi Wang, Dingyu Zhang, Shuai Ren, Hongsheng Li\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 2138-2156\nPublication doi: 10.18653/v1/2025.findings-acl.110\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. The publisher PDF, page 1, lists nine authors including Guozhi Wang; the retained arXiv snapshot lists eight. Each edition retains its own authors.\nPublication source: https://aclanthology.org/2025.findings-acl.110/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.17490v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-17490/"}
{"id": "arxiv:2407.18901", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e59087ecf28d07baf81b60e4166a866d4bea4d8cd934bea06c2149dd61f38f1a", "submitted_at": "2024-07-26T17:55:45Z", "updated_at": "2024-07-26T17:55:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.18901v1"}, "authors": ["Harsh Trivedi", "Tushar Khot", "Mareike Hartmann", "Ruskin Manku", "Vinty Dong", "Edward Li", "Shashank Gupta", "Ashish Sabharwal", "Niranjan Balasubramanian"], "canonical_source_url": "https://arxiv.org/abs/2407.18901", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-18901/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e59087ecf28d07baf81b60e4166a866d4bea4d8cd934bea06c2149dd61f38f1a", "submitted_at": "2024-07-26T17:55:45Z", "updated_at": "2024-07-26T17:55:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.18901v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.18901"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.18901"}], "name": "AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents", "publication": null, "published_at": "2024-07-26", "record_sha256": "c5042f846c9ef8c12a74bc79ad9eecf4473d3652a91bede5c1e792ba0ca395ca", "relations": [], "resource_id": "arxiv:2407.18901", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2407.18901v1", "topics": ["evaluation", "tool-integration", "cross-app"], "updated_at": "2024-07-26"}, "text": "AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces simulated application APIs and state-based task evaluation for interactive coding agents performing everyday digital workflows.\nPlatforms: unspecified\nTopics: evaluation, tool-integration, cross-app\nPublished: 2024-07-26\nAuthors: Harsh Trivedi, Tushar Khot, Mareike Hartmann, Ruskin Manku, Vinty Dong, Edward Li, Shashank Gupta, Ashish Sabharwal, Niranjan Balasubramanian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.18901v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-18901/"}
{"id": "arxiv:2407.19056", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c443e6a508c438332ac5990cf254a4441f5bdcc24354fd8aeca36bcbf4604a29", "submitted_at": "2024-07-26T19:27:17Z", "updated_at": "2024-07-26T19:27:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.19056v1"}, "authors": ["Zilong Wang", "Yuedong Cui", "Li Zhong", "Zimin Zhang", "Da Yin", "Bill Yuchen Lin", "Jingbo Shang"], "canonical_source_url": "https://arxiv.org/abs/2407.19056", "citation_url": "https://awesomecua.github.io/resources/arxiv-2407-19056/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c443e6a508c438332ac5990cf254a4441f5bdcc24354fd8aeca36bcbf4604a29", "submitted_at": "2024-07-26T19:27:17Z", "updated_at": "2024-07-26T19:27:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2407.19056v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2407.19056"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2407.19056"}], "name": "OfficeBench: Benchmarking Language Agents across Multiple Applications for Office Automation", "publication": null, "published_at": "2024-07-26", "record_sha256": "0b4c0bf2d4e6c4e1bf0abd45dba882fe866fc80cd730a5f5347e634372c289f3", "relations": [], "resource_id": "arxiv:2407.19056", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2407.19056v1", "topics": ["evaluation", "cross-app", "long-horizon"], "updated_at": "2024-07-26"}, "text": "OfficeBench: Benchmarking Language Agents across Multiple Applications for Office Automation\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates language agents on office workflows requiring information integration, long-term planning, and switching between applications.\nPlatforms: desktop\nTopics: evaluation, cross-app, long-horizon\nPublished: 2024-07-26\nAuthors: Zilong Wang, Yuedong Cui, Li Zhong, Zimin Zhang, Da Yin, Bill Yuchen Lin, Jingbo Shang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2407.19056v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2407-19056/"}
{"id": "arxiv:2408.00203", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfe55dd83d2f5ac48f9e3efb6bdf4aab7f11d81c0118725555456cce276adae3", "submitted_at": "2024-08-01T00:00:43Z", "updated_at": "2024-08-01T00:00:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.00203v1"}, "authors": ["Yadong Lu", "Jianwei Yang", "Yelong Shen", "Ahmed Awadallah"], "canonical_source_url": "https://arxiv.org/abs/2408.00203", "citation_url": "https://awesomecua.github.io/resources/arxiv-2408-00203/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfe55dd83d2f5ac48f9e3efb6bdf4aab7f11d81c0118725555456cce276adae3", "submitted_at": "2024-08-01T00:00:43Z", "updated_at": "2024-08-01T00:00:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.00203v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2408.00203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.00203"}], "name": "OmniParser for Pure Vision Based GUI Agent", "publication": null, "published_at": "2024-08-01", "record_sha256": "64aa9cba28aeb6f42dbb03e144c8a56ea243e336bab88da9079b4d5fdfd0a918", "relations": [], "resource_id": "arxiv:2408.00203", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2408.00203v1", "topics": ["grounding", "screen-understanding"], "updated_at": "2024-08-01"}, "text": "OmniParser for Pure Vision Based GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nParses screenshots into interactable regions and semantic element descriptions to support visually grounded agent actions.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\nPublished: 2024-08-01\nAuthors: Yadong Lu, Jianwei Yang, Yelong Shen, Ahmed Awadallah\nRecorded arXiv version: v1\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2408.00203v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2408-00203/"}
{"id": "arxiv:2408.02544", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d66c0c2102dd2734d5334faeb12b4971d50a2d2e50a6671ede2fa78a84a0c03f", "submitted_at": "2024-08-05T15:16:22Z", "updated_at": "2025-09-05T09:21:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2408.02544v3"}, "authors": ["Xinbei Ma", "Yiting Wang", "Yao Yao", "Tongxin Yuan", "Aston Zhang", "Zhuosheng Zhang", "Hai Zhao"], "canonical_source_url": "https://arxiv.org/abs/2408.02544", "citation_url": "https://awesomecua.github.io/resources/arxiv-2408-02544/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d66c0c2102dd2734d5334faeb12b4971d50a2d2e50a6671ede2fa78a84a0c03f", "submitted_at": "2024-08-05T15:16:22Z", "updated_at": "2025-09-05T09:21:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2408.02544v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2408.02544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.02544"}], "name": "Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental Distractions", "publication": null, "published_at": "2024-08-05", "record_sha256": "69252f04ce71ded8db04848516ec641e3ba5e629aa93f2de46a41514e494e1bc", "relations": [], "resource_id": "arxiv:2408.02544", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2408.02544v3", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-09-05"}, "text": "Caution for the Environment: Multimodal LLM Agents are Susceptible to Environmental Distractions\nResource type: paper\nReview status: curated\nScope: core\n\nMeasures how irrelevant GUI content affects multimodal agents using a simulated dataset and three perception patterns. The study also examines adversarial environment injection and possible ways to improve instruction faithfulness.\nPlatforms: unspecified\nTopics: safety-security, robustness, evaluation\nPublished: 2024-08-05\nAuthors: Xinbei Ma, Yiting Wang, Yao Yao, Tongxin Yuan, Aston Zhang, Zhuosheng Zhang, Hai Zhao\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract describes a simulated dataset; real-browser deployment is not asserted.\n\nPrimary source: https://arxiv.org/abs/2408.02544v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2408-02544/"}
{"id": "arxiv:2408.06327", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "603bfa4357617bfeb7a718065e77c71d77614f7240258684938dd4e6fd1a20df", "submitted_at": "2024-08-12T17:44:17Z", "updated_at": "2024-08-12T17:44:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.06327v1"}, "authors": ["Xiao Liu", "Tianjie Zhang", "Yu Gu", "Iat Long Iong", "Yifan Xu", "Xixuan Song", "Shudan Zhang", "Hanyu Lai", "Xinyi Liu", "Hanlin Zhao", "Jiadai Sun", "Xinyue Yang", "Yu Yang", "Zehan Qi", "Shuntian Yao", "Xueqiao Sun", "Siyi Cheng", "Qinkai Zheng", "Hao Yu", "Hanchen Zhang", "Wenyi Hong", "Ming Ding", "Lihang Pan", "Xiaotao Gu", "Aohan Zeng", "Zhengxiao Du", "Chan Hee Song", "Yu Su", "Yuxiao Dong", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2408.06327", "citation_url": "https://awesomecua.github.io/resources/arxiv-2408-06327/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "603bfa4357617bfeb7a718065e77c71d77614f7240258684938dd4e6fd1a20df", "submitted_at": "2024-08-12T17:44:17Z", "updated_at": "2024-08-12T17:44:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.06327v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2408.06327"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.06327"}], "name": "VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents", "publication": null, "published_at": "2024-08-12", "record_sha256": "fc058b923fe3eb7575b45594c3c43effdd1f94e21b7fe146d1cf2397ce9e8837", "relations": [], "resource_id": "arxiv:2408.06327", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2408.06327v1", "topics": ["evaluation", "data-generation"], "updated_at": "2024-08-12"}, "text": "VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nBenchmarks visual agents across GUI, embodied, and design tasks, with training trajectories collected through solvers, agents, and human demonstrations.\nPlatforms: cross-platform\nTopics: evaluation, data-generation\nPublished: 2024-08-12\nAuthors: Xiao Liu, Tianjie Zhang, Yu Gu, Iat Long Iong, Yifan Xu, Xixuan Song, Shudan Zhang, Hanyu Lai, Xinyi Liu, Hanlin Zhao, Jiadai Sun, Xinyue Yang, Yu Yang, Zehan Qi, Shuntian Yao, Xueqiao Sun, Siyi Cheng, Qinkai Zheng, Hao Yu, Hanchen Zhang, Wenyi Hong, Ming Ding, Lihang Pan, Xiaotao Gu, Aohan Zeng, Zhengxiao Du, Chan Hee Song, Yu Su, Yuxiao Dong, Jie Tang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2408.06327v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2408-06327/"}
{"id": "arxiv:2408.11824", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5b471c61c2ac9d4932bd71052ea35585699d70e9710f8366bc4fa8e4f2ebbae", "submitted_at": "2024-08-05T06:31:39Z", "updated_at": "2025-09-17T09:06:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2408.11824v4"}, "authors": ["Yanda Li", "Chi Zhang", "Wenjia Jiang", "Wanqi Yang", "Bin Fu", "Pei Cheng", "Xin Chen", "Ling Chen", "Yunchao Wei"], "canonical_source_url": "https://arxiv.org/abs/2408.11824", "citation_url": "https://awesomecua.github.io/resources/arxiv-2408-11824/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5b471c61c2ac9d4932bd71052ea35585699d70e9710f8366bc4fa8e4f2ebbae", "submitted_at": "2024-08-05T06:31:39Z", "updated_at": "2025-09-17T09:06:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2408.11824v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2408.11824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.11824"}], "name": "AppAgent v2: Advanced Agent for Flexible Mobile Interactions", "publication": null, "published_at": "2024-08-05", "record_sha256": "d82db1d9c502fafc6896ee2b91a293f7a5a3614b4073dab4818cd9c27bc73b85", "relations": [], "resource_id": "arxiv:2408.11824", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2408.11824v4", "topics": ["mobile-navigation", "memory"], "updated_at": "2025-09-17"}, "text": "AppAgent v2: Advanced Agent for Flexible Mobile Interactions\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates exploration and deployment while using a structured application knowledge base and retrieval to guide mobile workflows.\nPlatforms: android\nTopics: mobile-navigation, memory\nPublished: 2024-08-05\nAuthors: Yanda Li, Chi Zhang, Wenjia Jiang, Wanqi Yang, Bin Fu, Pei Cheng, Xin Chen, Ling Chen, Yunchao Wei\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2408.11824v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2408-11824/"}
{"id": "arxiv:2408.15978", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e3a34f16101df7a96ffdabbdf86b060e195494f16499f20bdf54110a0394cae", "submitted_at": "2024-08-28T17:49:29Z", "updated_at": "2024-08-28T17:49:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.15978v1"}, "authors": ["Yao Zhang", "Zijian Ma", "Yunpu Ma", "Zhen Han", "Yu Wu", "Volker Tresp"], "canonical_source_url": "https://arxiv.org/abs/2408.15978", "citation_url": "https://awesomecua.github.io/resources/arxiv-2408-15978/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e3a34f16101df7a96ffdabbdf86b060e195494f16499f20bdf54110a0394cae", "submitted_at": "2024-08-28T17:49:29Z", "updated_at": "2024-08-28T17:49:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2408.15978v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2408.15978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2408.15978"}], "name": "WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration", "publication": null, "published_at": "2024-08-28", "record_sha256": "80fba9f63cab578fe9344e651230cabf29ede07eace7cc57f14f000cf45a11cf", "relations": [], "resource_id": "arxiv:2408.15978", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2408.15978v1", "topics": ["planning", "web-navigation", "long-horizon"], "updated_at": "2024-08-28"}, "text": "WebPilot: A Versatile and Autonomous Multi-Agent System for Web Task Execution with Strategic Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nWebPilot separates global subtask planning from local Monte Carlo tree search during web execution. It continually refines the high-level plan and adapts local search to uncertain transitions and incomplete observations, evaluated on WebArena and MiniWoB++.\nPlatforms: web\nTopics: planning, web-navigation, long-horizon\nPublished: 2024-08-28\nAuthors: Yao Zhang, Zijian Ma, Yunpu Ma, Zhen Han, Yu Wu, Volker Tresp\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2408.15978v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2408-15978/"}
{"id": "arxiv:2409.01927", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cdb228e652218d7f8981d9d16ca29b465004675c510d1e9c5c7b543bec9dc454", "submitted_at": "2024-09-03T14:17:09Z", "updated_at": "2024-09-03T14:17:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.01927v1"}, "authors": ["Segev Shlomov", "Ben wiesel", "Aviad Sela", "Ido Levy", "Liane Galanti", "Roy Abitbol"], "canonical_source_url": "https://arxiv.org/abs/2409.01927", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-01927/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cdb228e652218d7f8981d9d16ca29b465004675c510d1e9c5c7b543bec9dc454", "submitted_at": "2024-09-03T14:17:09Z", "updated_at": "2024-09-03T14:17:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.01927v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2409.01927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.01927"}], "name": "From Grounding to Planning: Benchmarking Bottlenecks in Web Agents", "publication": null, "published_at": "2024-09-03", "record_sha256": "ad2370e44e0aedce4990057d376daed368abab9dc11248f7de04781497e1e849", "relations": [], "resource_id": "arxiv:2409.01927", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.01927v1", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2024-09-03"}, "text": "From Grounding to Planning: Benchmarking Bottlenecks in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates web-agent planning and grounding evaluations on Mind2Web to analyze which components limit task performance.\nPlatforms: web\nTopics: evaluation, grounding, planning\nPublished: 2024-09-03\nAuthors: Segev Shlomov, Ben wiesel, Aviad Sela, Ido Levy, Liane Galanti, Roy Abitbol\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.01927v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-01927/"}
{"id": "arxiv:2409.07429", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4c142be3a9c37bbb0d4f401a02de25bd25b4c3968f42ad9a61513abff2b5460", "submitted_at": "2024-09-11T17:21:00Z", "updated_at": "2024-09-11T17:21:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.07429v1"}, "authors": ["Zora Zhiruo Wang", "Jiayuan Mao", "Daniel Fried", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2409.07429", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-07429/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4c142be3a9c37bbb0d4f401a02de25bd25b4c3968f42ad9a61513abff2b5460", "submitted_at": "2024-09-11T17:21:00Z", "updated_at": "2024-09-11T17:21:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.07429v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.07429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.07429"}], "name": "Agent Workflow Memory", "publication": null, "published_at": "2024-09-11", "record_sha256": "dcc49da571ad4fd512ddba6bad77d62abb96c45911e301e29a37fd7e6ec5613f", "relations": [], "resource_id": "arxiv:2409.07429", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.07429v1", "topics": ["memory", "workflow", "web-navigation"], "updated_at": "2024-09-11"}, "text": "Agent Workflow Memory\nResource type: paper\nReview status: curated\nScope: core\n\nAgent Workflow Memory extracts reusable task routines and selectively supplies them to guide later web actions. It studies both offline workflow induction from training examples and online induction from test experience on Mind2Web and WebArena.\nPlatforms: web\nTopics: memory, workflow, web-navigation\nPublished: 2024-09-11\nAuthors: Zora Zhiruo Wang, Jiayuan Mao, Daniel Fried, Graham Neubig\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.07429v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-07429/"}
{"id": "arxiv:2409.08264", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe77f7f8ffd7490192ab54b7f6d49c0cb5b10be6df0a645bfb181b3be724118a", "submitted_at": "2024-09-12T17:56:43Z", "updated_at": "2024-09-13T20:17:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.08264v2"}, "authors": ["Rogerio Bonatti", "Dan Zhao", "Francesco Bonacci", "Dillon Dupont", "Sara Abdali", "Yinheng Li", "Yadong Lu", "Justin Wagle", "Kazuhito Koishida", "Arthur Bucker", "Lawrence Jang", "Zack Hui"], "canonical_source_url": "https://arxiv.org/abs/2409.08264", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-08264/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe77f7f8ffd7490192ab54b7f6d49c0cb5b10be6df0a645bfb181b3be724118a", "submitted_at": "2024-09-12T17:56:43Z", "updated_at": "2024-09-13T20:17:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.08264v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.08264"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.08264"}], "name": "Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale", "publication": null, "published_at": "2024-09-12", "record_sha256": "aa1e1adf1dbabcb0603c3389c7d98eb60d67d4f23f6322833df02522692e7530", "relations": [], "resource_id": "arxiv:2409.08264", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.08264v2", "topics": ["evaluation", "cross-app"], "updated_at": "2024-09-13"}, "text": "Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts real-desktop evaluation to Windows applications and supports parallel execution of benchmark tasks.\nPlatforms: desktop, windows\nTopics: evaluation, cross-app\nPublished: 2024-09-12\nAuthors: Rogerio Bonatti, Dan Zhao, Francesco Bonacci, Dillon Dupont, Sara Abdali, Yinheng Li, Yadong Lu, Justin Wagle, Kazuhito Koishida, Arthur Bucker, Lawrence Jang, Zack Hui\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.08264v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-08264/"}
{"id": "arxiv:2409.09354", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b46dabe4135f0caca15d54f84cfe805cdb7fe7f1abcce6b004483d99eb21e637", "submitted_at": "2024-09-14T07:54:25Z", "updated_at": "2024-09-14T07:54:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.09354v1"}, "authors": ["Kelin Fu", "Yang Tian", "Kaigui Bian"], "canonical_source_url": "https://arxiv.org/abs/2409.09354", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-09354/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b46dabe4135f0caca15d54f84cfe805cdb7fe7f1abcce6b004483d99eb21e637", "submitted_at": "2024-09-14T07:54:25Z", "updated_at": "2024-09-14T07:54:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.09354v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.09354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.09354"}], "name": "PeriGuru: A Peripheral Robotic Mobile App Operation Assistant based on GUI Image Understanding and Prompting with LLM", "publication": null, "published_at": "2024-09-14", "record_sha256": "891b11724575828f507f11238754d7ac2c9acb9300637bce11e42c2a27415ea3", "relations": [], "resource_id": "arxiv:2409.09354", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.09354v1", "topics": ["mobile-navigation", "screen-understanding", "automation"], "updated_at": "2024-09-14"}, "text": "PeriGuru: A Peripheral Robotic Mobile App Operation Assistant based on GUI Image Understanding and Prompting with LLM\nResource type: paper\nReview status: curated\nScope: core\n\nPeriGuru interprets smartphone GUI screenshots with computer vision and uses a language model to decide the next action. A peripheral robotic arm executes the phone interactions, providing an alternative mechanism for operating mobile apps.\nPlatforms: unspecified\nTopics: mobile-navigation, screen-understanding, automation\nPublished: 2024-09-14\nAuthors: Kelin Fu, Yang Tian, Kaigui Bian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.09354v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-09354/"}
{"id": "arxiv:2409.10741", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5bdbd80e252cfa37dc62d666a202e647f93813549c55dde82daf36afe309c7f2", "submitted_at": "2024-09-16T21:18:39Z", "updated_at": "2024-09-16T21:18:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.10741v1"}, "authors": ["Mobina Shahbandeh", "Parsa Alian", "Noor Nashid", "Ali Mesbah"], "canonical_source_url": "https://arxiv.org/abs/2409.10741", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-10741/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5bdbd80e252cfa37dc62d666a202e647f93813549c55dde82daf36afe309c7f2", "submitted_at": "2024-09-16T21:18:39Z", "updated_at": "2024-09-16T21:18:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.10741v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.10741"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.10741"}], "name": "NaviQAte: Functionality-Guided Web Application Navigation", "publication": null, "published_at": "2024-09-16", "record_sha256": "cc41475e1b8708af1e220c71526694b80e452c866bb9e703c1cae0c99d213878", "relations": [], "resource_id": "arxiv:2409.10741", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.10741v1", "topics": ["web-navigation", "automation"], "updated_at": "2024-09-16"}, "text": "NaviQAte: Functionality-Guided Web Application Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nFrames web-application exploration as functionality-oriented question answering, generating navigation actions from multimodal context.\nPlatforms: web\nTopics: web-navigation, automation\nPublished: 2024-09-16\nAuthors: Mobina Shahbandeh, Parsa Alian, Noor Nashid, Ali Mesbah\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.10741v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-10741/"}
{"id": "arxiv:2409.10811", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a32462e4cb09763750f8f22d7f4be32a6bdfe3db6c9b3de9c386dc7cd6b6f30", "submitted_at": "2024-09-17T00:58:00Z", "updated_at": "2025-10-01T17:58:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2409.10811v4"}, "authors": ["Shuqing Li", "Binchang Li", "Yepang Liu", "Cuiyun Gao", "Jianping Zhang", "Shing-Chi Cheung", "Michael R. Lyu"], "canonical_source_url": "https://arxiv.org/abs/2409.10811", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-10811/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a32462e4cb09763750f8f22d7f4be32a6bdfe3db6c9b3de9c386dc7cd6b6f30", "submitted_at": "2024-09-17T00:58:00Z", "updated_at": "2025-10-01T17:58:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2409.10811v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2409.10811"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.10811"}], "name": "Grounded GUI Understanding for Vision-Based Spatial Intelligent Agent: Exemplified by Extended Reality Apps", "publication": null, "published_at": "2024-09-17", "record_sha256": "42f303be4aafb81914577a044358636ee33df43bfc86b47f25b723e39fe2b147", "relations": [], "resource_id": "arxiv:2409.10811", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.10811v4", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-10-01"}, "text": "Grounded GUI Understanding for Vision-Based Spatial Intelligent Agent: Exemplified by Extended Reality Apps\nResource type: paper\nReview status: curated\nScope: core\n\nOrienter detects interactable GUI elements in XR applications by interpreting scene context and revisiting candidate elements through a validation and reflection loop. It addresses open-vocabulary element types and context-dependent interactability.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\nPublished: 2024-09-17\nAuthors: Shuqing Li, Binchang Li, Yepang Liu, Cuiyun Gao, Jianping Zhang, Shing-Chi Cheung, Michael R. Lyu\nRecorded arXiv version: v4\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced. The described environment is XR/VR applications; the operating system is not specified in the abstract.\n\nPrimary source: https://arxiv.org/abs/2409.10811v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-10811/"}
{"id": "arxiv:2409.11295", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "897cdbb6aa52478b260ecba830e0a13527c2614454f7c766c9f5c8c9bdbfd4e3", "submitted_at": "2024-09-17T15:49:44Z", "updated_at": "2025-03-12T20:54:00Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2409.11295v5"}, "authors": ["Zeyi Liao", "Lingbo Mo", "Chejian Xu", "Mintong Kang", "Jiawei Zhang", "Chaowei Xiao", "Yuan Tian", "Bo Li", "Huan Sun"], "canonical_source_url": "https://arxiv.org/abs/2409.11295", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-11295/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "897cdbb6aa52478b260ecba830e0a13527c2614454f7c766c9f5c8c9bdbfd4e3", "submitted_at": "2024-09-17T15:49:44Z", "updated_at": "2025-03-12T20:54:00Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2409.11295v5"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2409.11295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.11295"}], "name": "EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage", "publication": null, "published_at": "2024-09-17", "record_sha256": "43dc45786af29cfb507009dc1d3c85aba384916ce1db930526753877f5d0b3d5", "relations": [], "resource_id": "arxiv:2409.11295", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.11295v5", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-03-12"}, "text": "EIA: Environmental Injection Attack on Generalist Web Agents for Privacy Leakage\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates privacy-directed webpage instruction injection against generalist web agents using PII-related interaction steps from Mind2Web, including attempts to expose specific user information or entire requests.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2024-09-17\nAuthors: Zeyi Liao, Lingbo Mo, Chejian Xu, Mintong Kang, Jiawei Zhang, Chaowei Xiao, Yuan Tian, Bo Li, Huan Sun\nRecorded arXiv version: v5\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.11295v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-11295/"}
{"id": "arxiv:2409.12089", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f59976cec7adb9f9c88bd79bdfbeed040e4874b6d9ab0c229136f98101aa98", "submitted_at": "2024-09-18T16:04:10Z", "updated_at": "2024-10-06T21:25:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.12089v3"}, "authors": ["Wayne Chi", "Ameet Talwalkar", "Chris Donahue"], "canonical_source_url": "https://arxiv.org/abs/2409.12089", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-12089/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2f59976cec7adb9f9c88bd79bdfbeed040e4874b6d9ab0c229136f98101aa98", "submitted_at": "2024-09-18T16:04:10Z", "updated_at": "2024-10-06T21:25:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.12089v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2409.12089"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.12089"}], "name": "The Impact of Element Ordering on LM Agent Performance", "publication": null, "published_at": "2024-09-18", "record_sha256": "30dd2c60eeb806b73276dc7ec5c0cd3a8402be53bf706bac810e942df750c9f1", "relations": [], "resource_id": "arxiv:2409.12089", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.12089v3", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2024-10-06"}, "text": "The Impact of Element Ordering on LM Agent Performance\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how the ordering of UI elements in an agent observation affects web and desktop performance. A pixel-based element detector and dimensionality-reduction ordering are evaluated on OmniACT when document hierarchy is unavailable.\nPlatforms: web, desktop\nTopics: screen-understanding, grounding, evaluation\nPublished: 2024-09-18\nAuthors: Wayne Chi, Ameet Talwalkar, Chris Donahue\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.12089v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-12089/"}
{"id": "arxiv:2409.13711", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca44e6c98c09a2261b2cf8f3b6d18584b2d2471934b2dcac8d5b72c17b32e61", "submitted_at": "2024-09-06T18:44:25Z", "updated_at": "2024-09-24T18:38:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.13711v2"}, "authors": ["Maria Wang", "Srinivas Sunkara", "Gilles Baechler", "Jason Lin", "Yun Zhu", "Fedir Zubach", "Lei Shu", "Jindong Chen"], "canonical_source_url": "https://arxiv.org/abs/2409.13711", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-13711/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca44e6c98c09a2261b2cf8f3b6d18584b2d2471934b2dcac8d5b72c17b32e61", "submitted_at": "2024-09-06T18:44:25Z", "updated_at": "2024-09-24T18:38:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.13711v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.13711"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.13711"}], "name": "WebQuest: A Benchmark for Multimodal QA on Web Page Sequences", "publication": null, "published_at": "2024-09-06", "record_sha256": "bb6d34b3dd156190fa2f0e8cfaf1b9814c1cd8c06e8fd881e8dced45e804da42", "relations": [], "resource_id": "arxiv:2409.13711", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.13711v2", "topics": ["evaluation", "screen-understanding"], "updated_at": "2024-09-24"}, "text": "WebQuest: A Benchmark for Multimodal QA on Web Page Sequences\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces question answering over individual screens, multiple web pages, and navigation traces to test information synthesis across views.\nPlatforms: web\nTopics: evaluation, screen-understanding\nPublished: 2024-09-06\nAuthors: Maria Wang, Srinivas Sunkara, Gilles Baechler, Jason Lin, Yun Zhu, Fedir Zubach, Lei Shu, Jindong Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.13711v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-13711/"}
{"id": "arxiv:2409.14337", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bb00ad1be49c9bd51881827556b3a168d34c69a45e31053406f780ad2a62af1", "submitted_at": "2024-09-22T06:45:38Z", "updated_at": "2025-11-25T02:43:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.14337v3"}, "authors": ["Longxi Gao", "Li Zhang", "Shihe Wang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shangguang Wang", "Yuanchun Li", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2409.14337", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-14337/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bb00ad1be49c9bd51881827556b3a168d34c69a45e31053406f780ad2a62af1", "submitted_at": "2024-09-22T06:45:38Z", "updated_at": "2025-11-25T02:43:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2409.14337v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.14337"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.14337"}], "name": "MobileViews: A Million-scale and Diverse Mobile GUI Dataset", "publication": null, "published_at": "2024-09-22", "record_sha256": "5547b423f8fdb99840470a05fc50005226289d0c04703dc929e50ae1440d037f", "relations": [], "resource_id": "arxiv:2409.14337", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.14337v3", "topics": ["screen-understanding", "data-generation"], "updated_at": "2025-11-25"}, "text": "MobileViews: A Million-scale and Diverse Mobile GUI Dataset\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds mobile screenshot and view-hierarchy pairs through automated application traversal and evaluates their usefulness for GUI grounding.\nPlatforms: android\nTopics: screen-understanding, data-generation\nPublished: 2024-09-22\nAuthors: Longxi Gao, Li Zhang, Shihe Wang, Pengzhi Gao, Wei Liu, Jian Luan, Shangguang Wang, Yuanchun Li, Mengwei Xu\nRecorded arXiv version: v3\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.14337v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-14337/"}
{"id": "arxiv:2409.14818", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc22337ef29946dffcf7149df1008fd6cf02350c68d21ec0386e6485f5a14ebf", "submitted_at": "2024-09-23T08:47:54Z", "updated_at": "2024-10-03T05:23:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.14818v2"}, "authors": ["Qinzhuo Wu", "Weikai Xu", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Shuo Shang"], "canonical_source_url": "https://arxiv.org/abs/2409.14818", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-14818/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc22337ef29946dffcf7149df1008fd6cf02350c68d21ec0386e6485f5a14ebf", "submitted_at": "2024-09-23T08:47:54Z", "updated_at": "2024-10-03T05:23:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.14818v2"}, "publication": {"authors": ["Qinzhuo Wu", "Weikai Xu", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Shuo Shang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.599", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.599/", "https://aclanthology.org/2024.findings-emnlp.599.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. The landing page and its BibTeX list the fifth author as Liujian Liujianfeng; the PDF names Jianfeng Liu.", "pages": "10231-10251", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.599.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.599/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "c7f0e00d6102bac259b21a9422fba11f3da72da06535f37581b2670082c82a7b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.599/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000149+00:00", "role": "primary", "sha256": "131e26bc7168b529a0db7c4bda0747cdf318b7ef16bfaf3b4dede3826cb85c1a", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2024.findings-emnlp.599.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.14818"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.14818"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.599/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.599.pdf"}], "name": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding", "publication": {"authors": ["Qinzhuo Wu", "Weikai Xu", "Wei Liu", "Tao Tan", "Jianfeng Liu", "Ang Li", "Jian Luan", "Bin Wang", "Shuo Shang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.599", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.599/", "https://aclanthology.org/2024.findings-emnlp.599.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. The landing page and its BibTeX list the fifth author as Liujian Liujianfeng; the PDF names Jianfeng Liu.", "pages": "10231-10251", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.599.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.599/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}, "published_at": "2024-09-23", "record_sha256": "18608dc186407dc8486b49ed089e0cc81b5890a15281df895241f9661462fd61", "relations": [], "resource_id": "arxiv:2409.14818", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.14818v2", "topics": ["screen-understanding", "data-generation", "action-prediction"], "updated_at": "2024-10-03"}, "text": "MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nMobileVLM adds two pretraining stages and four UI tasks to learn element detail and relationships between successive screens. Its Mobile3M dataset contains three million Chinese mobile UI pages connected by recorded transition actions.\nPlatforms: unspecified\nTopics: screen-understanding, data-generation, action-prediction\nPublished: 2024-09-23\nAuthors: Qinzhuo Wu, Weikai Xu, Wei Liu, Tao Tan, Jianfeng Liu, Ang Li, Jian Luan, Bin Wang, Shuo Shang\nRecorded arXiv version: v2\n\nPublication edition: MobileVLM: A Vision-Language Model for Better Intra- and Inter-UI Understanding\nPublication authors: Qinzhuo Wu, Weikai Xu, Wei Liu, Tao Tan, Jianfeng Liu, Ang Li, Jian Luan, Bin Wang, Shuo Shang\nPublication edition date (recorded precision): 2024-11\nPublication venue: Findings of the Association for Computational Linguistics: EMNLP 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 10231-10251\nPublication doi: 10.18653/v1/2024.findings-emnlp.599\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. The landing page and its BibTeX list the fifth author as Liujian Liujianfeng; the PDF names Jianfeng Liu.\nPublication source: https://aclanthology.org/2024.findings-emnlp.599/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.14818v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-14818/"}
{"id": "arxiv:2409.15441", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "377d46c738f1a6f3e1c22ea99971a2fcfcf25ded82b2d2a24f728e614e22a1b3", "submitted_at": "2024-09-23T18:06:32Z", "updated_at": "2024-09-23T18:06:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.15441v1"}, "authors": ["Brian Tang", "Kang G. Shin"], "canonical_source_url": "https://arxiv.org/abs/2409.15441", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-15441/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "377d46c738f1a6f3e1c22ea99971a2fcfcf25ded82b2d2a24f728e614e22a1b3", "submitted_at": "2024-09-23T18:06:32Z", "updated_at": "2024-09-23T18:06:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2409.15441v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.15441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.15441"}], "name": "Steward: Natural Language Web Automation", "publication": null, "published_at": "2024-09-23", "record_sha256": "2c4b165d8a76e1463fcb041b5a5d1cb6bb8e4b72c64e8082e314d1b3d15fef17", "relations": [], "resource_id": "arxiv:2409.15441", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.15441v1", "topics": ["web-navigation", "automation", "efficiency"], "updated_at": "2024-09-23"}, "text": "Steward: Natural Language Web Automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Steward, a natural-language browser automation system that repeatedly plans and executes website actions. The paper examines completion detection, state representation, responsiveness, and caching in real-site tasks.\nPlatforms: web\nTopics: web-navigation, automation, efficiency\nPublished: 2024-09-23\nAuthors: Brian Tang, Kang G. Shin\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.15441v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-15441/"}
{"id": "arxiv:2409.15637", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c19b869b10f0b4c985b88b032f78e0e72f83e7ca8d9f346417054cd2a9002e3", "submitted_at": "2024-09-24T00:51:45Z", "updated_at": "2024-11-27T16:34:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.15637v2"}, "authors": ["Tianyue Ou", "Frank F. Xu", "Aman Madaan", "Jiarui Liu", "Robert Lo", "Abishek Sridhar", "Sudipta Sengupta", "Dan Roth", "Graham Neubig", "Shuyan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2409.15637", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-15637/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c19b869b10f0b4c985b88b032f78e0e72f83e7ca8d9f346417054cd2a9002e3", "submitted_at": "2024-09-24T00:51:45Z", "updated_at": "2024-11-27T16:34:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.15637v2"}, "publication": {"authors": ["Tianyue Ou", "Frank F. Xu", "Aman Madaan", "Jiarui Liu", "Robert Lo", "Abishek Sridhar", "Sudipta Sengupta", "Dan Roth", "Graham Neubig", "Shuyan Zhou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2908", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "91618-91652", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a6a6891cf1dfc64d664f086cf5976e93-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:04Z", "role": "primary", "sha256": "bf42d07f0959d7ad5ed0f4e00c34ef6cb7e8b1f10f17ac1e7a0ee1951523f821", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.15637"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.15637"}, {"rel": "publication", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a6a6891cf1dfc64d664f086cf5976e93-Paper-Conference.pdf"}], "name": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale", "publication": {"authors": ["Tianyue Ou", "Frank F. Xu", "Aman Madaan", "Jiarui Liu", "Robert Lo", "Abishek Sridhar", "Sudipta Sengupta", "Dan Roth", "Graham Neubig", "Shuyan Zhou"], "date_basis": "Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 12, 16], "doi": "10.52202/079017-2908", "event_dates": null, "evidence_urls": ["https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "91618-91652", "pdf_url": "https://proceedings.neurips.cc/paper_files/paper/2024/file/a6a6891cf1dfc64d664f086cf5976e93-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale", "type": "conference-paper", "url": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html", "venue": "Advances in Neural Information Processing Systems", "volume": "37"}, "published_at": "2024-09-24", "record_sha256": "ddebc42ea938156e0f2094dfc5849efca1c9b825f86c1610c311315c649acfe2", "relations": [], "resource_id": "arxiv:2409.15637", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.15637v2", "topics": ["data-generation", "imitation-learning", "web-navigation"], "updated_at": "2024-11-27"}, "text": "Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale\nResource type: paper\nReview status: curated\nScope: core\n\nSynatra converts indirect procedural knowledge, including human-oriented tutorials, into direct demonstrations for digital agents. It trains a 7B CodeLlama on 100,000 synthesized demonstrations and evaluates the resulting policy on Mind2Web, MiniWoB++, and WebArena.\nPlatforms: web\nTopics: data-generation, imitation-learning, web-navigation\nPublished: 2024-09-24\nAuthors: Tianyue Ou, Frank F. Xu, Aman Madaan, Jiarui Liu, Robert Lo, Abishek Sridhar, Sudipta Sengupta, Dan Roth, Graham Neubig, Shuyan Zhou\nRecorded arXiv version: v2\n\nPublication edition: Synatra: Turning Indirect Knowledge into Direct Demonstrations for Digital Agents at Scale\nPublication authors: Tianyue Ou, Frank F. Xu, Aman Madaan, Jiarui Liu, Robert Lo, Abishek Sridhar, Sudipta Sengupta, Dan Roth, Graham Neubig, Shuyan Zhou\nPublication edition date (recorded precision): 2024-12-16\nPublication venue: Advances in Neural Information Processing Systems\nPublication volume: 37\nPublication pages: 91618-91652\nPublication doi: 10.52202/079017-2908\nPublication date basis: Explicit publisher citation_publication_date: 2024-12-16. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.neurips.cc/paper_files/paper/2024/hash/a6a6891cf1dfc64d664f086cf5976e93-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.15637v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-15637/"}
{"id": "arxiv:2409.17140", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "618f41c2253a532780766b38396f241a48b587c580d6b4d7065d1519e6149751", "submitted_at": "2024-09-25T17:58:08Z", "updated_at": "2025-05-19T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.17140v2"}, "authors": ["Junting Lu", "Zhiyang Zhang", "Fangkai Yang", "Jue Zhang", "Lu Wang", "Chao Du", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2409.17140", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-17140/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "618f41c2253a532780766b38396f241a48b587c580d6b4d7065d1519e6149751", "submitted_at": "2024-09-25T17:58:08Z", "updated_at": "2025-05-19T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.17140v2"}, "publication": {"authors": ["Junting Lu", "Zhiyang Zhang", "Fangkai Yang", "Jue Zhang", "Lu Wang", "Chao Du", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.381", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.381/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "7711-7743", "pdf_url": "https://aclanthology.org/2025.acl-long.381.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.381/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:52Z", "role": "primary", "sha256": "31ab26306d5c41b25731db1ad7f73495330e637c1694c9da644f4e238a73917b", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.381/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {"arxiv": "2409.17140"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.17140"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.381/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.381.pdf"}], "name": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents", "publication": {"authors": ["Junting Lu", "Zhiyang Zhang", "Fangkai Yang", "Jue Zhang", "Lu Wang", "Chao Du", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.381", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.381/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "7711-7743", "pdf_url": "https://aclanthology.org/2025.acl-long.381.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.381/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-09-25", "record_sha256": "9704f7565cb9e4c9749809ec0f630c8734bea3f90d9c1b20e6e0b40740b38b3e", "relations": [], "resource_id": "arxiv:2409.17140", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2409.17140v2", "topics": ["tool-integration", "automation", "efficiency"], "updated_at": "2025-05-19"}, "text": "AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AXIS, an API-first application agent that can use UI actions and expand its available APIs through automated application exploration. The work evaluates the approach on Microsoft Word tasks.\nPlatforms: unspecified\nTopics: tool-integration, automation, efficiency\nPublished: 2024-09-25\nAuthors: Junting Lu, Zhiyang Zhang, Fangkai Yang, Jue Zhang, Lu Wang, Chao Du, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nRecorded arXiv version: v2\n\nPublication edition: AXIS: Efficient Human-Agent-Computer Interaction with API-First LLM-Based Agents\nPublication authors: Junting Lu, Zhiyang Zhang, Fangkai Yang, Jue Zhang, Lu Wang, Chao Du, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 7711-7743\nPublication doi: 10.18653/v1/2025.acl-long.381\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.381/\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2409.17140v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-17140/"}
{"id": "arxiv:2409.18060", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d927b7644f0c6b3a49fbdf1aa976a00f2409adf9a72e1c04333438f66f1014b", "submitted_at": "2024-09-26T17:01:33Z", "updated_at": "2024-10-07T19:04:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.18060v2"}, "authors": ["Sabrina Haque", "Christoph Csallner"], "canonical_source_url": "https://arxiv.org/abs/2409.18060", "citation_url": "https://awesomecua.github.io/resources/arxiv-2409-18060/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d927b7644f0c6b3a49fbdf1aa976a00f2409adf9a72e1c04333438f66f1014b", "submitted_at": "2024-09-26T17:01:33Z", "updated_at": "2024-10-07T19:04:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2409.18060v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2409.18060"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2409.18060"}], "name": "Inferring Alt-text For UI Icons With Large Language Models During App Development", "publication": null, "published_at": "2024-09-26", "record_sha256": "e6499e3e82a7301ca6d0eee011c45e2ae9ef1d14cfcb029f650061d1f9299013", "relations": [], "resource_id": "arxiv:2409.18060", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2409.18060v2", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": "2024-10-07"}, "text": "Inferring Alt-text For UI Icons With Large Language Models During App Development\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces IconDesc, which generates mobile icon alt-text from partial UI context such as element identifiers, bounds, nearby text and hierarchy. Model evaluation and a user study examine the resulting accessibility descriptions.\nPlatforms: unspecified\nTopics: screen-understanding, ui-design, evaluation\nPublished: 2024-09-26\nAuthors: Sabrina Haque, Christoph Csallner\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. This is a UI accessibility and element-description method; autonomous GUI control is not evaluated.\n\nPrimary source: https://arxiv.org/abs/2409.18060v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2409-18060/"}
{"id": "arxiv:2410.00079", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4c873d1dd443ffc18fc4b63595fdfc726eb84388eddbfced0dcde67c5e19a7", "submitted_at": "2024-09-30T16:52:51Z", "updated_at": "2024-09-30T16:52:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00079v1"}, "authors": ["Wenyue Hua", "Mengting Wan", "Shashank Vadrevu", "Ryan Nadel", "Yongfeng Zhang", "Chi Wang"], "canonical_source_url": "https://arxiv.org/abs/2410.00079", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-00079/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4c873d1dd443ffc18fc4b63595fdfc726eb84388eddbfced0dcde67c5e19a7", "submitted_at": "2024-09-30T16:52:51Z", "updated_at": "2024-09-30T16:52:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00079v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2410.00079"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00079"}], "name": "Interactive Speculative Planning: Enhance Agent Efficiency through Co-design of System and User Interface", "publication": null, "published_at": "2024-09-30", "record_sha256": "2ae6acb7f71320829bf141d472f7b3ebeef198698507673a6532023499bb1036", "relations": [], "resource_id": "arxiv:2410.00079", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2410.00079v1", "topics": ["planning", "efficiency", "human-in-the-loop"], "updated_at": "2024-09-30"}, "text": "Interactive Speculative Planning: Enhance Agent Efficiency through Co-design of System and User Interface\nResource type: paper\nReview status: curated\nScope: adjacent\n\nStudies agent-system and interface co-design that incorporates user interruptions and intermediate feedback to reduce planning latency.\nPlatforms: unspecified\nTopics: planning, efficiency, human-in-the-loop\nPublished: 2024-09-30\nAuthors: Wenyue Hua, Mengting Wan, Shashank Vadrevu, Ryan Nadel, Yongfeng Zhang, Chi Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.00079v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-00079/"}
{"id": "arxiv:2410.00201", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85dca895f00b442a60123c7c85cd26a90e3d1c2a810c91c32385f877bcd755a", "submitted_at": "2024-09-30T19:55:54Z", "updated_at": "2024-09-30T19:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00201v1"}, "authors": ["Yi-Hao Peng", "Faria Huq", "Yue Jiang", "Jason Wu", "Amanda Xin Yue Li", "Jeffrey Bigham", "Amy Pavel"], "canonical_source_url": "https://arxiv.org/abs/2410.00201", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-00201/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85dca895f00b442a60123c7c85cd26a90e3d1c2a810c91c32385f877bcd755a", "submitted_at": "2024-09-30T19:55:54Z", "updated_at": "2024-09-30T19:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.00201v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2410.00201"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00201"}], "name": "DreamStruct: Understanding Slides and User Interfaces via Synthetic Data Generation", "publication": null, "published_at": "2024-09-30", "record_sha256": "6440538471d649132f80b4a32f33cd746e35b0b2c8667c88869105ffc7523a1c", "relations": [], "resource_id": "arxiv:2410.00201", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.00201v1", "topics": ["data-generation", "screen-understanding"], "updated_at": "2024-09-30"}, "text": "DreamStruct: Understanding Slides and User Interfaces via Synthetic Data Generation\nResource type: paper\nReview status: curated\nScope: core\n\nGenerates labeled slides and interface visuals through code to support visual element detection, description, and classification with less manual annotation.\nPlatforms: cross-platform\nTopics: data-generation, screen-understanding\nPublished: 2024-09-30\nAuthors: Yi-Hao Peng, Faria Huq, Yue Jiang, Jason Wu, Amanda Xin Yue Li, Jeffrey Bigham, Amy Pavel\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.00201v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-00201/"}
{"id": "arxiv:2410.00467", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f370d105dc1fc1a377fd1b1fd5d1fc3d0f035694cc7c364e6326c587607c98c6", "submitted_at": "2024-10-01T07:49:24Z", "updated_at": "2024-12-19T14:29:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.00467v3"}, "authors": ["Shaoqing Zhang", "Zhuosheng Zhang", "Kehai Chen", "Xinbei Ma", "Muyun Yang", "Tiejun Zhao", "Min Zhang"], "canonical_source_url": "https://arxiv.org/abs/2410.00467", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-00467/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f370d105dc1fc1a377fd1b1fd5d1fc3d0f035694cc7c364e6326c587607c98c6", "submitted_at": "2024-10-01T07:49:24Z", "updated_at": "2024-12-19T14:29:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.00467v3"}, "publication": {"authors": ["Shaoqing Zhang", "Zhuosheng Zhang", "Kehai Chen (陈科海)", "Xinbei Ma", "Muyun Yang (杨沐昀)", "Tiejun Zhao (赵铁军)", "Min Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.70", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.70/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1304-1320", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.70.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Dynamic Planning for LLM-based Graphical User Interface Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.70/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "1d89601c2a7a81756b9e7441c48a312132f385e5370780507d27c9848aa1176d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2024.findings-emnlp.70/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.00467"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00467"}, {"rel": "publication", "url": "https://aclanthology.org/2024.findings-emnlp.70/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2024.findings-emnlp.70.pdf"}], "name": "Dynamic Planning for LLM-based Graphical User Interface Automation", "publication": {"authors": ["Shaoqing Zhang", "Zhuosheng Zhang", "Kehai Chen (陈科海)", "Xinbei Ma", "Muyun Yang (杨沐昀)", "Tiejun Zhao (赵铁军)", "Min Zhang"], "date_basis": "Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2024, 11], "doi": "10.18653/v1/2024.findings-emnlp.70", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2024.findings-emnlp.70/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "1304-1320", "pdf_url": "https://aclanthology.org/2024.findings-emnlp.70.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Dynamic Planning for LLM-based Graphical User Interface Automation", "type": "conference-paper", "url": "https://aclanthology.org/2024.findings-emnlp.70/", "venue": "Findings of the Association for Computational Linguistics: EMNLP 2024", "volume": null}, "published_at": "2024-10-01", "record_sha256": "93b3ddd9b3622133cdf49cb2b59f1c5ef8989fc818e31beebea52ee73de682c0", "relations": [], "resource_id": "arxiv:2410.00467", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.00467v3", "topics": ["planning", "interaction-history", "mobile-navigation", "robustness"], "updated_at": "2024-12-19"}, "text": "Dynamic Planning for LLM-based Graphical User Interface Automation\nResource type: paper\nReview status: curated\nScope: core\n\nDynamic Planning of Thoughts revises a mobile GUI plan using current environmental feedback and execution history. The study compares this planning strategy with long-dialogue prompting and examines its behavior on unfamiliar tasks and different language-model backbones.\nPlatforms: unspecified\nTopics: planning, interaction-history, mobile-navigation, robustness\nPublished: 2024-10-01\nAuthors: Shaoqing Zhang, Zhuosheng Zhang, Kehai Chen, Xinbei Ma, Muyun Yang, Tiejun Zhao, Min Zhang\nRecorded arXiv version: v3\n\nPublication edition: Dynamic Planning for LLM-based Graphical User Interface Automation\nPublication authors: Shaoqing Zhang, Zhuosheng Zhang, Kehai Chen (陈科海), Xinbei Ma, Muyun Yang (杨沐昀), Tiejun Zhao (赵铁军), Min Zhang\nPublication edition date (recorded precision): 2024-11\nPublication venue: Findings of the Association for Computational Linguistics: EMNLP 2024\nPublication publisher: Association for Computational Linguistics\nPublication pages: 1304-1320\nPublication doi: 10.18653/v1/2024.findings-emnlp.70\nPublication date basis: Explicit publisher citation_publication_date: 2024/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2024.findings-emnlp.70/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.00467v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-00467/"}
{"id": "arxiv:2410.00689", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69febc7aabfe38a6da439f1b4f673e0c36bc34c6afd1e8b06481433c1f43b1aa", "submitted_at": "2024-10-01T13:43:55Z", "updated_at": "2024-10-11T15:42:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.00689v2"}, "authors": ["Ruhana Azam", "Tamer Abuelsaad", "Aditya Vempaty", "Ashish Jagmohan"], "canonical_source_url": "https://arxiv.org/abs/2410.00689", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-00689/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69febc7aabfe38a6da439f1b4f673e0c36bc34c6afd1e8b06481433c1f43b1aa", "submitted_at": "2024-10-01T13:43:55Z", "updated_at": "2024-10-11T15:42:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.00689v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.00689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.00689"}], "name": "Multimodal Auto Validation For Self-Refinement in Web Agents", "publication": null, "published_at": "2024-10-01", "record_sha256": "0cc3c9ba7b3549ba7e0515f70e173bde82fc4645f1d85f0b87a06d451d9d8bac", "relations": [], "resource_id": "arxiv:2410.00689", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.00689v2", "topics": ["evaluation", "robustness", "web-navigation"], "updated_at": "2024-10-11"}, "text": "Multimodal Auto Validation For Self-Refinement in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies text and visual validation for web agents and uses the resulting validator to detect failures and guide workflow self-refinement.\nPlatforms: web\nTopics: evaluation, robustness, web-navigation\nPublished: 2024-10-01\nAuthors: Ruhana Azam, Tamer Abuelsaad, Aditya Vempaty, Ashish Jagmohan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.00689v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-00689/"}
{"id": "arxiv:2410.02810", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a2e16e978f50c8d7161a11a11fba6aeda9617663de1746422151d38c4874cc9", "submitted_at": "2024-09-21T05:54:35Z", "updated_at": "2025-04-08T06:37:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.02810v3"}, "authors": ["Nikolai Rozanov", "Marek Rei"], "canonical_source_url": "https://arxiv.org/abs/2410.02810", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-02810/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a2e16e978f50c8d7161a11a11fba6aeda9617663de1746422151d38c4874cc9", "submitted_at": "2024-09-21T05:54:35Z", "updated_at": "2025-04-08T06:37:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.02810v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.02810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.02810"}], "name": "StateAct: Enhancing LLM Base Agents via Self-prompting and State-tracking", "publication": null, "published_at": "2024-09-21", "record_sha256": "4981ed2934c190e8d6e854d7ecf8c8d9961169d0ee19bdc4bb2430d568c8209b", "relations": [], "resource_id": "arxiv:2410.02810", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2410.02810v3", "topics": ["planning", "memory"], "updated_at": "2025-04-08"}, "text": "StateAct: Enhancing LLM Base Agents via Self-prompting and State-tracking\nResource type: paper\nReview status: curated\nScope: adjacent\n\nMaintains explicit task goals and evolving state descriptions to improve base-agent decisions across interactive tasks, including web shopping.\nPlatforms: web\nTopics: planning, memory\nPublished: 2024-09-21\nAuthors: Nikolai Rozanov, Marek Rei\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.02810v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-02810/"}
{"id": "arxiv:2410.02907", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e227f3686b3234745ca1749cb8fb7c2b19383ac9dfebf68976f61f07d1ddf3b", "submitted_at": "2024-10-03T18:56:51Z", "updated_at": "2025-02-05T18:56:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.02907v2"}, "authors": ["Shikhar Murty", "Hao Zhu", "Dzmitry Bahdanau", "Christopher D. Manning"], "canonical_source_url": "https://arxiv.org/abs/2410.02907", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-02907/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e227f3686b3234745ca1749cb8fb7c2b19383ac9dfebf68976f61f07d1ddf3b", "submitted_at": "2024-10-03T18:56:51Z", "updated_at": "2025-02-05T18:56:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.02907v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.02907"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.02907"}], "name": "NNetNav: Unsupervised Learning of Browser Agents Through Environment Interaction in the Wild", "publication": null, "published_at": "2024-10-03", "record_sha256": "379720df46453f88cdd7d322fc261210a388affd5782c44e8a5dc511a58686f1", "relations": [], "resource_id": "arxiv:2410.02907", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.02907v2", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-02-05"}, "text": "NNetNav: Unsupervised Learning of Browser Agents Through Environment Interaction in the Wild\nResource type: paper\nReview status: curated\nScope: core\n\nGenerates browser-agent demonstrations through autonomous exploration and retrospective task labeling, pruning trajectories that lack meaningful subtasks.\nPlatforms: web\nTopics: data-generation, web-navigation\nPublished: 2024-10-03\nAuthors: Shikhar Murty, Hao Zhu, Dzmitry Bahdanau, Christopher D. Manning\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.02907v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-02907/"}
{"id": "arxiv:2410.05243", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ece826d135b9ab9dfed2a3baf9deeab99489120da749cd67517b636649c5a6", "submitted_at": "2024-10-07T17:47:50Z", "updated_at": "2025-06-17T15:06:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.05243v3"}, "authors": ["Boyu Gou", "Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2410.05243", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-05243/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ece826d135b9ab9dfed2a3baf9deeab99489120da749cd67517b636649c5a6", "submitted_at": "2024-10-07T17:47:50Z", "updated_at": "2025-06-17T15:06:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.05243v3"}, "publication": {"authors": ["Boyu Gou", "Demi Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "30851-30883", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "4e7384659700d8bc8e1f56de671c0d943f15b8c682bdff0fbfb20f41c707b7ff", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.05243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.05243"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf"}], "name": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "publication": {"authors": ["Boyu Gou", "Demi Ruohan Wang", "Boyuan Zheng", "Yanan Xie", "Cheng Chang", "Yiheng Shu", "Huan Sun", "Yu Su"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "30851-30883", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/4ca0e369689dadb25a5345ba9755ad6f-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-07", "record_sha256": "eaa07790cff437e918a2b541e83c747edecd6eb865feb37e317fcf9b84bc8778", "relations": [], "resource_id": "arxiv:2410.05243", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.05243v3", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-17"}, "text": "Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UGround, a visual target-localization model trained with large-scale synthetic references to elements in GUI screenshots.\nPlatforms: cross-platform\nTopics: grounding, data-generation\nPublished: 2024-10-07\nAuthors: Boyu Gou, Ruohan Wang, Boyuan Zheng, Yanan Xie, Cheng Chang, Yiheng Shu, Huan Sun, Yu Su\nRecorded arXiv version: v3\n\nPublication edition: Navigating the Digital World as Humans Do: Universal Visual Grounding for GUI Agents\nPublication authors: Boyu Gou, Demi Ruohan Wang, Boyuan Zheng, Yanan Xie, Cheng Chang, Yiheng Shu, Huan Sun, Yu Su\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 30851-30883\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/4ca0e369689dadb25a5345ba9755ad6f-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.05243v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-05243/"}
{"id": "arxiv:2410.06703", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ba2b599c2b3f09974ee41d329545b69773a3b408b6db0124c5b1073d5fe8f1cf", "submitted_at": "2024-10-09T09:13:38Z", "updated_at": "2026-06-04T09:59:28Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2410.06703v7"}, "authors": ["Ido Levy", "Ben Wiesel", "Sami Marreed", "Alon Oved", "Avi Yaeli", "Nir Mashkif", "Segev Shlomov"], "canonical_source_url": "https://arxiv.org/abs/2410.06703", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-06703/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ba2b599c2b3f09974ee41d329545b69773a3b408b6db0124c5b1073d5fe8f1cf", "submitted_at": "2024-10-09T09:13:38Z", "updated_at": "2026-06-04T09:59:28Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2410.06703v7"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.06703"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.06703"}], "name": "ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents", "publication": null, "published_at": "2024-10-09", "record_sha256": "33e44ecc5b86b2b2534dda161ab3f88a45f4e7fb155caa1fd0fccaabbe1313a3", "relations": [], "resource_id": "arxiv:2410.06703", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.06703v7", "topics": ["safety-security", "evaluation"], "updated_at": "2026-06-04"}, "text": "ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates web-task completion under explicit enterprise policies, measuring policy-compliant completion and safety violations separately.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2024-10-09\nAuthors: Ido Levy, Ben Wiesel, Sami Marreed, Alon Oved, Avi Yaeli, Nir Mashkif, Segev Shlomov\nRecorded arXiv version: v7\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.06703v7\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-06703/"}
{"id": "arxiv:2410.08164", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "284888d16b049767f76448f23f5d35fda25b109fd67f450105950173ae81c97b", "submitted_at": "2024-10-10T17:43:51Z", "updated_at": "2024-10-10T17:43:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.08164v1"}, "authors": ["Saaket Agashe", "Jiuzhou Han", "Shuyu Gan", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2410.08164", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-08164/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "284888d16b049767f76448f23f5d35fda25b109fd67f450105950173ae81c97b", "submitted_at": "2024-10-10T17:43:51Z", "updated_at": "2024-10-10T17:43:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.08164v1"}, "publication": {"authors": ["Saaket Agashe", "Jiuzhou Han", "Shuyu Gan", "Jiachen Yang", "Ang Li", "Xin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "22924-22946", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/394c7c30ea87b5c3521b4d9e9d419071-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Agent S: An Open Agentic Framework that Uses Computers Like a Human", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:01Z", "role": "primary", "sha256": "032cfd2f08acc7f85921748e0546bf6e16d038da412c207105ca9f6d433c70b3", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.08164"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.08164"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/394c7c30ea87b5c3521b4d9e9d419071-Paper-Conference.pdf"}], "name": "Agent S: An Open Agentic Framework that Uses Computers Like a Human", "publication": {"authors": ["Saaket Agashe", "Jiuzhou Han", "Shuyu Gan", "Jiachen Yang", "Ang Li", "Xin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "22924-22946", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/394c7c30ea87b5c3521b4d9e9d419071-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Agent S: An Open Agentic Framework that Uses Computers Like a Human", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-10", "record_sha256": "0a1b3e7664b06237fa094dde85515ea3c2ede368f1e6b4605f477f923bfdde25", "relations": [], "resource_id": "arxiv:2410.08164", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.08164v1", "topics": ["planning", "memory"], "updated_at": "2024-10-10"}, "text": "Agent S: An Open Agentic Framework that Uses Computers Like a Human\nResource type: paper\nReview status: curated\nScope: core\n\nCombines hierarchical planning, retrieved experience, and a computer interaction interface to execute multistep GUI tasks.\nPlatforms: desktop\nTopics: planning, memory\nPublished: 2024-10-10\nAuthors: Saaket Agashe, Jiuzhou Han, Shuyu Gan, Jiachen Yang, Ang Li, Xin Eric Wang\nRecorded arXiv version: v1\n\nPublication edition: Agent S: An Open Agentic Framework that Uses Computers Like a Human\nPublication authors: Saaket Agashe, Jiuzhou Han, Shuyu Gan, Jiachen Yang, Ang Li, Xin Wang\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 22924-22946\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/394c7c30ea87b5c3521b4d9e9d419071-Abstract-Conference.html\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.08164v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-08164/"}
{"id": "arxiv:2410.09006", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff649a0792864bd40b7bf2f825f6a8a815eb60cc795ed4dc0c6b55a942477758", "submitted_at": "2024-10-11T17:24:00Z", "updated_at": "2025-03-22T18:01:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.09006v2"}, "authors": ["Zhuohao Jerry Zhang", "Eldon Schoop", "Jeffrey Nichols", "Anuj Mahajan", "Amanda Swearngin"], "canonical_source_url": "https://arxiv.org/abs/2410.09006", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-09006/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff649a0792864bd40b7bf2f825f6a8a815eb60cc795ed4dc0c6b55a942477758", "submitted_at": "2024-10-11T17:24:00Z", "updated_at": "2025-03-22T18:01:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.09006v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.09006"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.09006"}], "name": "From Interaction to Impact: Towards Safer AI Agents Through Understanding and Evaluating Mobile UI Operation Impacts", "publication": null, "published_at": "2024-10-11", "record_sha256": "fdee754a6d63d3fc41e81f26a1769c53a1f195500e74baf3f7cc085f9f276394", "relations": [], "resource_id": "arxiv:2410.09006", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.09006v2", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-03-22"}, "text": "From Interaction to Impact: Towards Safer AI Agents Through Understanding and Evaluating Mobile UI Operation Impacts\nResource type: paper\nReview status: curated\nScope: core\n\nDevelops a taxonomy of mobile-action consequences and annotated interaction data to evaluate whether models recognize impactful or risky UI operations.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2024-10-11\nAuthors: Zhuohao Jerry Zhang, Eldon Schoop, Jeffrey Nichols, Anuj Mahajan, Amanda Swearngin\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.09006v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-09006/"}
{"id": "arxiv:2410.11871", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f1594a51709f2a1283d3ac14734f08fa7894b436b85bafbbbe9b2eaba332129", "submitted_at": "2024-10-09T12:06:43Z", "updated_at": "2025-05-21T06:52:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.11871v3"}, "authors": ["Pawel Pawlowski", "Krystian Zawistowski", "Wojciech Lapacz", "Adam Wiacek", "Marcin Skorupa", "Sebastien Postansque", "Jakub Hoscilowicz"], "canonical_source_url": "https://arxiv.org/abs/2410.11871", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-11871/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f1594a51709f2a1283d3ac14734f08fa7894b436b85bafbbbe9b2eaba332129", "submitted_at": "2024-10-09T12:06:43Z", "updated_at": "2025-05-21T06:52:31Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.11871v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.11871"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11871"}], "name": "TinyClick: Single-Turn Agent for Empowering GUI Automation", "publication": null, "published_at": "2024-10-09", "record_sha256": "8875afaf1d24fa5680cf5f4a804acb3530eee336ca8d2a7bee18dc602b0e6ef4", "relations": [], "resource_id": "arxiv:2410.11871", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.11871v3", "topics": ["grounding", "efficiency"], "updated_at": "2025-05-21"}, "text": "TinyClick: Single-Turn Agent for Empowering GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts a compact Florence-2 model to map user instructions to screen coordinates using multitask training and augmented data.\nPlatforms: cross-platform\nTopics: grounding, efficiency\nPublished: 2024-10-09\nAuthors: Pawel Pawlowski, Krystian Zawistowski, Wojciech Lapacz, Adam Wiacek, Marcin Skorupa, Sebastien Postansque, Jakub Hoscilowicz\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.11871v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-11871/"}
{"id": "arxiv:2410.11872", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e804944da205076ac6cf88c7966c471697e526f657bda339e09df240257f7217", "submitted_at": "2024-10-09T14:49:02Z", "updated_at": "2024-10-17T07:12:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.11872v2"}, "authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "canonical_source_url": "https://arxiv.org/abs/2410.11872", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-11872/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e804944da205076ac6cf88c7966c471697e526f657bda339e09df240257f7217", "submitted_at": "2024-10-09T14:49:02Z", "updated_at": "2024-10-17T07:12:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.11872v2"}, "publication": {"authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "date_basis": "Explicit publisher citation_publication_date: 2025/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 8], "doi": "10.18653/v1/2025.sigdial-1.38", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.sigdial-1.38/", "https://aclanthology.org/2025.sigdial-1.38.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the five-author list on page 1 of the publisher PDF. Landing-page citation metadata lists only Jakub Hoscilowicz and Artur Janicki.", "pages": "471-476", "pdf_url": "https://aclanthology.org/2025.sigdial-1.38.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.sigdial-1.38/", "venue": "Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "7c710272d48754149cc5947f5382b8123eac5855db7655410259c4a3129b8b5b", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:09.676954+00:00", "role": "primary", "sha256": "c5b5cd07e3ca2f2911e183c4752e01c721971c47fd20c836605452f2b80e5260", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.11872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.11872"}, {"rel": "publication", "url": "https://aclanthology.org/2025.sigdial-1.38/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.sigdial-1.38.pdf"}], "name": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "publication": {"authors": ["Jakub Hoscilowicz", "Bartosz Maj", "Bartosz Kozakiewicz", "Oleksii Tymoshchuk", "Artur Janicki"], "date_basis": "Explicit publisher citation_publication_date: 2025/8. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 8], "doi": "10.18653/v1/2025.sigdial-1.38", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.sigdial-1.38/", "https://aclanthology.org/2025.sigdial-1.38.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the five-author list on page 1 of the publisher PDF. Landing-page citation metadata lists only Jakub Hoscilowicz and Artur Janicki.", "pages": "471-476", "pdf_url": "https://aclanthology.org/2025.sigdial-1.38.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.sigdial-1.38/", "venue": "Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue", "volume": null}, "published_at": "2024-10-09", "record_sha256": "a427a7aa3800366ffa34eb5377208ebfb8fee794201eee0a648d91e3c582c06f", "relations": [], "resource_id": "arxiv:2410.11872", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.11872v2", "topics": ["mobile-navigation", "grounding"], "updated_at": "2024-10-17"}, "text": "ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates multimodal reasoning and action planning from a dedicated UI localization model, with evaluation on Android devices.\nPlatforms: android\nTopics: mobile-navigation, grounding\nPublished: 2024-10-09\nAuthors: Jakub Hoscilowicz, Bartosz Maj, Bartosz Kozakiewicz, Oleksii Tymoshchuk, Artur Janicki\nRecorded arXiv version: v2\n\nPublication edition: ClickAgent: Enhancing UI Location Capabilities of Autonomous Agents\nPublication authors: Jakub Hoscilowicz, Bartosz Maj, Bartosz Kozakiewicz, Oleksii Tymoshchuk, Artur Janicki\nPublication edition date (recorded precision): 2025-08\nPublication venue: Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue\nPublication publisher: Association for Computational Linguistics\nPublication pages: 471-476\nPublication doi: 10.18653/v1/2025.sigdial-1.38\nPublication date basis: Explicit publisher citation_publication_date: 2025/8. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the five-author list on page 1 of the publisher PDF. Landing-page citation metadata lists only Jakub Hoscilowicz and Artur Janicki.\nPublication source: https://aclanthology.org/2025.sigdial-1.38/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.11872v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-11872/"}
{"id": "arxiv:2410.13232", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cae86bff4412bbfa7b969ecd60fe7b695f4b7c5ea1be6a67e97919df762d1108", "submitted_at": "2024-10-17T05:37:00Z", "updated_at": "2025-03-29T08:59:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13232v2"}, "authors": ["Hyungjoo Chae", "Namyoung Kim", "Kai Tzu-iunn Ong", "Minju Gwak", "Gwanwoo Song", "Jihoon Kim", "Sunghwan Kim", "Dongha Lee", "Jinyoung Yeo"], "canonical_source_url": "https://arxiv.org/abs/2410.13232", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-13232/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cae86bff4412bbfa7b969ecd60fe7b695f4b7c5ea1be6a67e97919df762d1108", "submitted_at": "2024-10-17T05:37:00Z", "updated_at": "2025-03-29T08:59:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13232v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.13232"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13232"}], "name": "Web Agents with World Models: Learning and Leveraging Environment Dynamics in Web Navigation", "publication": null, "published_at": "2024-10-17", "record_sha256": "14d357878e8f5ebe870cb9a73a79ccabb21e5df1b234436a16c76ea47862e0ec", "relations": [], "resource_id": "arxiv:2410.13232", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.13232v2", "topics": ["world-models", "planning"], "updated_at": "2025-03-29"}, "text": "Web Agents with World Models: Learning and Leveraging Environment Dynamics in Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nUses a learned web world model to predict important state changes and compare candidate actions before executing them.\nPlatforms: web\nTopics: world-models, planning\nPublished: 2024-10-17\nAuthors: Hyungjoo Chae, Namyoung Kim, Kai Tzu-iunn Ong, Minju Gwak, Gwanwoo Song, Jihoon Kim, Sunghwan Kim, Dongha Lee, Jinyoung Yeo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.13232v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-13232/"}
{"id": "arxiv:2410.13757", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0ef02f9c31bf2f095330cf018e7525aecc7644eb1552fb6b511ffced3889122", "submitted_at": "2024-10-17T16:53:50Z", "updated_at": "2025-05-13T06:25:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13757v3"}, "authors": ["Zichen Zhu", "Hao Tang", "Yansi Li", "Dingye Liu", "Hongshen Xu", "Kunyao Lan", "Danyang Zhang", "Yixuan Jiang", "Hao Zhou", "Chenrun Wang", "Situo Zhang", "Liangtai Sun", "Yixiao Wang", "Yuheng Sun", "Lu Chen", "Kai Yu"], "canonical_source_url": "https://arxiv.org/abs/2410.13757", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-13757/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0ef02f9c31bf2f095330cf018e7525aecc7644eb1552fb6b511ffced3889122", "submitted_at": "2024-10-17T16:53:50Z", "updated_at": "2025-05-13T06:25:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13757v3"}, "publication": {"authors": ["Zichen Zhu", "Hao Tang", "Yansi Li", "Dingye Liu", "Hongshen Xu", "Kunyao Lan", "Danyang Zhang", "Yixuan Jiang", "Hao Zhou", "Chenrun Wang", "Situo Zhang", "Liangtai Sun", "Yixiao Wang", "Yuheng Sun", "Lu Chen", "Kai Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.43", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.43/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "535-549", "pdf_url": "https://aclanthology.org/2025.naacl-demo.43.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.43/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:53Z", "role": "primary", "sha256": "a873b34264be14f64599c0897219ff2796ab74d111df8a2fdc667fd22ad87028", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-demo.43/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.13757"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13757"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.43/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.43.pdf"}], "name": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation", "publication": {"authors": ["Zichen Zhu", "Hao Tang", "Yansi Li", "Dingye Liu", "Hongshen Xu", "Kunyao Lan", "Danyang Zhang", "Yixuan Jiang", "Hao Zhou", "Chenrun Wang", "Situo Zhang", "Liangtai Sun", "Yixiao Wang", "Yuheng Sun", "Lu Chen", "Kai Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.43", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.43/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "535-549", "pdf_url": "https://aclanthology.org/2025.naacl-demo.43.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.43/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}, "published_at": "2024-10-17", "record_sha256": "aee833a27ac08bb23ee529ef1de27e3e68c3a0db609517acc0298d57dac4820d", "relations": [], "resource_id": "arxiv:2410.13757", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.13757v3", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2025-05-13"}, "text": "MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation\nResource type: paper\nReview status: curated\nScope: core\n\nCombines adaptive mobile-task planning, reflection-based recovery, and multiple memory components, evaluated with MobBench and AndroidArena.\nPlatforms: android\nTopics: memory, planning, mobile-navigation\nPublished: 2024-10-17\nAuthors: Zichen Zhu, Hao Tang, Yansi Li, Dingye Liu, Hongshen Xu, Kunyao Lan, Danyang Zhang, Yixuan Jiang, Hao Zhou, Chenrun Wang, Situo Zhang, Liangtai Sun, Yixiao Wang, Yuheng Sun, Lu Chen, Kai Yu\nRecorded arXiv version: v3\n\nPublication edition: MobA: Multifaceted Memory-Enhanced Adaptive Planning for Efficient Mobile Task Automation\nPublication authors: Zichen Zhu, Hao Tang, Yansi Li, Dingye Liu, Hongshen Xu, Kunyao Lan, Danyang Zhang, Yixuan Jiang, Hao Zhou, Chenrun Wang, Situo Zhang, Liangtai Sun, Yixiao Wang, Yuheng Sun, Lu Chen, Kai Yu\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 535-549\nPublication doi: 10.18653/v1/2025.naacl-demo.43\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.naacl-demo.43/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.13757v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-13757/"}
{"id": "arxiv:2410.13824", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a3bcd26dff4f1d21a8a2772929c0c97cd1cc180a7606b0744e630dabae94609", "submitted_at": "2024-10-17T17:48:54Z", "updated_at": "2024-11-06T08:29:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13824v3"}, "authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "canonical_source_url": "https://arxiv.org/abs/2410.13824", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-13824/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a3bcd26dff4f1d21a8a2772929c0c97cd1cc180a7606b0744e630dabae94609", "submitted_at": "2024-10-17T17:48:54Z", "updated_at": "2024-11-06T08:29:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.13824v3"}, "publication": {"authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "17971-18005", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "899100e1d0cad8d949c9075f62ab3506995587b01aab3981ac4e26eaf6d88339", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.13824"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13824"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf"}], "name": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "publication": {"authors": ["Junpeng Liu", "Tianyue Ou", "Yifan Song", "Yuxiao Qu", "Wai Lam", "Chenyan Xiong", "Wenhu Chen", "Graham Neubig", "Xiang Yue"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "17971-18005", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/2da53cd1abdae59150e35f4693834f32-Paper-Conference.pdf", "publisher": null, "series": null, "title": "Harnessing Webpage UIs for Text-Rich Visual Understanding", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-17", "record_sha256": "44e2afc613865bd89cae59a91125781d682919eaf8ef58fe537b2e1d0cb89f24", "relations": [], "resource_id": "arxiv:2410.13824", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.13824v3", "topics": ["data-generation", "screen-understanding", "grounding"], "updated_at": "2024-11-06"}, "text": "Harnessing Webpage UIs for Text-Rich Visual Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds MultiUI by generating instructions from webpage accessibility trees and pairing them with screenshots for multimodal training. Evaluation examines web-interface understanding, including VisualWebBench and Mind2Web, and transfer to other text-rich visual tasks.\nPlatforms: web\nTopics: data-generation, screen-understanding, grounding\nPublished: 2024-10-17\nAuthors: Junpeng Liu, Tianyue Ou, Yifan Song, Yuxiao Qu, Wai Lam, Chenyan Xiong, Wenhu Chen, Graham Neubig, Xiang Yue\nRecorded arXiv version: v3\n\nPublication edition: Harnessing Webpage UIs for Text-Rich Visual Understanding\nPublication authors: Junpeng Liu, Tianyue Ou, Yifan Song, Yuxiao Qu, Wai Lam, Chenyan Xiong, Wenhu Chen, Graham Neubig, Xiang Yue\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 17971-18005\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/2da53cd1abdae59150e35f4693834f32-Abstract-Conference.html\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.13824v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-13824/"}
{"id": "arxiv:2410.13825", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09268d51b3d810ffc0024ae332d840afc4501c8f6a2017128fca525be4809a13", "submitted_at": "2024-10-17T17:50:38Z", "updated_at": "2025-05-24T03:55:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13825v2"}, "authors": ["Ke Yang", "Yao Liu", "Sapana Chaudhary", "Rasool Fakoor", "Pratik Chaudhari", "George Karypis", "Huzefa Rangwala"], "canonical_source_url": "https://arxiv.org/abs/2410.13825", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-13825/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09268d51b3d810ffc0024ae332d840afc4501c8f6a2017128fca525be4809a13", "submitted_at": "2024-10-17T17:50:38Z", "updated_at": "2025-05-24T03:55:22Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13825v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.13825"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13825"}], "name": "AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents", "publication": null, "published_at": "2024-10-17", "record_sha256": "b9ee354525191733f5e5dd105a5d39d5d10da1ffda24df62c95df3b8e1c9b0cf", "relations": [], "resource_id": "arxiv:2410.13825", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.13825v2", "topics": ["web-navigation", "planning"], "updated_at": "2025-05-24"}, "text": "AgentOccam: A Simple Yet Strong Baseline for LLM-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how simplifying and aligning a web agent's observations and actions with language-model capabilities affects task completion.\nPlatforms: web\nTopics: web-navigation, planning\nPublished: 2024-10-17\nAuthors: Ke Yang, Yao Liu, Sapana Chaudhary, Rasool Fakoor, Pratik Chaudhari, George Karypis, Huzefa Rangwala\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.13825v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-13825/"}
{"id": "arxiv:2410.13886", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4922d9362b23bc0e41d47dacf590c6e9bd6e88cfa6418cf3517814131295270c", "submitted_at": "2024-10-11T06:54:12Z", "updated_at": "2024-10-21T18:06:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13886v2"}, "authors": ["Priyanshu Kumar", "Elaine Lau", "Saranya Vijayakumar", "Tu Trinh", "Scale Red Team", "Elaine Chang", "Vaughn Robinson", "Sean Hendryx", "Shuyan Zhou", "Matt Fredrikson", "Summer Yue", "Zifan Wang"], "canonical_source_url": "https://arxiv.org/abs/2410.13886", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-13886/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4922d9362b23bc0e41d47dacf590c6e9bd6e88cfa6418cf3517814131295270c", "submitted_at": "2024-10-11T06:54:12Z", "updated_at": "2024-10-21T18:06:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.13886v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.13886"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.13886"}], "name": "Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents", "publication": null, "published_at": "2024-10-11", "record_sha256": "5da4b11ea07b22c33ffc941e7dc2ab206826f006d8951876f7ae8d023b8d67f0", "relations": [], "resource_id": "arxiv:2410.13886", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.13886v2", "topics": ["safety-security", "evaluation"], "updated_at": "2024-10-21"}, "text": "Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces BrowserART and studies whether chat-model refusal behavior transfers to browser agents carrying out potentially harmful instructions.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2024-10-11\nAuthors: Priyanshu Kumar, Elaine Lau, Saranya Vijayakumar, Tu Trinh, Scale Red Team, Elaine Chang, Vaughn Robinson, Sean Hendryx, Shuyan Zhou, Matt Fredrikson, Summer Yue, Zifan Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.13886v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-13886/"}
{"id": "arxiv:2410.14803", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9e3d8350cd0be5d4f3e82b3b35a15fe89403cd602e8bb20b7cd783c7a7067a0", "submitted_at": "2024-10-18T18:19:56Z", "updated_at": "2025-02-21T16:23:59Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2410.14803v5"}, "authors": ["Taiyi Wang", "Zhihao Wu", "Jianheng Liu", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2410.14803", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-14803/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9e3d8350cd0be5d4f3e82b3b35a15fe89403cd602e8bb20b7cd783c7a7067a0", "submitted_at": "2024-10-18T18:19:56Z", "updated_at": "2025-02-21T16:23:59Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2410.14803v5"}, "publication": {"authors": ["Taiyi Wang", "Zhihao Wu", "Jianheng Liu", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher title uses singular Agent; the preprint uses Agents. The authors and DistRL contribution agree.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "74757-74782", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b9e472cd579c83e2f6aa3459f46aac28-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agent", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "f3c834388e958e41ea7a8419b354435c1c522ef256c00d3af7f72bebb89c7fdc", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.14803"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.14803"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b9e472cd579c83e2f6aa3459f46aac28-Paper-Conference.pdf"}], "name": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agents", "publication": {"authors": ["Taiyi Wang", "Zhihao Wu", "Jianheng Liu", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher title uses singular Agent; the preprint uses Agents. The authors and DistRL contribution agree.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "74757-74782", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b9e472cd579c83e2f6aa3459f46aac28-Paper-Conference.pdf", "publisher": null, "series": null, "title": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agent", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-18", "record_sha256": "5ca3e2553a4972c41b5256e2d8247a0943a53c8ae8c0f3bceec9a2d4e9ccce32", "relations": [], "resource_id": "arxiv:2410.14803", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.14803v5", "topics": ["reinforcement-learning", "efficiency"], "updated_at": "2025-02-21"}, "text": "DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces asynchronous distributed reinforcement learning with centralized updates and decentralized interaction collection for device-control agents.\nPlatforms: android\nTopics: reinforcement-learning, efficiency\nPublished: 2024-10-18\nAuthors: Taiyi Wang, Zhihao Wu, Jianheng Liu, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v5\n\nPublication edition: DistRL: An Asynchronous Distributed Reinforcement Learning Framework for On-Device Control Agent\nPublication authors: Taiyi Wang, Zhihao Wu, Jianheng Liu, Jianye HAO, Jun Wang, Kun Shao\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 74757-74782\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/b9e472cd579c83e2f6aa3459f46aac28-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.14803v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-14803/"}
{"id": "arxiv:2410.15164", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "206059754652205f1e036fac8ee117f71dd356bfe2d35fd84aff8d2935ea242d", "submitted_at": "2024-10-19T17:28:48Z", "updated_at": "2025-03-31T20:39:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.15164v3"}, "authors": ["Jingxuan Chen", "Derek Yuen", "Bin Xie", "Yuhao Yang", "Gongwei Chen", "Zhihao Wu", "Li Yixing", "Xurui Zhou", "Weiwen Liu", "Shuai Wang", "Kaiwen Zhou", "Rui Shao", "Liqiang Nie", "Yasheng Wang", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2410.15164", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-15164/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "206059754652205f1e036fac8ee117f71dd356bfe2d35fd84aff8d2935ea242d", "submitted_at": "2024-10-19T17:28:48Z", "updated_at": "2025-03-31T20:39:17Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.15164v3"}, "publication": {"authors": ["Jingxuan Chen", "Derek Yuen", "Bin Xie", "Yuhao Yang", "Gongwei Chen", "Zhihao Wu", "Li Yixing", "Xurui Zhou", "Weiwen Liu", "Shuai Wang", "Kaiwen Zhou", "Rui Shao", "Liqiang Nie", "Yasheng Wang", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "61769-61806", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/9a75f4dd9679aa4ff80a4e6f0a1dc700-Paper-Conference.pdf", "publisher": null, "series": null, "title": "SPA-BENCH: A COMPREHENSIVE BENCHMARK FOR SMARTPHONE AGENT EVALUATION", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "56efdfea12b96edd94ad366f941d57668cdb75f6889fc03297e749497ff03e92", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.15164"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.15164"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/9a75f4dd9679aa4ff80a4e6f0a1dc700-Paper-Conference.pdf"}], "name": "SPA-Bench: A Comprehensive Benchmark for SmartPhone Agent Evaluation", "publication": {"authors": ["Jingxuan Chen", "Derek Yuen", "Bin Xie", "Yuhao Yang", "Gongwei Chen", "Zhihao Wu", "Li Yixing", "Xurui Zhou", "Weiwen Liu", "Shuai Wang", "Kaiwen Zhou", "Rui Shao", "Liqiang Nie", "Yasheng Wang", "Jianye HAO", "Jun Wang", "Kun Shao"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "61769-61806", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/9a75f4dd9679aa4ff80a4e6f0a1dc700-Paper-Conference.pdf", "publisher": null, "series": null, "title": "SPA-BENCH: A COMPREHENSIVE BENCHMARK FOR SMARTPHONE AGENT EVALUATION", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-19", "record_sha256": "f836c4fe29212ed65c4399f626d923325166c5856baecc675a36d139c35ebbbc", "relations": [], "resource_id": "arxiv:2410.15164", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.15164v3", "topics": ["evaluation", "efficiency", "mobile-navigation"], "updated_at": "2025-03-31"}, "text": "SPA-Bench: A Comprehensive Benchmark for SmartPhone Agent Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a smartphone-agent benchmark with English and Chinese tasks, real-time Android interaction, and completion and resource-use metrics.\nPlatforms: android\nTopics: evaluation, efficiency, mobile-navigation\nPublished: 2024-10-19\nAuthors: Jingxuan Chen, Derek Yuen, Bin Xie, Yuhao Yang, Gongwei Chen, Zhihao Wu, Li Yixing, Xurui Zhou, Weiwen Liu, Shuai Wang, Kaiwen Zhou, Rui Shao, Liqiang Nie, Yasheng Wang, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v3\n\nPublication edition: SPA-BENCH: A COMPREHENSIVE BENCHMARK FOR SMARTPHONE AGENT EVALUATION\nPublication authors: Jingxuan Chen, Derek Yuen, Bin Xie, Yuhao Yang, Gongwei Chen, Zhihao Wu, Li Yixing, Xurui Zhou, Weiwen Liu, Shuai Wang, Kaiwen Zhou, Rui Shao, Liqiang Nie, Yasheng Wang, Jianye HAO, Jun Wang, Kun Shao\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 61769-61806\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/9a75f4dd9679aa4ff80a4e6f0a1dc700-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.15164v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-15164/"}
{"id": "arxiv:2410.16464", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0944387dd3c65ea7b2f5ca52721496cf10b3c9860b303063201e7811318f4950", "submitted_at": "2024-10-21T19:46:06Z", "updated_at": "2025-06-16T22:23:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.16464v3"}, "authors": ["Yueqi Song", "Frank Xu", "Shuyan Zhou", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2410.16464", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-16464/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0944387dd3c65ea7b2f5ca52721496cf10b3c9860b303063201e7811318f4950", "submitted_at": "2024-10-21T19:46:06Z", "updated_at": "2025-06-16T22:23:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.16464v3"}, "publication": {"authors": ["Yueqi Song", "Frank F. Xu", "Shuyan Zhou", "Graham Neubig"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.577", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.577/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "11066-11085", "pdf_url": "https://aclanthology.org/2025.findings-acl.577.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Beyond Browsing: API-Based Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.577/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "12c4aaa6eb73e55e905fb507db47a14cc245c933504453365b06242c39e92481", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.577/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.16464"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.16464"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.577/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.577.pdf"}], "name": "Beyond Browsing: API-Based Web Agents", "publication": {"authors": ["Yueqi Song", "Frank F. Xu", "Shuyan Zhou", "Graham Neubig"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.577", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.577/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "11066-11085", "pdf_url": "https://aclanthology.org/2025.findings-acl.577.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Beyond Browsing: API-Based Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.577/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-10-21", "record_sha256": "4197fa96f0a1608779275676647961370037abf0a5f4ff9b0828836559009cd1", "relations": [], "resource_id": "arxiv:2410.16464", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.16464v3", "topics": ["web-navigation", "tool-integration", "evaluation"], "updated_at": "2025-06-16"}, "text": "Beyond Browsing: API-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCompares browsing-only, API-only, and hybrid web agents on WebArena, examining how access to application APIs changes task execution alongside browser interaction.\nPlatforms: web\nTopics: web-navigation, tool-integration, evaluation\nPublished: 2024-10-21\nAuthors: Yueqi Song, Frank Xu, Shuyan Zhou, Graham Neubig\nRecorded arXiv version: v3\n\nPublication edition: Beyond Browsing: API-Based Web Agents\nPublication authors: Yueqi Song, Frank F. Xu, Shuyan Zhou, Graham Neubig\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 11066-11085\nPublication doi: 10.18653/v1/2025.findings-acl.577\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2025.findings-acl.577/\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.16464v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-16464/"}
{"id": "arxiv:2410.17401", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "226a4acbbaab42c8cbc44a897651d8c745119426ea1223f1b849558af32fab02", "submitted_at": "2024-10-22T20:18:26Z", "updated_at": "2025-05-31T18:34:01Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2410.17401v4"}, "authors": ["Chejian Xu", "Mintong Kang", "Jiawei Zhang", "Zeyi Liao", "Lingbo Mo", "Mengqi Yuan", "Huan Sun", "Bo Li"], "canonical_source_url": "https://arxiv.org/abs/2410.17401", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-17401/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "226a4acbbaab42c8cbc44a897651d8c745119426ea1223f1b849558af32fab02", "submitted_at": "2024-10-22T20:18:26Z", "updated_at": "2025-05-31T18:34:01Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2410.17401v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.17401"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17401"}], "name": "AdvAgent: Controllable Blackbox Red-teaming on Web Agents", "publication": null, "published_at": "2024-10-22", "record_sha256": "aacf89f06854c2299892fd0e1ca02b411679bf9263859f3327fe3a82ba37401e", "relations": [], "resource_id": "arxiv:2410.17401", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.17401v4", "topics": ["safety-security", "evaluation"], "updated_at": "2025-05-31"}, "text": "AdvAgent: Controllable Blackbox Red-teaming on Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a black-box web-agent red-teaming framework that learns adversarial prompts from agent feedback to study security weaknesses.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2024-10-22\nAuthors: Chejian Xu, Mintong Kang, Jiawei Zhang, Zeyi Liao, Lingbo Mo, Mengqi Yuan, Huan Sun, Bo Li\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.17401v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-17401/"}
{"id": "arxiv:2410.17520", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362cbf7ea267ddc4a64505aa95c4f16a9f3631f95319ce10eef9981e2c85937b", "submitted_at": "2024-10-23T02:51:43Z", "updated_at": "2026-01-27T01:51:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.17520v3"}, "authors": ["Juyong Lee", "Dongyoon Hahm", "June Suk Choi", "W. Bradley Knox", "Kimin Lee"], "canonical_source_url": "https://arxiv.org/abs/2410.17520", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-17520/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362cbf7ea267ddc4a64505aa95c4f16a9f3631f95319ce10eef9981e2c85937b", "submitted_at": "2024-10-23T02:51:43Z", "updated_at": "2026-01-27T01:51:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.17520v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.17520"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17520"}], "name": "MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control", "publication": null, "published_at": "2024-10-23", "record_sha256": "14763fb7fba561f55c01a871b93e19261071a7f385e83e3763463d303c92b22f", "relations": [], "resource_id": "arxiv:2410.17520", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.17520v3", "topics": ["safety-security", "evaluation"], "updated_at": "2026-01-27"}, "text": "MobileSafetyBench: Evaluating Safety of Autonomous Agents in Mobile Device Control\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates mobile agents on misuse, side effects, and indirect prompt-injection risks in an Android emulator environment.\nPlatforms: android\nTopics: safety-security, evaluation\nPublished: 2024-10-23\nAuthors: Juyong Lee, Dongyoon Hahm, June Suk Choi, W. Bradley Knox, Kimin Lee\nRecorded arXiv version: v3\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.17520v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-17520/"}
{"id": "arxiv:2410.17883", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f6eb76587ef91a21942d34ba9ee7ba0ebc846277b12c69ecb179c2b2d1720df", "submitted_at": "2024-10-23T13:57:00Z", "updated_at": "2025-02-12T17:51:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.17883v2"}, "authors": ["Filippos Christianos", "Georgios Papoudakis", "Thomas Coste", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2410.17883", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-17883/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9f6eb76587ef91a21942d34ba9ee7ba0ebc846277b12c69ecb179c2b2d1720df", "submitted_at": "2024-10-23T13:57:00Z", "updated_at": "2025-02-12T17:51:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.17883v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.17883"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.17883"}], "name": "Lightweight Neural App Control", "publication": null, "published_at": "2024-10-23", "record_sha256": "440e005214bbea896f79f3cbcf5baa238efe3856dcac2c81074798ad5a99b700", "relations": [], "resource_id": "arxiv:2410.17883", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.17883v2", "topics": ["mobile-navigation", "action-prediction", "efficiency", "interaction-history"], "updated_at": "2025-02-12"}, "text": "Lightweight Neural App Control\nResource type: paper\nReview status: curated\nScope: core\n\nLiMAC combines a compact Action Transformer with a fine-tuned vision-language model for Android app control. It uses textual goals and histories of screenshots and UI trees to predict actions, with evaluation on two mobile-control datasets.\nPlatforms: android\nTopics: mobile-navigation, action-prediction, efficiency, interaction-history\nPublished: 2024-10-23\nAuthors: Filippos Christianos, Georgios Papoudakis, Thomas Coste, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.17883v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-17883/"}
{"id": "arxiv:2410.18603", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ece9f64cc037dad8633388edf1b54d88818d8092b77e19b15780e929b23aa593", "submitted_at": "2024-10-24T09:58:40Z", "updated_at": "2024-10-24T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18603v1"}, "authors": ["Chengyou Jia", "Minnan Luo", "Zhuohang Dang", "Qiushi Sun", "Fangzhi Xu", "Junlin Hu", "Tianbao Xie", "Zhiyong Wu"], "canonical_source_url": "https://arxiv.org/abs/2410.18603", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-18603/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ece9f64cc037dad8633388edf1b54d88818d8092b77e19b15780e929b23aa593", "submitted_at": "2024-10-24T09:58:40Z", "updated_at": "2024-10-24T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18603v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.18603"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18603"}], "name": "AgentStore: Scalable Integration of Heterogeneous Agents As Specialized Generalist Computer Assistant", "publication": null, "published_at": "2024-10-24", "record_sha256": "77b0e4a374d79b66db66993335d2f810014581cedf84babec298da4288f557b5", "relations": [], "resource_id": "arxiv:2410.18603", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.18603v1", "topics": ["automation", "tool-integration"], "updated_at": "2024-10-24"}, "text": "AgentStore: Scalable Integration of Heterogeneous Agents As Specialized Generalist Computer Assistant\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a platform that integrates specialized agents under a coordinating MetaAgent for varied computer tasks.\nPlatforms: desktop\nTopics: automation, tool-integration\nPublished: 2024-10-24\nAuthors: Chengyou Jia, Minnan Luo, Zhuohang Dang, Qiushi Sun, Fangzhi Xu, Junlin Hu, Tianbao Xie, Zhiyong Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.18603v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-18603/"}
{"id": "arxiv:2410.18963", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66fa51b108ac79cc43a895ba757ca463d5008bfe6db5cde4bdf3a3bb38c444eb", "submitted_at": "2024-10-24T17:58:08Z", "updated_at": "2024-10-24T17:58:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18963v1"}, "authors": ["Xiaoqiang Wang", "Bang Liu"], "canonical_source_url": "https://arxiv.org/abs/2410.18963", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-18963/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "66fa51b108ac79cc43a895ba757ca463d5008bfe6db5cde4bdf3a3bb38c444eb", "submitted_at": "2024-10-24T17:58:08Z", "updated_at": "2024-10-24T17:58:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.18963v1"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.18963"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18963"}], "name": "OSCAR: Operating System Control via State-Aware Reasoning and Re-Planning", "publication": null, "published_at": "2024-10-24", "record_sha256": "d88593febf3f4a70c118210569eae3e030808e1d606e981a2e77ee0dc0e37633", "relations": [], "resource_id": "arxiv:2410.18963", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.18963v1", "topics": ["planning", "state-tracking", "tool-integration", "robustness"], "updated_at": "2024-10-24"}, "text": "OSCAR: Operating System Control via State-Aware Reasoning and Re-Planning\nResource type: paper\nReview status: curated\nScope: core\n\nOSCAR translates user instructions into Python control code while observing screen images and acting through mouse and keyboard operations. A state-machine controller handles errors and replans tasks during execution across desktop and mobile applications.\nPlatforms: desktop, cross-platform\nTopics: planning, state-tracking, tool-integration, robustness\nPublished: 2024-10-24\nAuthors: Xiaoqiang Wang, Bang Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.18963v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-18963/"}
{"id": "arxiv:2410.18967", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "299b7c631468e259ac9d25a1c4ebf507be57b7ee478c611d0a2597b588f6c4c8", "submitted_at": "2024-10-24T17:58:31Z", "updated_at": "2025-02-28T00:29:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.18967v2"}, "authors": ["Zhangheng Li", "Keen You", "Haotian Zhang", "Di Feng", "Harsh Agrawal", "Xiujun Li", "Mohana Prasad Sathya Moorthy", "Jeff Nichols", "Yinfei Yang", "Zhe Gan"], "canonical_source_url": "https://arxiv.org/abs/2410.18967", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-18967/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "299b7c631468e259ac9d25a1c4ebf507be57b7ee478c611d0a2597b588f6c4c8", "submitted_at": "2024-10-24T17:58:31Z", "updated_at": "2025-02-28T00:29:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.18967v2"}}, "environments": ["ios", "android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.18967"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.18967"}], "name": "Ferret-UI 2: Mastering Universal User Interface Understanding Across Platforms", "publication": null, "published_at": "2024-10-24", "record_sha256": "bbb0d6b687fa799b9ef217b44ffb58908e45becf167b62480f5677f8ac299184", "relations": [], "resource_id": "arxiv:2410.18967", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.18967v2", "topics": ["screen-understanding", "grounding"], "updated_at": "2025-02-28"}, "text": "Ferret-UI 2: Mastering Universal User Interface Understanding Across Platforms\nResource type: paper\nReview status: curated\nScope: core\n\nExtends Ferret-UI across device platforms with adaptive visual scaling and synthetic data for referring, grounding, and user-centered UI tasks.\nPlatforms: ios, android, web, cross-platform\nTopics: screen-understanding, grounding\nPublished: 2024-10-24\nAuthors: Zhangheng Li, Keen You, Haotian Zhang, Di Feng, Harsh Agrawal, Xiujun Li, Mohana Prasad Sathya Moorthy, Jeff Nichols, Yinfei Yang, Zhe Gan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.18967v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-18967/"}
{"id": "arxiv:2410.19100", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f413943bd5c2a395ddb5cf38a1a5ad61754b35a9eb5de43e8d980bd83e4adff", "submitted_at": "2024-10-24T19:03:01Z", "updated_at": "2025-02-15T05:19:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.19100v3"}, "authors": ["Lawrence Jang", "Yinheng Li", "Dan Zhao", "Charles Ding", "Justin Lin", "Paul Pu Liang", "Rogerio Bonatti", "Kazuhito Koishida"], "canonical_source_url": "https://arxiv.org/abs/2410.19100", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-19100/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f413943bd5c2a395ddb5cf38a1a5ad61754b35a9eb5de43e8d980bd83e4adff", "submitted_at": "2024-10-24T19:03:01Z", "updated_at": "2025-02-15T05:19:38Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.19100v3"}, "publication": {"authors": ["Lawrence Jang", "Yinheng Li", "Dan Zhao", "Charles Ding", "Justin Lin", "Paul Pu Liang", "Rogerio Bonatti", "Kazuhito Koishida"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "36934-36958", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5b555804d495321df2e3208cc27f4fbc-Paper-Conference.pdf", "publisher": null, "series": null, "title": "VideoWebArena:  Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:02Z", "role": "primary", "sha256": "fdb42289bd00ad80cb51b66a01b91e13e60f134e809bb6c46181b4024ccabad2", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.19100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19100"}, {"rel": "publication", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"}, {"rel": "publication_pdf", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5b555804d495321df2e3208cc27f4fbc-Paper-Conference.pdf"}], "name": "VideoWebArena: Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks", "publication": {"authors": ["Lawrence Jang", "Yinheng Li", "Dan Zhao", "Charles Ding", "Justin Lin", "Paul Pu Liang", "Rogerio Bonatti", "Kazuhito Koishida"], "date_basis": "Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 5, 1], "doi": null, "event_dates": null, "evidence_urls": ["https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "36934-36958", "pdf_url": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5b555804d495321df2e3208cc27f4fbc-Paper-Conference.pdf", "publisher": null, "series": null, "title": "VideoWebArena:  Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks", "type": "conference-paper", "url": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html", "venue": "International Conference on Learning Representations", "volume": "2025"}, "published_at": "2024-10-24", "record_sha256": "896949c261f8d763928d893ca5aafcac0aa072bbb8e5e6ac793b026b87e1a8b1", "relations": [], "resource_id": "arxiv:2410.19100", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.19100v3", "topics": ["evaluation", "video", "memory"], "updated_at": "2025-02-15"}, "text": "VideoWebArena: Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates whether web agents can retain procedural skills and factual information from long video tutorials and use them during tasks.\nPlatforms: web\nTopics: evaluation, video, memory\nPublished: 2024-10-24\nAuthors: Lawrence Jang, Yinheng Li, Dan Zhao, Charles Ding, Justin Lin, Paul Pu Liang, Rogerio Bonatti, Kazuhito Koishida\nRecorded arXiv version: v3\n\nPublication edition: VideoWebArena:  Evaluating Long Context Multimodal Agents with Video Understanding Web Tasks\nPublication authors: Lawrence Jang, Yinheng Li, Dan Zhao, Charles Ding, Justin Lin, Paul Pu Liang, Rogerio Bonatti, Kazuhito Koishida\nPublication edition date (recorded precision): 2025-05-01\nPublication venue: International Conference on Learning Representations\nPublication volume: 2025\nPublication pages: 36934-36958\nPublication date basis: Explicit publisher citation_publication_date: 2025-05-01. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://proceedings.iclr.cc/paper_files/paper/2025/hash/5b555804d495321df2e3208cc27f4fbc-Abstract-Conference.html\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.19100v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-19100/"}
{"id": "arxiv:2410.19461", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dce2e22787c6e2cce95ecc4f5583bf94d7670b4b87ab75d2b8c7efb66fbcd1a8", "submitted_at": "2024-10-25T10:46:17Z", "updated_at": "2024-11-02T08:54:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.19461v2"}, "authors": ["Xuetian Chen", "Hangcheng Li", "Jiaqing Liang", "Sihang Jiang", "Deqing Yang"], "canonical_source_url": "https://arxiv.org/abs/2410.19461", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-19461/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dce2e22787c6e2cce95ecc4f5583bf94d7670b4b87ab75d2b8c7efb66fbcd1a8", "submitted_at": "2024-10-25T10:46:17Z", "updated_at": "2024-11-02T08:54:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.19461v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.19461"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19461"}], "name": "EDGE: Enhanced Grounded GUI Understanding with Enriched Multi-Granularity Synthetic Data", "publication": null, "published_at": "2024-10-25", "record_sha256": "31b98dc94a6c3117cf6293d66df0cc2c730d3f2402269666e7130698546efe65", "relations": [], "resource_id": "arxiv:2410.19461", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.19461v2", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2024-11-02"}, "text": "EDGE: Enhanced Grounded GUI Understanding with Enriched Multi-Granularity Synthetic Data\nResource type: paper\nReview status: curated\nScope: core\n\nSynthesizes GUI training examples at multiple levels of detail from webpages and studies transfer of learned understanding to other interface environments.\nPlatforms: web, cross-platform\nTopics: data-generation, grounding, screen-understanding\nPublished: 2024-10-25\nAuthors: Xuetian Chen, Hangcheng Li, Jiaqing Liang, Sihang Jiang, Deqing Yang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.19461v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-19461/"}
{"id": "arxiv:2410.19609", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db2a115281f831ea9333a03fc2a9ebb2f09528eb082060f00a33b125f9ef7762", "submitted_at": "2024-10-25T15:01:27Z", "updated_at": "2024-10-25T15:01:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.19609v1"}, "authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Tianqing Fang", "Zhenzhong Lan", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2410.19609", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-19609/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db2a115281f831ea9333a03fc2a9ebb2f09528eb082060f00a33b125f9ef7762", "submitted_at": "2024-10-25T15:01:27Z", "updated_at": "2024-10-25T15:01:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.19609v1"}, "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Tianqing Fang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1336", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1336/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "27545-27564", "pdf_url": "https://aclanthology.org/2025.acl-long.1336.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1336/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "efd9fe9277a980d42a991abe428ccc3983299c6a8a4d91e18d19cc07a37ed22c", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.1336/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.19609"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.19609"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.1336/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.1336.pdf"}], "name": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization", "publication": {"authors": ["Hongliang He", "Wenlin Yao", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Tianqing Fang", "Zhenzhong Lan", "Dong Yu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1336", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1336/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "27545-27564", "pdf_url": "https://aclanthology.org/2025.acl-long.1336.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1336/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-10-25", "record_sha256": "ecbfa8d37daec52fe25ee7adf541f800e64dc6f76e31db77a690fc028bd9f9b2", "relations": [], "resource_id": "arxiv:2410.19609", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.19609v1", "topics": ["web-navigation", "data-generation"], "updated_at": "2024-10-25"}, "text": "OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nDevelops multimodal browser agents through repeated cycles of real-web exploration, trajectory feedback, and policy optimization.\nPlatforms: web\nTopics: web-navigation, data-generation\nPublished: 2024-10-25\nAuthors: Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Hongming Zhang, Tianqing Fang, Zhenzhong Lan, Dong Yu\nRecorded arXiv version: v1\n\nPublication edition: OpenWebVoyager: Building Multimodal Web Agents via Iterative Real-World Exploration, Feedback and Optimization\nPublication authors: Hongliang He, Wenlin Yao, Kaixin Ma, Wenhao Yu, Hongming Zhang, Tianqing Fang, Zhenzhong Lan, Dong Yu\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 27545-27564\nPublication doi: 10.18653/v1/2025.acl-long.1336\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.1336/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.19609v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-19609/"}
{"id": "arxiv:2410.22552", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf1fb8aa2e289a72f3c44c99e2d531eaa5eeea08b78ff6f3a21e139d47502b17", "submitted_at": "2024-10-29T21:37:04Z", "updated_at": "2024-10-29T21:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.22552v1"}, "authors": ["Jaekyeom Kim", "Dong-Ki Kim", "Lajanugen Logeswaran", "Sungryull Sohn", "Honglak Lee"], "canonical_source_url": "https://arxiv.org/abs/2410.22552", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-22552/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf1fb8aa2e289a72f3c44c99e2d531eaa5eeea08b78ff6f3a21e139d47502b17", "submitted_at": "2024-10-29T21:37:04Z", "updated_at": "2024-10-29T21:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.22552v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.22552"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.22552"}], "name": "Auto-Intent: Automated Intent Discovery and Self-Exploration for Large Language Model Web Agents", "publication": null, "published_at": "2024-10-29", "record_sha256": "27e956bbe6218acedda678d9bde9c87bed65f76a150b9de8f249ba0640b05751", "relations": [], "resource_id": "arxiv:2410.22552", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.22552v1", "topics": ["planning", "web-navigation"], "updated_at": "2024-10-29"}, "text": "Auto-Intent: Automated Intent Discovery and Self-Exploration for Large Language Model Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDiscovers compact intents from demonstrations and uses intent predictions to guide a language-model web agent's decisions and exploration.\nPlatforms: web\nTopics: planning, web-navigation\nPublished: 2024-10-29\nAuthors: Jaekyeom Kim, Dong-Ki Kim, Lajanugen Logeswaran, Sungryull Sohn, Honglak Lee\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.22552v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-22552/"}
{"id": "arxiv:2410.23218", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0dbd8ef8807e7886c02ed6296c9d188503fe7b676062fcff1890fd0652f10f3", "submitted_at": "2024-10-30T17:10:19Z", "updated_at": "2024-10-30T17:10:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23218v1"}, "authors": ["Zhiyong Wu", "Zhenyu Wu", "Fangzhi Xu", "Yian Wang", "Qiushi Sun", "Chengyou Jia", "Kanzhi Cheng", "Zichen Ding", "Liheng Chen", "Paul Pu Liang", "Yu Qiao"], "canonical_source_url": "https://arxiv.org/abs/2410.23218", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-23218/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e0dbd8ef8807e7886c02ed6296c9d188503fe7b676062fcff1890fd0652f10f3", "submitted_at": "2024-10-30T17:10:19Z", "updated_at": "2024-10-30T17:10:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23218v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.23218"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23218"}], "name": "OS-ATLAS: A Foundation Action Model for Generalist GUI Agents", "publication": null, "published_at": "2024-10-30", "record_sha256": "f2ffee6790016f0b3e7d09539033a10c35e8e10c2012ef3b068d77cd441a98d9", "relations": [], "resource_id": "arxiv:2410.23218", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.23218v1", "topics": ["grounding", "data-generation"], "updated_at": "2024-10-30"}, "text": "OS-ATLAS: A Foundation Action Model for Generalist GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines cross-platform GUI grounding data synthesis with action-model training to study generalization to unfamiliar interfaces.\nPlatforms: cross-platform\nTopics: grounding, data-generation\nPublished: 2024-10-30\nAuthors: Zhiyong Wu, Zhenyu Wu, Fangzhi Xu, Yian Wang, Qiushi Sun, Chengyou Jia, Kanzhi Cheng, Zichen Ding, Liheng Chen, Paul Pu Liang, Yu Qiao\nRecorded arXiv version: v1\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.23218v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-23218/"}
{"id": "arxiv:2410.23252", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26fac89c78ab03fcbaf25bbfe93ebce090dec9adbf5c6dd5440641f154cffac9", "submitted_at": "2024-10-30T17:35:44Z", "updated_at": "2025-03-08T23:37:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.23252v3"}, "authors": ["Haoyi Qiu", "Alexander R. Fabbri", "Divyansh Agarwal", "Kung-Hsiang Huang", "Sarah Tan", "Nanyun Peng", "Chien-Sheng Wu"], "canonical_source_url": "https://arxiv.org/abs/2410.23252", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-23252/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26fac89c78ab03fcbaf25bbfe93ebce090dec9adbf5c6dd5440641f154cffac9", "submitted_at": "2024-10-30T17:35:44Z", "updated_at": "2025-03-08T23:37:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2410.23252v3"}, "publication": {"authors": ["Haoyi Qiu", "Alexander Richard Fabbri", "Divyansh Agarwal", "Kung-Hsiang Huang", "Sarah Tan", "Nanyun Peng", "Chien-Sheng Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.findings-naacl.222", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-naacl.222/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "3978-4005", "pdf_url": "https://aclanthology.org/2025.findings-naacl.222.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Evaluating Cultural and Social Awareness of LLM Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-naacl.222/", "venue": "Findings of the Association for Computational Linguistics: NAACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:55Z", "role": "primary", "sha256": "6563b807356c30b69a1c5bb9328358c139d171777cb893079ddc123cb82a282e", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-naacl.222/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2410.23252"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23252"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-naacl.222/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-naacl.222.pdf"}], "name": "Evaluating Cultural and Social Awareness of LLM Web Agents", "publication": {"authors": ["Haoyi Qiu", "Alexander Richard Fabbri", "Divyansh Agarwal", "Kung-Hsiang Huang", "Sarah Tan", "Nanyun Peng", "Chien-Sheng Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.findings-naacl.222", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-naacl.222/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "3978-4005", "pdf_url": "https://aclanthology.org/2025.findings-naacl.222.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Evaluating Cultural and Social Awareness of LLM Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-naacl.222/", "venue": "Findings of the Association for Computational Linguistics: NAACL 2025", "volume": null}, "published_at": "2024-10-30", "record_sha256": "ed1651ece6a5e7aa9d640012f2239961538e53f396cbc5000ea4721e82be855d", "relations": [], "resource_id": "arxiv:2410.23252", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.23252v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-08"}, "text": "Evaluating Cultural and Social Awareness of LLM Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CASA to assess web agents' responses to cultural and social norms in shopping and forum tasks, including misleading observations.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2024-10-30\nAuthors: Haoyi Qiu, Alexander R. Fabbri, Divyansh Agarwal, Kung-Hsiang Huang, Sarah Tan, Nanyun Peng, Chien-Sheng Wu\nRecorded arXiv version: v3\n\nPublication edition: Evaluating Cultural and Social Awareness of LLM Web Agents\nPublication authors: Haoyi Qiu, Alexander Richard Fabbri, Divyansh Agarwal, Kung-Hsiang Huang, Sarah Tan, Nanyun Peng, Chien-Sheng Wu\nPublication edition date (recorded precision): 2025-04\nPublication venue: Findings of the Association for Computational Linguistics: NAACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 3978-4005\nPublication doi: 10.18653/v1/2025.findings-naacl.222\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2025.findings-naacl.222/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.23252v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-23252/"}
{"id": "arxiv:2410.23555", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c67b1559410152607fc9d54abf25c37ef939c9f755f089bdf87d4f39393c085", "submitted_at": "2024-10-31T01:51:41Z", "updated_at": "2024-10-31T01:51:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23555v1"}, "authors": ["Nalin Tiwary", "Vardhan Dongre", "Sanil Arun Chawla", "Ashwin Lamani", "Dilek Hakkani-Tür"], "canonical_source_url": "https://arxiv.org/abs/2410.23555", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-23555/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c67b1559410152607fc9d54abf25c37ef939c9f755f089bdf87d4f39393c085", "submitted_at": "2024-10-31T01:51:41Z", "updated_at": "2024-10-31T01:51:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2410.23555v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "fced2234354232b8834eef0cf2557b2932990e74b3530fc95459cd3fa0e11792", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2410.23555v1"}], "identifiers": {"arxiv": "2410.23555"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.23555"}], "name": "From Context to Action: Analysis of the Impact of State Representation and Context on the Generalization of Multi-Turn Web Navigation Agents", "publication": null, "published_at": "2024-10-31", "record_sha256": "ec9d9b5bb9b69d11250ccc535480d2e82e814b9010c6088ac8bef9fa2677e6d1", "relations": [], "resource_id": "arxiv:2410.23555", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.23555v1", "topics": ["web-navigation", "interaction-history", "state-tracking"], "updated_at": "2024-10-31"}, "text": "From Context to Action: Analysis of the Impact of State Representation and Context on the Generalization of Multi-Turn Web Navigation Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes how interaction history and webpage representation affect multiturn web-agent generalization across unfamiliar websites, task categories, and geographic settings.\nPlatforms: web\nTopics: web-navigation, interaction-history, state-tracking\nPublished: 2024-10-31\nAuthors: Nalin Tiwary, Vardhan Dongre, Sanil Arun Chawla, Ashwin Lamani, Dilek Hakkani-Tür\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.23555v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-23555/"}
{"id": "arxiv:2410.24024", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29fe5262adf8677bbd5fead628c5fc28707af263aa7e800d1dee1f7e499784c", "submitted_at": "2024-10-31T15:25:20Z", "updated_at": "2024-11-04T05:57:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.24024v2"}, "authors": ["Yifan Xu", "Xiao Liu", "Xueqiao Sun", "Siyi Cheng", "Hao Yu", "Hanyu Lai", "Shudan Zhang", "Dan Zhang", "Jie Tang", "Yuxiao Dong"], "canonical_source_url": "https://arxiv.org/abs/2410.24024", "citation_url": "https://awesomecua.github.io/resources/arxiv-2410-24024/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29fe5262adf8677bbd5fead628c5fc28707af263aa7e800d1dee1f7e499784c", "submitted_at": "2024-10-31T15:25:20Z", "updated_at": "2024-11-04T05:57:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2410.24024v2"}, "publication": {"authors": ["Yifan Xu", "Xiao Liu", "Xueqiao Sun", "Siyi Cheng", "Hao Yu (余浩)", "Hanyu Lai", "Shudan Zhang", "Dan Zhang", "Jie Tang", "Yuxiao Dong"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.107", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.107/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "2144-2166", "pdf_url": "https://aclanthology.org/2025.acl-long.107.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.107/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "52b82854b5f4703e6e7f3cd3e9fc49adcd011ca3ba0a8fd1aced77214c452043", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.107/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2410.24024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2410.24024"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.107/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.107.pdf"}], "name": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents", "publication": {"authors": ["Yifan Xu", "Xiao Liu", "Xueqiao Sun", "Siyi Cheng", "Hao Yu (余浩)", "Hanyu Lai", "Shudan Zhang", "Dan Zhang", "Jie Tang", "Yuxiao Dong"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.107", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.107/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "2144-2166", "pdf_url": "https://aclanthology.org/2025.acl-long.107.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.107/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-10-31", "record_sha256": "42d30ac7f024cbaa15f5641bfa25fe31515fbcb35df1f0f188126b64d5b55f9e", "relations": [], "resource_id": "arxiv:2410.24024", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2410.24024v2", "topics": ["evaluation", "data-generation", "mobile-navigation"], "updated_at": "2024-11-04"}, "text": "AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProvides a shared Android action environment, reproducible tasks, and instruction data for training and comparing language and multimodal agents.\nPlatforms: android\nTopics: evaluation, data-generation, mobile-navigation\nPublished: 2024-10-31\nAuthors: Yifan Xu, Xiao Liu, Xueqiao Sun, Siyi Cheng, Hao Yu, Hanyu Lai, Shudan Zhang, Dan Zhang, Jie Tang, Yuxiao Dong\nRecorded arXiv version: v2\n\nPublication edition: AndroidLab: Training and Systematic Benchmarking of Android Autonomous Agents\nPublication authors: Yifan Xu, Xiao Liu, Xueqiao Sun, Siyi Cheng, Hao Yu (余浩), Hanyu Lai, Shudan Zhang, Dan Zhang, Jie Tang, Yuxiao Dong\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 2144-2166\nPublication doi: 10.18653/v1/2025.acl-long.107\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2025.acl-long.107/\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2410.24024v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2410-24024/"}
{"id": "arxiv:2411.00820", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cfe1aa31818f06567aa3d619ad90ac4bdb9f9f9003a1a8a71294cd4f8ccf8dc", "submitted_at": "2024-10-28T17:05:10Z", "updated_at": "2024-10-28T17:05:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.00820v1"}, "authors": ["Xiao Liu", "Bo Qin", "Dongzhu Liang", "Guang Dong", "Hanyu Lai", "Hanchen Zhang", "Hanlin Zhao", "Iat Long Iong", "Jiadai Sun", "Jiaqi Wang", "Junjie Gao", "Junjun Shan", "Kangning Liu", "Shudan Zhang", "Shuntian Yao", "Siyi Cheng", "Wentao Yao", "Wenyi Zhao", "Xinghan Liu", "Xinyi Liu", "Xinying Chen", "Xinyue Yang", "Yang Yang", "Yifan Xu", "Yu Yang", "Yujia Wang", "Yulin Xu", "Zehan Qi", "Yuxiao Dong", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2411.00820", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-00820/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cfe1aa31818f06567aa3d619ad90ac4bdb9f9f9003a1a8a71294cd4f8ccf8dc", "submitted_at": "2024-10-28T17:05:10Z", "updated_at": "2024-10-28T17:05:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.00820v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2411.00820"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.00820"}], "name": "AutoGLM: Autonomous Foundation Agents for GUIs", "publication": null, "published_at": "2024-10-28", "record_sha256": "5054cd73e303de653428a1e66171f0cf7b226d1293071abca5e0dd387214deaa", "relations": [], "resource_id": "arxiv:2411.00820", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.00820v1", "topics": ["automation", "reinforcement-learning", "grounding"], "updated_at": "2024-10-28"}, "text": "AutoGLM: Autonomous Foundation Agents for GUIs\nResource type: paper\nReview status: curated\nScope: core\n\nPresents AutoGLM agents for browser and phone control, separating planning from grounding and training through an evolving interaction curriculum.\nPlatforms: web, android\nTopics: automation, reinforcement-learning, grounding\nPublished: 2024-10-28\nAuthors: Xiao Liu, Bo Qin, Dongzhu Liang, Guang Dong, Hanyu Lai, Hanchen Zhang, Hanlin Zhao, Iat Long Iong, Jiadai Sun, Jiaqi Wang, Junjie Gao, Junjun Shan, Kangning Liu, Shudan Zhang, Shuntian Yao, Siyi Cheng, Wentao Yao, Wenyi Zhao, Xinghan Liu, Xinyi Liu, Xinying Chen, Xinyue Yang, Yang Yang, Yifan Xu, Yu Yang, Yujia Wang, Yulin Xu, Zehan Qi, Yuxiao Dong, Jie Tang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.00820v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-00820/"}
{"id": "arxiv:2411.02006", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de860a29f6c54317586c2218688436898f1d7a8b61207b1d5f4a65b05b3a4e6e", "submitted_at": "2024-11-04T11:50:58Z", "updated_at": "2025-09-14T09:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02006v3"}, "authors": ["Biao Wu", "Yanda Li", "Zhiwei Zhang", "Yunchao Wei", "Meng Fang", "Ling Chen"], "canonical_source_url": "https://arxiv.org/abs/2411.02006", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-02006/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de860a29f6c54317586c2218688436898f1d7a8b61207b1d5f4a65b05b3a4e6e", "submitted_at": "2024-11-04T11:50:58Z", "updated_at": "2025-09-14T09:13:39Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02006v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:27Z", "role": "primary", "sha256": "bda47fcb119c581a798dbc791b4929af91832d5053ac073a02dab8fd4e10ace3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2411.02006v3"}], "identifiers": {"arxiv": "2411.02006"}, "kind": "survey", "last_verified_at": "2026-10-17T09:13:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02006"}], "name": "Foundations and Recent Trends in Multimodal Mobile Agents: A Survey", "publication": null, "published_at": "2024-11-04", "record_sha256": "f9da592ace7b5fd2c30366884adb4dd0a061367313b87b2bf7dbcf49cfefeaa6", "relations": [], "resource_id": "arxiv:2411.02006", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.02006v3", "topics": ["survey", "mobile-navigation", "evaluation"], "updated_at": "2025-09-14"}, "text": "Foundations and Recent Trends in Multimodal Mobile Agents: A Survey\nResource type: survey\nReview status: curated\nScope: core\n\nSurveys multimodal mobile agents, organizing prompt-based and training-based approaches alongside evaluation benchmarks, supporting technologies, and open challenges in mobile task execution.\nPlatforms: unspecified\nTopics: survey, mobile-navigation, evaluation\nPublished: 2024-11-04\nAuthors: Biao Wu, Yanda Li, Zhiwei Zhang, Yunchao Wei, Meng Fang, Ling Chen\nRecorded arXiv version: v3\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.02006v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-02006/"}
{"id": "arxiv:2411.02337", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db039a2592cd79a1f8f7f1c985c2ddf42447c63d1f58d2815ba1d6ef2369da47", "submitted_at": "2024-11-04T17:59:58Z", "updated_at": "2025-01-27T11:56:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02337v3"}, "authors": ["Zehan Qi", "Xiao Liu", "Iat Long Iong", "Hanyu Lai", "Xueqiao Sun", "Wenyi Zhao", "Yu Yang", "Xinyue Yang", "Jiadai Sun", "Shuntian Yao", "Tianjie Zhang", "Wei Xu", "Jie Tang", "Yuxiao Dong"], "canonical_source_url": "https://arxiv.org/abs/2411.02337", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-02337/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db039a2592cd79a1f8f7f1c985c2ddf42447c63d1f58d2815ba1d6ef2369da47", "submitted_at": "2024-11-04T17:59:58Z", "updated_at": "2025-01-27T11:56:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.02337v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.02337"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02337"}], "name": "WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning", "publication": null, "published_at": "2024-11-04", "record_sha256": "df0cc8c863ef0c51c2d1814613d8fcb1c950431edde305f05b42ce1a2053f590", "relations": [], "resource_id": "arxiv:2411.02337", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.02337v3", "topics": ["reinforcement-learning", "web-navigation"], "updated_at": "2025-01-27"}, "text": "WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nTrains web agents with an evolving task curriculum, outcome supervision, and reinforcement-learning updates intended to manage policy drift.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation\nPublished: 2024-11-04\nAuthors: Zehan Qi, Xiao Liu, Iat Long Iong, Hanyu Lai, Xueqiao Sun, Wenyi Zhao, Yu Yang, Xinyue Yang, Jiadai Sun, Shuntian Yao, Tianjie Zhang, Wei Xu, Jie Tang, Yuxiao Dong\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.02337v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-02337/"}
{"id": "arxiv:2411.02391", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "085c64595365cc6439c8c5fa8a479259a3d59b7e9740e0f9d43e48d00e1439c0", "submitted_at": "2024-11-04T18:56:42Z", "updated_at": "2025-05-24T16:15:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.02391v2"}, "authors": ["Yanzhe Zhang", "Tao Yu", "Diyi Yang"], "canonical_source_url": "https://arxiv.org/abs/2411.02391", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-02391/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "085c64595365cc6439c8c5fa8a479259a3d59b7e9740e0f9d43e48d00e1439c0", "submitted_at": "2024-11-04T18:56:42Z", "updated_at": "2025-05-24T16:15:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.02391v2"}, "publication": {"authors": ["Yanzhe Zhang", "Tao Yu", "Diyi Yang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.411", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.411/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8387-8401", "pdf_url": "https://aclanthology.org/2025.acl-long.411.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Attacking Vision-Language Computer Agents via Pop-ups", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.411/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "a28a603d74e08954a05265f54c8765bd874a6cb2fc0c5a68a35e15e68692b0a1", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.411/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2411.02391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.02391"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.411/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.411.pdf"}], "name": "Attacking Vision-Language Computer Agents via Pop-ups", "publication": {"authors": ["Yanzhe Zhang", "Tao Yu", "Diyi Yang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.411", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.411/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "8387-8401", "pdf_url": "https://aclanthology.org/2025.acl-long.411.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Attacking Vision-Language Computer Agents via Pop-ups", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.411/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-11-04", "record_sha256": "7c0809b194c874dc2f451c32ddceeba158415988a31fe8a26ba560f05f0127fd", "relations": [], "resource_id": "arxiv:2411.02391", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.02391v2", "topics": ["safety-security", "evaluation"], "updated_at": "2025-05-24"}, "text": "Attacking Vision-Language Computer Agents via Pop-ups\nResource type: paper\nReview status: curated\nScope: core\n\nTests whether adversarial visual pop-ups divert computer agents from intended tasks in desktop and web benchmark environments.\nPlatforms: desktop, web\nTopics: safety-security, evaluation\nPublished: 2024-11-04\nAuthors: Yanzhe Zhang, Tao Yu, Diyi Yang\nRecorded arXiv version: v2\n\nPublication edition: Attacking Vision-Language Computer Agents via Pop-ups\nPublication authors: Yanzhe Zhang, Tao Yu, Diyi Yang\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 8387-8401\nPublication doi: 10.18653/v1/2025.acl-long.411\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.411/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.02391v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-02391/"}
{"id": "arxiv:2411.03817", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd631a31f72dabe3a3bebf322031f9d686082073a8e4aa71eb65745498327f4e", "submitted_at": "2024-11-06T10:35:11Z", "updated_at": "2024-12-09T09:20:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.03817v3"}, "authors": ["Zhirui Deng", "Zhicheng Dou", "Yutao Zhu", "Ji-Rong Wen", "Ruibin Xiong", "Mang Wang", "Weipeng Chen"], "canonical_source_url": "https://arxiv.org/abs/2411.03817", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-03817/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd631a31f72dabe3a3bebf322031f9d686082073a8e4aa71eb65745498327f4e", "submitted_at": "2024-11-06T10:35:11Z", "updated_at": "2024-12-09T09:20:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2411.03817v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:28Z", "role": "primary", "sha256": "b105dc6cc1b32ab71467b2701b6a85506f698e73f04cfbd63ab41eddfbca48b6", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://arxiv.org/abs/2411.03817v3"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:31Z", "role": "primary", "sha256": "ce801ec21577b4f62d5b0ecec7c50e87744e5b38db612a2aa4017e27fe37102f", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2411.03817v3"}], "identifiers": {"arxiv": "2411.03817"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.03817"}], "name": "From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning", "publication": null, "published_at": "2024-11-06", "record_sha256": "c73ab5d919ed6657ea803943e6ac44de3bff676338a7912604e6601b9f27d408", "relations": [], "resource_id": "arxiv:2411.03817", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2411.03817v3", "topics": ["reinforcement-learning", "reward-models", "self-reflection"], "updated_at": "2024-12-09"}, "text": "From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces StepAgent, which compares agent and expert actions to supply intermediate rewards for policy learning, with implicit-reward and inverse-RL variants. Evaluation includes WebShop and Mind2Web alongside non-web tasks.\nPlatforms: web\nTopics: reinforcement-learning, reward-models, self-reflection\nPublished: 2024-11-06\nAuthors: Zhirui Deng, Zhicheng Dou, Yutao Zhu, Ji-Rong Wen, Ruibin Xiong, Mang Wang, Weipeng Chen\nRecorded arXiv version: v3\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced. Official full-paper experimental sections explicitly evaluate WebShop and Mind2Web; the overall optimization method also covers embodied and question-answering tasks.\n\nPrimary source: https://arxiv.org/abs/2411.03817v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-03817/"}
{"id": "arxiv:2411.04468", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a2c499237dbcbe1482824a2710f5185dee05c9ed66a0882d1e422f30afe056a", "submitted_at": "2024-11-07T06:36:19Z", "updated_at": "2024-11-07T06:36:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.04468v1"}, "authors": ["Adam Fourney", "Gagan Bansal", "Hussein Mozannar", "Cheng Tan", "Eduardo Salinas", "Erkang", "Zhu", "Friederike Niedtner", "Grace Proebsting", "Griffin Bassman", "Jack Gerrits", "Jacob Alber", "Peter Chang", "Ricky Loynd", "Robert West", "Victor Dibia", "Ahmed Awadallah", "Ece Kamar", "Rafah Hosn", "Saleema Amershi"], "canonical_source_url": "https://arxiv.org/abs/2411.04468", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-04468/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4a2c499237dbcbe1482824a2710f5185dee05c9ed66a0882d1e422f30afe056a", "submitted_at": "2024-11-07T06:36:19Z", "updated_at": "2024-11-07T06:36:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.04468v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2411.04468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.04468"}], "name": "Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks", "publication": null, "published_at": "2024-11-07", "record_sha256": "03b70f9c55745ca2f10eea5e08e11caa7336483e81ea1989e11fbd118e827029", "relations": [], "resource_id": "arxiv:2411.04468", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2411.04468v1", "topics": ["planning", "tool-integration"], "updated_at": "2024-11-07"}, "text": "Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces an orchestrated team of browsing, file, and coding agents with progress tracking and error recovery for general digital tasks.\nPlatforms: web, desktop\nTopics: planning, tool-integration\nPublished: 2024-11-07\nAuthors: Adam Fourney, Gagan Bansal, Hussein Mozannar, Cheng Tan, Eduardo Salinas, Erkang, Zhu, Friederike Niedtner, Grace Proebsting, Griffin Bassman, Jack Gerrits, Jacob Alber, Peter Chang, Ricky Loynd, Robert West, Victor Dibia, Ahmed Awadallah, Ece Kamar, Rafah Hosn, Saleema Amershi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.04468v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-04468/"}
{"id": "arxiv:2411.04890", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57670445f9b0041bfb505a195d01a9ded6234e01e44c9aa43a6aee5ebd57dd93", "submitted_at": "2024-11-07T17:28:10Z", "updated_at": "2025-02-13T10:09:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.04890v2"}, "authors": ["Shuai Wang", "Weiwen Liu", "Jingxuan Chen", "Yuqi Zhou", "Weinan Gan", "Xingshan Zeng", "Yuhan Che", "Shuai Yu", "Xinlong Hao", "Kun Shao", "Bin Wang", "Chuhan Wu", "Yasheng Wang", "Ruiming Tang", "Jianye Hao"], "canonical_source_url": "https://arxiv.org/abs/2411.04890", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-04890/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57670445f9b0041bfb505a195d01a9ded6234e01e44c9aa43a6aee5ebd57dd93", "submitted_at": "2024-11-07T17:28:10Z", "updated_at": "2025-02-13T10:09:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.04890v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.04890"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.04890"}], "name": "GUI Agents with Foundation Models: A Comprehensive Survey", "publication": null, "published_at": "2024-11-07", "record_sha256": "3abab2c1f4f949baa649115df7df7c99a9e254636e022e58402b12b6c4fcbe70", "relations": [], "resource_id": "arxiv:2411.04890", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.04890v2", "topics": ["survey"], "updated_at": "2025-02-13"}, "text": "GUI Agents with Foundation Models: A Comprehensive Survey\nResource type: survey\nReview status: curated\nScope: core\n\nSurvey of (M)LLM-based GUI agents organized around their datasets and benchmarks, a unified agent framework with a taxonomy, and commercial applications.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2024-11-07\nAuthors: Shuai Wang, Weiwen Liu, Jingxuan Chen, Yuqi Zhou, Weinan Gan, Xingshan Zeng, Yuhan Che, Shuai Yu, Xinlong Hao, Kun Shao, Bin Wang, Chuhan Wu, Yasheng Wang, Ruiming Tang, Jianye Hao\nRecorded arXiv version: v2\n\nPrimary source: https://arxiv.org/abs/2411.04890v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-04890/"}
{"id": "arxiv:2411.06559", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ecaca6448aae5c8849d008d820d700bfa438ed0248f85be585b0b831408b9b6", "submitted_at": "2024-11-10T18:50:51Z", "updated_at": "2025-04-01T05:04:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.06559v2"}, "authors": ["Yu Gu", "Kai Zhang", "Yuting Ning", "Boyuan Zheng", "Boyu Gou", "Tianci Xue", "Cheng Chang", "Sanjari Srivastava", "Yanan Xie", "Peng Qi", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2411.06559", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-06559/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ecaca6448aae5c8849d008d820d700bfa438ed0248f85be585b0b831408b9b6", "submitted_at": "2024-11-10T18:50:51Z", "updated_at": "2025-04-01T05:04:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.06559v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.06559"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.06559"}], "name": "Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents", "publication": null, "published_at": "2024-11-10", "record_sha256": "fa538f8370c7704b24b74f8c61afc66c39a2c6c064dde24f04cd2d3e2f987707", "relations": [], "resource_id": "arxiv:2411.06559", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.06559v2", "topics": ["world-models", "planning"], "updated_at": "2025-04-01"}, "text": "Is Your LLM Secretly a World Model of the Internet? Model-Based Planning for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebDreamer, which simulates candidate web actions with language-model world models and value estimates before acting.\nPlatforms: web\nTopics: world-models, planning\nPublished: 2024-11-10\nAuthors: Yu Gu, Kai Zhang, Yuting Ning, Boyuan Zheng, Boyu Gou, Tianci Xue, Cheng Chang, Sanjari Srivastava, Yanan Xie, Peng Qi, Huan Sun, Yu Su\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.06559v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-06559/"}
{"id": "arxiv:2411.10323", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "255e101e263c4e3a0cb491d4c34b3e75a8f3660cec2141e3a507a30cf5514dbd", "submitted_at": "2024-11-15T16:23:52Z", "updated_at": "2024-11-15T16:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10323v1"}, "authors": ["Siyuan Hu", "Mingyu Ouyang", "Difei Gao", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2411.10323", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-10323/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "255e101e263c4e3a0cb491d4c34b3e75a8f3660cec2141e3a507a30cf5514dbd", "submitted_at": "2024-11-15T16:23:52Z", "updated_at": "2024-11-15T16:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10323v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.10323"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.10323"}], "name": "The Dawn of GUI Agent: A Preliminary Case Study with Claude 3.5 Computer Use", "publication": null, "published_at": "2024-11-15", "record_sha256": "8b4a43c0e44014c3d11ed247c8b8effc3c2aaa5045ed85b43b3eeb16888993ec", "relations": [], "resource_id": "arxiv:2411.10323", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.10323v1", "topics": ["evaluation", "automation", "planning"], "updated_at": "2024-11-15"}, "text": "The Dawn of GUI Agent: A Preliminary Case Study with Claude 3.5 Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a preliminary study of Claude 3.5 Computer Use across desktop tasks and software, analyzing planning, action, and critique behavior and providing a framework for deploying GUI-automation models.\nPlatforms: desktop\nTopics: evaluation, automation, planning\nPublished: 2024-11-15\nAuthors: Siyuan Hu, Mingyu Ouyang, Difei Gao, Mike Zheng Shou\nRecorded arXiv version: v1\n\nReviewed against the retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.10323v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-10323/"}
{"id": "arxiv:2411.10943", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa9b7092616e94a1d0124d36d8b75d0401b9647ec762c963524616a92def9a5", "submitted_at": "2024-11-17T02:44:56Z", "updated_at": "2024-11-17T02:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10943v1"}, "authors": ["Minghe Gao", "Wendong Bu", "Bingchen Miao", "Yang Wu", "Yunfei Li", "Juncheng Li", "Siliang Tang", "Qi Wu", "Yueting Zhuang", "Meng Wang"], "canonical_source_url": "https://arxiv.org/abs/2411.10943", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-10943/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa9b7092616e94a1d0124d36d8b75d0401b9647ec762c963524616a92def9a5", "submitted_at": "2024-11-17T02:44:56Z", "updated_at": "2024-11-17T02:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.10943v1"}}, "environments": ["web", "desktop", "android", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:29:13.577033+00:00", "role": "primary", "sha256": "92f18020e481d6f42a6d35e96dff820594e9370597f5392d038e9ee908c49f1e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2411.10943v1"}], "identifiers": {"arxiv": "2411.10943"}, "kind": "survey", "last_verified_at": "2026-10-18T04:29:13.577033+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.10943"}], "name": "Generalist Virtual Agents: A Survey on Autonomous Agents Across Digital Platforms", "publication": null, "published_at": "2024-11-17", "record_sha256": "c77fe98a748c471cc6e4699d79e96a8575f6de9d8766a2045cb7c037bff93995", "relations": [], "resource_id": "arxiv:2411.10943", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.10943v1", "topics": ["survey", "web-navigation", "mobile-navigation", "evaluation", "planning"], "updated_at": "2024-11-17"}, "text": "Generalist Virtual Agents: A Survey on Autonomous Agents Across Digital Platforms\nResource type: survey\nReview status: curated\nScope: core\n\nSurveys virtual agents for browsers, applications and operating systems, organizing prior work by task type, observation and action spaces, model strategies and evaluation methods. It examines screen-based interaction and keyboard, mouse and touch control, alongside transferability and long-sequence decision-making challenges.\nPlatforms: web, desktop, android, windows\nTopics: survey, web-navigation, mobile-navigation, evaluation, planning\nPublished: 2024-11-17\nAuthors: Minghe Gao, Wendong Bu, Bingchen Miao, Yang Wu, Yunfei Li, Juncheng Li, Siliang Tang, Qi Wu, Yueting Zhuang, Meng Wang\nRecorded arXiv version: v1\n\nReviewed the complete retained arXiv abstract and exact-version official HTML sections on browser/application/OS environments, screen observations, input actions and evaluation. The HTML abstract fingerprint and ordered authors match the retained Atom entry. Literature claims and experiments were not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.10943v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-10943/"}
{"id": "arxiv:2411.13451", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94d908dd89ff27434805fe7915c5a93d5c2b1d873449babc1c566c5582cca734", "submitted_at": "2024-11-20T16:54:15Z", "updated_at": "2024-11-20T16:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.13451v1"}, "authors": ["Gaurav Verma", "Rachneet Kaur", "Nishan Srishankar", "Zhen Zeng", "Tucker Balch", "Manuela Veloso"], "canonical_source_url": "https://arxiv.org/abs/2411.13451", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-13451/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "94d908dd89ff27434805fe7915c5a93d5c2b1d873449babc1c566c5582cca734", "submitted_at": "2024-11-20T16:54:15Z", "updated_at": "2024-11-20T16:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.13451v1"}, "publication": {"authors": ["Gaurav Verma", "Rachneet Kaur", "Nishan Srishankar", "Zhen Zeng", "Tucker Balch", "Manuela Veloso"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1008", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1008/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "20635-20651", "pdf_url": "https://aclanthology.org/2025.acl-long.1008.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1008/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:56Z", "role": "primary", "sha256": "e8930e90d2d78f0d6ef35acc487ba355d3798cdafbcc1d2637258bad81407d41", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.1008/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.13451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.13451"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.1008/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.1008.pdf"}], "name": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations", "publication": {"authors": ["Gaurav Verma", "Rachneet Kaur", "Nishan Srishankar", "Zhen Zeng", "Tucker Balch", "Manuela Veloso"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.1008", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.1008/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "20635-20651", "pdf_url": "https://aclanthology.org/2025.acl-long.1008.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.1008/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-11-20", "record_sha256": "7c1f8dd2ecc3683782c95f0f32c7693348d4d607fd3c54021f629ff5c6a08fb2", "relations": [], "resource_id": "arxiv:2411.13451", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.13451v1", "topics": ["web-navigation", "data-generation"], "updated_at": "2024-11-20"}, "text": "AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts multimodal web agents to unfamiliar websites using a small number of human demonstrations through prompting or meta-learning.\nPlatforms: web\nTopics: web-navigation, data-generation\nPublished: 2024-11-20\nAuthors: Gaurav Verma, Rachneet Kaur, Nishan Srishankar, Zhen Zeng, Tucker Balch, Manuela Veloso\nRecorded arXiv version: v1\n\nPublication edition: AdaptAgent: Adapting Multimodal Web Agents with Few-Shot Learning from Human Demonstrations\nPublication authors: Gaurav Verma, Rachneet Kaur, Nishan Srishankar, Zhen Zeng, Tucker Balch, Manuela Veloso\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 20635-20651\nPublication doi: 10.18653/v1/2025.acl-long.1008\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.1008/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.13451v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-13451/"}
{"id": "arxiv:2411.13591", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b639d016daf616a5f58656d2a291505cd82f7765c47638a515e359210fd56779", "submitted_at": "2024-11-18T05:47:12Z", "updated_at": "2025-09-11T16:37:00Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2411.13591v7"}, "authors": ["Anthony Nguyen"], "canonical_source_url": "https://arxiv.org/abs/2411.13591", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-13591/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b639d016daf616a5f58656d2a291505cd82f7765c47638a515e359210fd56779", "submitted_at": "2024-11-18T05:47:12Z", "updated_at": "2025-09-11T16:37:00Z", "version": 7, "versioned_url": "https://arxiv.org/abs/2411.13591v7"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.13591"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.13591"}], "name": "Improved GUI Grounding via Iterative Narrowing", "publication": null, "published_at": "2024-11-18", "record_sha256": "febc288eb9c4bcf6c94292f300ebb2af7beca5cc0cf94f0d41bb434bef77aaaf", "relations": [], "resource_id": "arxiv:2411.13591", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.13591v7", "topics": ["grounding"], "updated_at": "2025-09-11"}, "text": "Improved GUI Grounding via Iterative Narrowing\nResource type: paper\nReview status: curated\nScope: core\n\nUses iterative visual narrowing to improve localization of requested GUI elements with general or specially trained vision-language models.\nPlatforms: cross-platform\nTopics: grounding\nPublished: 2024-11-18\nAuthors: Anthony Nguyen\nRecorded arXiv version: v7\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.13591v7\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-13591/"}
{"id": "arxiv:2411.15004", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ef5e3848c3395a5620c5af422e6ebf198e405cfde1314cfac50d347daf68a429", "submitted_at": "2024-11-22T15:26:23Z", "updated_at": "2024-12-05T02:00:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.15004v2"}, "authors": ["Junhong Shen", "Atishay Jain", "Zedian Xiao", "Ishan Amlekar", "Mouad Hadji", "Aaron Podolny", "Ameet Talwalkar"], "canonical_source_url": "https://arxiv.org/abs/2411.15004", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-15004/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ef5e3848c3395a5620c5af422e6ebf198e405cfde1314cfac50d347daf68a429", "submitted_at": "2024-11-22T15:26:23Z", "updated_at": "2024-12-05T02:00:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2411.15004v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.15004"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.15004"}], "name": "ScribeAgent: Towards Specialized Web Agents Using Production-Scale Workflow Data", "publication": null, "published_at": "2024-11-22", "record_sha256": "dbb0911d8bd17a90d68580884007710a2921780ae366bb5cfba5586f6cdc7d4f", "relations": [], "resource_id": "arxiv:2411.15004", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.15004v2", "topics": ["workflow", "web-navigation"], "updated_at": "2024-12-05"}, "text": "ScribeAgent: Towards Specialized Web Agents Using Production-Scale Workflow Data\nResource type: paper\nReview status: curated\nScope: core\n\nStudies fine-tuning open language models on production web-workflow traces, including training choices for specialized web agents.\nPlatforms: web\nTopics: workflow, web-navigation\nPublished: 2024-11-22\nAuthors: Junhong Shen, Atishay Jain, Zedian Xiao, Ishan Amlekar, Mouad Hadji, Aaron Podolny, Ameet Talwalkar\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.15004v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-15004/"}
{"id": "arxiv:2411.17465", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c8f5c0a7fc5f194ffa6ca914619b9402c947d15f100b124505883ad8ff23ac7", "submitted_at": "2024-11-26T14:29:47Z", "updated_at": "2024-11-26T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.17465v1"}, "authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2411.17465", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-17465/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8c8f5c0a7fc5f194ffa6ca914619b9402c947d15f100b124505883ad8ff23ac7", "submitted_at": "2024-11-26T14:29:47Z", "updated_at": "2024-11-26T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2411.17465v1"}, "publication": {"authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Stan Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2025, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "19498-19508", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf", "publisher": null, "series": null, "title": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:10:00Z", "role": "primary", "sha256": "2a79195c42f55ef363988febb2b7cae1d348c4450a443462b75c96fd20c17256", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2411.17465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.17465"}, {"rel": "publication", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf"}], "name": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "publication": {"authors": ["Kevin Qinghong Lin", "Linjie Li", "Difei Gao", "Zhengyuan Yang", "Shiwei Wu", "Zechen Bai", "Stan Weixian Lei", "Lijuan Wang", "Mike Zheng Shou"], "date_basis": "Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.", "date_parts": [2025, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The publisher page directly links this arXiv identifier.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "19498-19508", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.pdf", "publisher": null, "series": null, "title": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition", "volume": null}, "published_at": "2024-11-26", "record_sha256": "81614c38ce061c38229d2857cd6353146174135e9a5008e9e34b92b758539da0", "relations": [], "resource_id": "arxiv:2411.17465", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.17465v1", "topics": ["grounding", "efficiency"], "updated_at": "2024-11-26"}, "text": "ShowUI: One Vision-Language-Action Model for GUI Visual Agent\nResource type: paper\nReview status: curated\nScope: core\n\nCombines visual token selection, interleaved visual-action history, and curated instruction data in a compact GUI action model.\nPlatforms: cross-platform\nTopics: grounding, efficiency\nPublished: 2024-11-26\nAuthors: Kevin Qinghong Lin, Linjie Li, Difei Gao, Zhengyuan Yang, Shiwei Wu, Zechen Bai, Weixian Lei, Lijuan Wang, Mike Zheng Shou\nRecorded arXiv version: v1\n\nPublication edition: ShowUI: One Vision-Language-Action Model for GUI Visual Agent\nPublication authors: Kevin Qinghong Lin, Linjie Li, Difei Gao, Zhengyuan Yang, Shiwei Wu, Zechen Bai, Stan Weixian Lei, Lijuan Wang, Mike Zheng Shou\nPublication edition date (recorded precision): 2025-06\nPublication venue: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition\nPublication pages: 19498-19508\nPublication date basis: Year from publisher citation metadata and month from the same page's embedded BibTeX; no day is supplied.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://openaccess.thecvf.com/content/CVPR2025/html/Lin_ShowUI_One_Vision-Language-Action_Model_for_GUI_Visual_Agent_CVPR_2025_paper.html\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2411.17465v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-17465/"}
{"id": "arxiv:2411.18279", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f692aa06152820ae60359fc2cc52aaed2ca8a97fe1efb9b8037de973956467a", "submitted_at": "2024-11-27T12:13:39Z", "updated_at": "2025-05-06T15:08:00Z", "version": 12, "versioned_url": "https://arxiv.org/abs/2411.18279v12"}, "authors": ["Chaoyun Zhang", "Shilin He", "Jiaxu Qian", "Bowen Li", "Liqun Li", "Si Qin", "Yu Kang", "Minghua Ma", "Guyue Liu", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2411.18279", "citation_url": "https://awesomecua.github.io/resources/arxiv-2411-18279/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f692aa06152820ae60359fc2cc52aaed2ca8a97fe1efb9b8037de973956467a", "submitted_at": "2024-11-27T12:13:39Z", "updated_at": "2025-05-06T15:08:00Z", "version": 12, "versioned_url": "https://arxiv.org/abs/2411.18279v12"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2411.18279"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2411.18279"}], "name": "Large Language Model-Brained GUI Agents: A Survey", "publication": null, "published_at": "2024-11-27", "record_sha256": "c60e34bc3e12b89b6b80e797d18f7f4dd293f7c476ef82307e19a0ff7964f8a3", "relations": [], "resource_id": "arxiv:2411.18279", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2411.18279v12", "topics": ["survey"], "updated_at": "2025-05-06"}, "text": "Large Language Model-Brained GUI Agents: A Survey\nResource type: survey\nReview status: curated\nScope: core\n\nSurvey of LLM-brained GUI agents that carry out natural-language instructions across web, mobile, and desktop interfaces, organizing research and industry progress.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2024-11-27\nAuthors: Chaoyun Zhang, Shilin He, Jiaxu Qian, Bowen Li, Liqun Li, Si Qin, Yu Kang, Minghua Ma, Guyue Liu, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nRecorded arXiv version: v12\n\nPrimary source: https://arxiv.org/abs/2411.18279v12\nCatalog record: https://awesomecua.github.io/resources/arxiv-2411-18279/"}
{"id": "arxiv:2412.00402", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c22d6a9aafb89e5628198fb4f6ea50fe40b1e955a2fa367571459753cea61df", "submitted_at": "2024-11-30T08:55:39Z", "updated_at": "2024-11-30T08:55:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.00402v1"}, "authors": ["Weikai Xie", "Li Zhang", "Shihe Wang", "Rongjie Yi", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2412.00402", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-00402/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9c22d6a9aafb89e5628198fb4f6ea50fe40b1e955a2fa367571459753cea61df", "submitted_at": "2024-11-30T08:55:39Z", "updated_at": "2024-11-30T08:55:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.00402v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {"arxiv": "2412.00402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.00402"}], "name": "DroidCall: A Dataset for LLM-powered Android Intent Invocation", "publication": null, "published_at": "2024-11-30", "record_sha256": "5b799123653e44b27a76db5dfb69e88331471e0bfb0453305c002260f67c41e0", "relations": [], "resource_id": "arxiv:2412.00402", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2412.00402v1", "topics": ["tool-integration", "data-generation"], "updated_at": "2024-11-30"}, "text": "DroidCall: A Dataset for LLM-powered Android Intent Invocation\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces data and a generation pipeline for translating natural-language requests into Android intent invocations using compact language models.\nPlatforms: android\nTopics: tool-integration, data-generation\nPublished: 2024-11-30\nAuthors: Weikai Xie, Li Zhang, Shihe Wang, Rongjie Yi, Mengwei Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.00402v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-00402/"}
{"id": "arxiv:2412.01268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a4c750017de421d3ca9bf119834dfb96d2fe8a0610f12b30d2fae603e26c66b", "submitted_at": "2024-12-02T08:35:31Z", "updated_at": "2024-12-02T08:35:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.01268v1"}, "authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "canonical_source_url": "https://arxiv.org/abs/2412.01268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-01268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a4c750017de421d3ca9bf119834dfb96d2fe8a0610f12b30d2fae603e26c66b", "submitted_at": "2024-12-02T08:35:31Z", "updated_at": "2024-12-02T08:35:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.01268v1"}, "publication": {"authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.76", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.76/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "1461-1473", "pdf_url": "https://aclanthology.org/2025.findings-acl.76.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.76/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform", "web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "0ec2dd7116411e80f842627d047d06cf60d5921d91350daeb251a9c95a8f7884", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.01268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.01268"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.76/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.76.pdf"}], "name": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "publication": {"authors": ["Yiqin Wang", "Haoji Zhang", "Jingqi Tian", "Yansong Tang"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.76", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.76/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "1461-1473", "pdf_url": "https://aclanthology.org/2025.findings-acl.76.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.76/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-12-02", "record_sha256": "e9dbd754561910311264dcc9f9591e6809176c71561ab3d88315c863135a50c7", "relations": [], "resource_id": "arxiv:2412.01268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.01268v1", "topics": ["grounding", "planning"], "updated_at": "2024-12-02"}, "text": "Ponder & Press: Advancing Visual GUI Agent towards General Computer Control\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates interpretation of user instructions from precise GUI localization using general-purpose and specialized visual-language components.\nPlatforms: cross-platform, web, desktop\nTopics: grounding, planning\nPublished: 2024-12-02\nAuthors: Yiqin Wang, Haoji Zhang, Jingqi Tian, Yansong Tang\nRecorded arXiv version: v1\n\nPublication edition: Ponder & Press: Advancing Visual GUI Agent towards General Computer Control\nPublication authors: Yiqin Wang, Haoji Zhang, Jingqi Tian, Yansong Tang\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 1461-1473\nPublication doi: 10.18653/v1/2025.findings-acl.76\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.findings-acl.76/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.01268v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-01268/"}
{"id": "arxiv:2412.04454", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89dbdf874ace87db12fa321e67ecf818d3964622b045ea7d95e109101820d407", "submitted_at": "2024-12-05T18:58:26Z", "updated_at": "2025-05-05T16:17:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04454v2"}, "authors": ["Yiheng Xu", "Zekun Wang", "Junli Wang", "Dunjie Lu", "Tianbao Xie", "Amrita Saha", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "canonical_source_url": "https://arxiv.org/abs/2412.04454", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-04454/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89dbdf874ace87db12fa321e67ecf818d3964622b045ea7d95e109101820d407", "submitted_at": "2024-12-05T18:58:26Z", "updated_at": "2025-05-05T16:17:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.04454v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.04454"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.04454"}], "name": "Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction", "publication": null, "published_at": "2024-12-05", "record_sha256": "425bdc1097cd70dd1641a725fd8a771df5d7cbfc81b43b564abd7085c53dec61", "relations": [], "resource_id": "arxiv:2412.04454", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.04454v2", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-05-05"}, "text": "Aguvis: Unified Pure Vision Agents for Autonomous GUI Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nUses screenshots, a unified action space, and staged grounding and reasoning training for cross-platform GUI interaction.\nPlatforms: cross-platform\nTopics: grounding, planning, data-generation\nPublished: 2024-12-05\nAuthors: Yiheng Xu, Zekun Wang, Junli Wang, Dunjie Lu, Tianbao Xie, Amrita Saha, Doyen Sahoo, Tao Yu, Caiming Xiong\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.04454v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-04454/"}
{"id": "arxiv:2412.05467", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572a993cd37a83bf7100dca71f0a44473ed6547f521409abf0f4ee22733a28cc", "submitted_at": "2024-12-06T23:43:59Z", "updated_at": "2025-02-28T16:02:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2412.05467v4"}, "authors": ["Thibault Le Sellier De Chezelles", "Maxime Gasse", "Alexandre Drouin", "Massimo Caccia", "Léo Boisvert", "Megh Thakkar", "Tom Marty", "Rim Assouel", "Sahar Omidi Shayegan", "Lawrence Keunho Jang", "Xing Han Lù", "Ori Yoran", "Dehan Kong", "Frank F. Xu", "Siva Reddy", "Quentin Cappart", "Graham Neubig", "Ruslan Salakhutdinov", "Nicolas Chapados", "Alexandre Lacoste"], "canonical_source_url": "https://arxiv.org/abs/2412.05467", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-05467/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572a993cd37a83bf7100dca71f0a44473ed6547f521409abf0f4ee22733a28cc", "submitted_at": "2024-12-06T23:43:59Z", "updated_at": "2025-02-28T16:02:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2412.05467v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.05467"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.05467"}], "name": "The BrowserGym Ecosystem for Web Agent Research", "publication": null, "published_at": "2024-12-06", "record_sha256": "8478c4d89b40df6573028f6ca61b124da1ff2a144ad15ad6d7ae0b803aefe3aa", "relations": [], "resource_id": "arxiv:2412.05467", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.05467v4", "topics": ["evaluation", "sandbox"], "updated_at": "2025-02-28"}, "text": "The BrowserGym Ecosystem for Web Agent Research\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes BrowserGym's shared web interaction environment and AgentLab tooling for consistent experiments across multiple agent benchmarks.\nPlatforms: web\nTopics: evaluation, sandbox\nPublished: 2024-12-06\nAuthors: Thibault Le Sellier De Chezelles, Maxime Gasse, Alexandre Drouin, Massimo Caccia, Léo Boisvert, Megh Thakkar, Tom Marty, Rim Assouel, Sahar Omidi Shayegan, Lawrence Keunho Jang, Xing Han Lù, Ori Yoran, Dehan Kong, Frank F. Xu, Siva Reddy, Quentin Cappart, Graham Neubig, Ruslan Salakhutdinov, Nicolas Chapados, Alexandre Lacoste\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.05467v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-05467/"}
{"id": "arxiv:2412.09362", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee5840bbd1404674f0e9509b305e995c4dece9257e3f2a79348dde9aee99928b", "submitted_at": "2024-12-12T15:29:36Z", "updated_at": "2024-12-12T15:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.09362v1"}, "authors": ["Huawen Shen", "Chang Liu", "Gengluo Li", "Xinlong Wang", "Yu Zhou", "Can Ma", "Xiangyang Ji"], "canonical_source_url": "https://arxiv.org/abs/2412.09362", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-09362/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ee5840bbd1404674f0e9509b305e995c4dece9257e3f2a79348dde9aee99928b", "submitted_at": "2024-12-12T15:29:36Z", "updated_at": "2024-12-12T15:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.09362v1"}}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.09362"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.09362"}], "name": "Falcon-UI: Understanding GUI Before Following User Instructions", "publication": null, "published_at": "2024-12-12", "record_sha256": "7c2ae8bc5d9d195bd4a13e9ab067496bfa23cbcdc582efe5b3bec5c524953fea", "relations": [], "resource_id": "arxiv:2412.09362", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.09362v1", "topics": ["screen-understanding", "data-generation"], "updated_at": "2024-12-12"}, "text": "Falcon-UI: Understanding GUI Before Following User Instructions\nResource type: paper\nReview status: curated\nScope: core\n\nPretrains GUI-context understanding with instruction-free web interactions before fine-tuning Falcon-UI on instructed GUI tasks.\nPlatforms: web, android, cross-platform\nTopics: screen-understanding, data-generation\nPublished: 2024-12-12\nAuthors: Huawen Shen, Chang Liu, Gengluo Li, Xinlong Wang, Yu Zhou, Can Ma, Xiangyang Ji\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.09362v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-09362/"}
{"id": "arxiv:2412.09605", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ee630d434cef63f63c368752aeae14b0a60d49c3fde33b3a79ecbfc8fc4d06e", "submitted_at": "2024-12-12T18:59:27Z", "updated_at": "2025-03-03T18:59:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.09605v2"}, "authors": ["Yiheng Xu", "Dunjie Lu", "Zhennan Shen", "Junli Wang", "Zekun Wang", "Yuchen Mao", "Caiming Xiong", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2412.09605", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-09605/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ee630d434cef63f63c368752aeae14b0a60d49c3fde33b3a79ecbfc8fc4d06e", "submitted_at": "2024-12-12T18:59:27Z", "updated_at": "2025-03-03T18:59:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.09605v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.09605"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.09605"}], "name": "AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials", "publication": null, "published_at": "2024-12-12", "record_sha256": "060066c48d6c937b9a6fc340464ccd9337a33063e192e015543e9bc83309cdf6", "relations": [], "resource_id": "arxiv:2412.09605", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.09605v2", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-03-03"}, "text": "AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials\nResource type: paper\nReview status: curated\nScope: core\n\nConverts public tutorials into task specifications and executes them with an agent to synthesize multimodal interaction trajectories.\nPlatforms: web\nTopics: data-generation, web-navigation\nPublished: 2024-12-12\nAuthors: Yiheng Xu, Dunjie Lu, Zhennan Shen, Junli Wang, Zekun Wang, Yuchen Mao, Caiming Xiong, Tao Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.09605v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-09605/"}
{"id": "arxiv:2412.10047", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54051d974e388ed2a26f1421b8c1fc79472bb941810519039ef48e90bf4761b6", "submitted_at": "2024-12-13T11:19:56Z", "updated_at": "2025-01-13T06:47:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10047v2"}, "authors": ["Lu Wang", "Fangkai Yang", "Chaoyun Zhang", "Junting Lu", "Jiaxu Qian", "Shilin He", "Pu Zhao", "Bo Qiao", "Ray Huang", "Si Qin", "Qisheng Su", "Jiayi Ye", "Yudi Zhang", "Jian-Guang Lou", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang", "Qi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2412.10047", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-10047/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54051d974e388ed2a26f1421b8c1fc79472bb941810519039ef48e90bf4761b6", "submitted_at": "2024-12-13T11:19:56Z", "updated_at": "2025-01-13T06:47:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10047v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:41Z", "role": "primary", "sha256": "0ba893415123ecc1763614ef5fc97e5c8bb7835f10b26b43f1b921ea860484cf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2410.06212%2C2410.06215%2C2410.07283%2C2410.07331%2C2410.07869%2C2410.08209%2C2410.08368%2C2410.10570%2C2410.12481%2C2410.12669%2C2410.15549%2C2410.16950%2C2410.21909%2C2410.22584%2C2410.22770%2C2410.23022%2C2410.23555%2C2411.02006%2C2411.02393%2C2411.02479%2C2411.03312%2C2411.03817%2C2411.05451%2C2411.11145%2C2411.11904%2C2411.11916%2C2411.12990%2C2411.14228%2C2411.15397%2C2412.01857%2C2412.04144%2C2412.04383%2C2412.05819%2C2412.06224%2C2412.06771%2C2412.09530%2C2412.09613%2C2412.10047%2C2412.10380%2C2412.10673%2C2412.12004%2C2412.13178%2C2412.13810%2C2412.14203%2C2412.14222%2C2412.16359%2C2412.19638%2C2412.20206%2C2501.03895%2C2501.05179&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:13:28Z", "role": "primary", "sha256": "aaddc3cae2fcadc9780390fceb663c0200ecb042639e2c69bccea19607cd78ea", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2412.10047v2"}], "identifiers": {"arxiv": "2412.10047"}, "kind": "paper", "last_verified_at": "2026-10-17T09:13:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10047"}], "name": "Large Action Models: From Inception to Implementation", "publication": null, "published_at": "2024-12-13", "record_sha256": "1d99dbf868c804211f22af0811bb2dc09e503af8ed3660498643836d4508b9a6", "relations": [], "resource_id": "arxiv:2412.10047", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.10047v2", "topics": ["automation", "data-generation", "evaluation"], "updated_at": "2025-01-13"}, "text": "Large Action Models: From Inception to Implementation\nResource type: paper\nReview status: curated\nScope: core\n\nUses a Windows-agent case study to describe a development workflow for large action models, covering data collection, training, grounding, environment integration, and evaluation.\nPlatforms: desktop, windows\nTopics: automation, data-generation, evaluation\nPublished: 2024-12-13\nAuthors: Lu Wang, Fangkai Yang, Chaoyun Zhang, Junting Lu, Jiaxu Qian, Shilin He, Pu Zhao, Bo Qiao, Ray Huang, Si Qin, Qisheng Su, Jiayi Ye, Yudi Zhang, Jian-Guang Lou, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang, Qi Zhang\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its normalized abstract SHA-256 matches the retained bibliographic snapshot. The original API response body was not locally available for this review. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.10047v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-10047/"}
{"id": "arxiv:2412.10342", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7cab57b118ab76a9026e7ee800709f4e93969e19ad18f549f8b75ee657622b", "submitted_at": "2024-12-13T18:40:10Z", "updated_at": "2025-02-03T15:23:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10342v2"}, "authors": ["Zhiqi Ge", "Juncheng Li", "Xinglei Pang", "Minghe Gao", "Kaihang Pan", "Wang Lin", "Hao Fei", "Wenqiao Zhang", "Siliang Tang", "Yueting Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2412.10342", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-10342/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e7cab57b118ab76a9026e7ee800709f4e93969e19ad18f549f8b75ee657622b", "submitted_at": "2024-12-13T18:40:10Z", "updated_at": "2025-02-03T15:23:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.10342v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.10342"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10342"}], "name": "Iris: Breaking GUI Complexity with Adaptive Focus and Self-Refining", "publication": null, "published_at": "2024-12-13", "record_sha256": "73bcc165d51233fd494c6741fe8fc7e551adee3f5e710eba507801f11f8e9d82", "relations": [], "resource_id": "arxiv:2412.10342", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.10342v2", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-02-03"}, "text": "Iris: Breaking GUI Complexity with Adaptive Focus and Self-Refining\nResource type: paper\nReview status: curated\nScope: core\n\nCombines information-sensitive image cropping with a learning loop between element description and grounding for complex interface understanding.\nPlatforms: web, desktop\nTopics: grounding, screen-understanding, efficiency\nPublished: 2024-12-13\nAuthors: Zhiqi Ge, Juncheng Li, Xinglei Pang, Minghe Gao, Kaihang Pan, Wang Lin, Hao Fei, Wenqiao Zhang, Siliang Tang, Yueting Zhuang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.10342v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-10342/"}
{"id": "arxiv:2412.10742", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2abbae042d20381aa218e6a2b14a1fc28074543e430ad09031237d22587be7a", "submitted_at": "2024-12-14T08:25:28Z", "updated_at": "2024-12-14T08:25:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10742v1"}, "authors": ["Jiarun Liu", "Jia Hao", "Chunhong Zhang", "Zheng Hu"], "canonical_source_url": "https://arxiv.org/abs/2412.10742", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-10742/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2abbae042d20381aa218e6a2b14a1fc28074543e430ad09031237d22587be7a", "submitted_at": "2024-12-14T08:25:28Z", "updated_at": "2024-12-14T08:25:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10742v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.10742"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10742"}], "name": "WEPO: Web Element Preference Optimization for LLM-based Web Navigation", "publication": null, "published_at": "2024-12-14", "record_sha256": "053ba97a4fe5536fd6ea4edf942ae3715a5d012b1fae73e9875d2f54f093fdc6", "relations": [], "resource_id": "arxiv:2412.10742", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.10742v1", "topics": ["grounding", "web-navigation"], "updated_at": "2024-12-14"}, "text": "WEPO: Web Element Preference Optimization for LLM-based Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nUses web-element preference optimization with sampled negative HTML elements to align high-level user intent with navigation actions.\nPlatforms: web\nTopics: grounding, web-navigation\nPublished: 2024-12-14\nAuthors: Jiarun Liu, Jia Hao, Chunhong Zhang, Zheng Hu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.10742v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-10742/"}
{"id": "arxiv:2412.10840", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e0aefcd3eca9e37d25a2c5195c0d502caec9be83b8f6b4b7ee5778bff9e150", "submitted_at": "2024-12-14T14:30:05Z", "updated_at": "2024-12-14T14:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10840v1"}, "authors": ["Hai-Ming Xu", "Qi Chen", "Lei Wang", "Lingqiao Liu"], "canonical_source_url": "https://arxiv.org/abs/2412.10840", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-10840/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e0aefcd3eca9e37d25a2c5195c0d502caec9be83b8f6b4b7ee5778bff9e150", "submitted_at": "2024-12-14T14:30:05Z", "updated_at": "2024-12-14T14:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.10840v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.10840"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.10840"}], "name": "Attention-driven GUI Grounding: Leveraging Pretrained Multimodal Large Language Models without Fine-Tuning", "publication": null, "published_at": "2024-12-14", "record_sha256": "620f7e78528565892436d1629dee62917406ff782d7adc2e0614930dd3fb5887", "relations": [], "resource_id": "arxiv:2412.10840", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.10840v1", "topics": ["grounding"], "updated_at": "2024-12-14"}, "text": "Attention-driven GUI Grounding: Leveraging Pretrained Multimodal Large Language Models without Fine-Tuning\nResource type: paper\nReview status: curated\nScope: core\n\nLocates GUI targets by identifying and aggregating attention maps in pretrained multimodal models without task-specific fine-tuning.\nPlatforms: cross-platform\nTopics: grounding\nPublished: 2024-12-14\nAuthors: Hai-Ming Xu, Qi Chen, Lei Wang, Lingqiao Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.10840v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-10840/"}
{"id": "arxiv:2412.13194", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362d2a470603bece43d2205e56dd633e7310cfad2ba2d91b0b765abcbcc41b2f", "submitted_at": "2024-12-17T18:59:50Z", "updated_at": "2024-12-17T18:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.13194v1"}, "authors": ["Yifei Zhou", "Qianlan Yang", "Kaixiang Lin", "Min Bai", "Xiong Zhou", "Yu-Xiong Wang", "Sergey Levine", "Erran Li"], "canonical_source_url": "https://arxiv.org/abs/2412.13194", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-13194/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362d2a470603bece43d2205e56dd633e7310cfad2ba2d91b0b765abcbcc41b2f", "submitted_at": "2024-12-17T18:59:50Z", "updated_at": "2024-12-17T18:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.13194v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.13194"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.13194"}], "name": "Proposer-Agent-Evaluator(PAE): Autonomous Skill Discovery For Foundation Model Internet Agents", "publication": null, "published_at": "2024-12-17", "record_sha256": "f3c4446a9a6e3e47fb740db19e63548d9662cdf0b2bc99e82e8c80fd1e6bee5c", "relations": [], "resource_id": "arxiv:2412.13194", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.13194v1", "topics": ["data-generation", "reinforcement-learning", "web-navigation"], "updated_at": "2024-12-17"}, "text": "Proposer-Agent-Evaluator(PAE): Autonomous Skill Discovery For Foundation Model Internet Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines autonomous task proposal, web interaction, and visual success evaluation to discover skills and improve browser-agent policies.\nPlatforms: web\nTopics: data-generation, reinforcement-learning, web-navigation\nPublished: 2024-12-17\nAuthors: Yifei Zhou, Qianlan Yang, Kaixiang Lin, Min Bai, Xiong Zhou, Yu-Xiong Wang, Sergey Levine, Erran Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.13194v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-13194/"}
{"id": "arxiv:2412.13501", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8f83d0490e70db1ae9c55c071935fb1f59e1358335c1bd3f50835611f4360c", "submitted_at": "2024-12-18T04:48:28Z", "updated_at": "2025-09-26T21:01:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.13501v3"}, "authors": ["Dang Nguyen", "Jian Chen", "Yu Wang", "Gang Wu", "Namyong Park", "Zhengmian Hu", "Hanjia Lyu", "Junda Wu", "Ryan Aponte", "Yu Xia", "Xintong Li", "Jing Shi", "Hongjie Chen", "Viet Dac Lai", "Zhouhang Xie", "Sungchul Kim", "Ruiyi Zhang", "Tong Yu", "Mehrab Tanjim", "Nesreen K. Ahmed", "Puneet Mathur", "Seunghyun Yoon", "Lina Yao", "Branislav Kveton", "Jihyung Kil", "Thien Huu Nguyen", "Trung Bui", "Tianyi Zhou", "Ryan A. Rossi", "Franck Dernoncourt"], "canonical_source_url": "https://arxiv.org/abs/2412.13501", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-13501/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8f83d0490e70db1ae9c55c071935fb1f59e1358335c1bd3f50835611f4360c", "submitted_at": "2024-12-18T04:48:28Z", "updated_at": "2025-09-26T21:01:59Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.13501v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.13501"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.13501"}], "name": "GUI Agents: A Survey", "publication": null, "published_at": "2024-12-18", "record_sha256": "332bae01dcdecbb7052e412d72780e4d62e958403f4e50745d248c83a55b79ee", "relations": [], "resource_id": "arxiv:2412.13501", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.13501v3", "topics": ["survey"], "updated_at": "2025-09-26"}, "text": "GUI Agents: A Survey\nResource type: survey\nReview status: curated\nScope: core\n\nCategorizes GUI-agent benchmarks, evaluation metrics, architectures, and training methods, and proposes a unified perception, reasoning, planning, and acting framework.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2024-12-18\nAuthors: Dang Nguyen, Jian Chen, Yu Wang, Gang Wu, Namyong Park, Zhengmian Hu, Hanjia Lyu, Junda Wu, Ryan Aponte, Yu Xia, Xintong Li, Jing Shi, Hongjie Chen, Viet Dac Lai, Zhouhang Xie, Sungchul Kim, Ruiyi Zhang, Tong Yu, Mehrab Tanjim, Nesreen K. Ahmed, Puneet Mathur, Seunghyun Yoon, Lina Yao, Branislav Kveton, Jihyung Kil, Thien Huu Nguyen, Trung Bui, Tianyi Zhou, Ryan A. Rossi, Franck Dernoncourt\nRecorded arXiv version: v3\n\nPrimary source: https://arxiv.org/abs/2412.13501v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-13501/"}
{"id": "arxiv:2412.14161", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f91dd9ec9f50fb5526a78cfe270a7a1705d5bd83a2aef26644453740c005d5", "submitted_at": "2024-12-18T18:55:40Z", "updated_at": "2025-09-10T08:35:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.14161v3"}, "authors": ["Frank F. Xu", "Yufan Song", "Boxuan Li", "Yuxuan Tang", "Kritanjali Jain", "Mengxue Bao", "Zora Z. Wang", "Xuhui Zhou", "Zhitong Guo", "Murong Cao", "Mingyang Yang", "Hao Yang Lu", "Amaad Martin", "Zhe Su", "Leander Maben", "Raj Mehta", "Wayne Chi", "Lawrence Jang", "Yiqing Xie", "Shuyan Zhou", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2412.14161", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-14161/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f91dd9ec9f50fb5526a78cfe270a7a1705d5bd83a2aef26644453740c005d5", "submitted_at": "2024-12-18T18:55:40Z", "updated_at": "2025-09-10T08:35:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.14161v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.14161"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.14161"}], "name": "TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks", "publication": null, "published_at": "2024-12-18", "record_sha256": "4844c0dd69e1f56de5086d9fb9a5c22d1ed5aa79718d960da2cc69c36aec05d8", "relations": [], "resource_id": "arxiv:2412.14161", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.14161v3", "topics": ["evaluation", "long-horizon", "workflow"], "updated_at": "2025-09-10"}, "text": "TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nSimulates a software-company workplace to evaluate agents on professional tasks involving websites, programs, files, and coworker communication.\nPlatforms: web, desktop\nTopics: evaluation, long-horizon, workflow\nPublished: 2024-12-18\nAuthors: Frank F. Xu, Yufan Song, Boxuan Li, Yuxuan Tang, Kritanjali Jain, Mengxue Bao, Zora Z. Wang, Xuhui Zhou, Zhitong Guo, Murong Cao, Mingyang Yang, Hao Yang Lu, Amaad Martin, Zhe Su, Leander Maben, Raj Mehta, Wayne Chi, Lawrence Jang, Yiqing Xie, Shuyan Zhou, Graham Neubig\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.14161v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-14161/"}
{"id": "arxiv:2412.16256", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b49d45aa539c224846a7e14cee9b537d032228e19e86a9cb3cdd9f5004aa429", "submitted_at": "2024-12-20T07:16:57Z", "updated_at": "2025-07-08T08:49:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.16256v2"}, "authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "canonical_source_url": "https://arxiv.org/abs/2412.16256", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-16256/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b49d45aa539c224846a7e14cee9b537d032228e19e86a9cb3cdd9f5004aa429", "submitted_at": "2024-12-20T07:16:57Z", "updated_at": "2025-07-08T08:49:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2412.16256v2"}, "publication": {"authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1152", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1152/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "22418-22433", "pdf_url": "https://aclanthology.org/2025.findings-acl.1152.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Aria-UI: Visual Grounding for GUI Instructions", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1152/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "e24a63769d700caa2d27341dd2634b990cdbc53c7cf557817b3679dc63faef6f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.16256"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.16256"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.1152/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.1152.pdf"}], "name": "Aria-UI: Visual Grounding for GUI Instructions", "publication": {"authors": ["Yuhao Yang", "Yue Wang", "Dongxu Li", "Ziyang Luo", "Bei Chen", "Chao Huang", "Junnan Li"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.1152", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.1152/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "22418-22433", "pdf_url": "https://aclanthology.org/2025.findings-acl.1152.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Aria-UI: Visual Grounding for GUI Instructions", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.1152/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2024-12-20", "record_sha256": "a3a5bdcf8359ffdad7508eddc938a5050e41a7987cc3c60bf19aed435e59e8f8", "relations": [], "resource_id": "arxiv:2412.16256", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.16256v2", "topics": ["grounding", "memory"], "updated_at": "2025-07-08"}, "text": "Aria-UI: Visual Grounding for GUI Instructions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a vision-based grounding model trained with diverse synthetic instructions and action histories for context-sensitive GUI targeting.\nPlatforms: cross-platform\nTopics: grounding, memory\nPublished: 2024-12-20\nAuthors: Yuhao Yang, Yue Wang, Dongxu Li, Ziyang Luo, Bei Chen, Chao Huang, Junnan Li\nRecorded arXiv version: v2\n\nPublication edition: Aria-UI: Visual Grounding for GUI Instructions\nPublication authors: Yuhao Yang, Yue Wang, Dongxu Li, Ziyang Luo, Bei Chen, Chao Huang, Junnan Li\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 22418-22433\nPublication doi: 10.18653/v1/2025.findings-acl.1152\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.findings-acl.1152/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.16256v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-16256/"}
{"id": "arxiv:2412.17589", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db567d35887fe5367e456b35bc347400b1bc2d147730e9d34c158e1fd7c8dd6c", "submitted_at": "2024-12-23T14:02:12Z", "updated_at": "2024-12-23T14:02:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.17589v1"}, "authors": ["Yanheng He", "Jiahe Jin", "Shijie Xia", "Jiadi Su", "Runze Fan", "Haoyang Zou", "Xiangkun Hu", "Pengfei Liu"], "canonical_source_url": "https://arxiv.org/abs/2412.17589", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-17589/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db567d35887fe5367e456b35bc347400b1bc2d147730e9d34c158e1fd7c8dd6c", "submitted_at": "2024-12-23T14:02:12Z", "updated_at": "2024-12-23T14:02:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.17589v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.17589"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.17589"}], "name": "PC Agent: While You Sleep, AI Works -- A Cognitive Journey into Digital World", "publication": null, "published_at": "2024-12-23", "record_sha256": "34fbff9913f07ea2cb32758a4337d7722a179360d6b48e4b53cf499ec42f1afe", "relations": [], "resource_id": "arxiv:2412.17589", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.17589v1", "topics": ["recording", "data-generation", "planning"], "updated_at": "2024-12-23"}, "text": "PC Agent: While You Sleep, AI Works -- A Cognitive Journey into Digital World\nResource type: paper\nReview status: curated\nScope: core\n\nCombines PC Tracker interaction capture, completion of action and reasoning annotations, and planning and grounding agents for desktop workflows.\nPlatforms: desktop\nTopics: recording, data-generation, planning\nPublished: 2024-12-23\nAuthors: Yanheng He, Jiahe Jin, Shijie Xia, Jiadi Su, Runze Fan, Haoyang Zou, Xiangkun Hu, Pengfei Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.17589v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-17589/"}
{"id": "arxiv:2412.18116", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3190c452dbfa24b131cc34f66192bbdb9c538d30a1919082a9fed334f6169fff", "submitted_at": "2024-12-24T02:54:56Z", "updated_at": "2025-05-06T12:24:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.18116v3"}, "authors": ["Hao Wen", "Shizuo Tian", "Borislav Pavlov", "Wenjie Du", "Yixuan Li", "Ge Chang", "Shanhui Zhao", "Jiacheng Liu", "Yunxin Liu", "Ya-Qin Zhang", "Yuanchun Li"], "canonical_source_url": "https://arxiv.org/abs/2412.18116", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-18116/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3190c452dbfa24b131cc34f66192bbdb9c538d30a1919082a9fed334f6169fff", "submitted_at": "2024-12-24T02:54:56Z", "updated_at": "2025-05-06T12:24:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.18116v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.18116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.18116"}], "name": "AutoDroid-V2: Boosting SLM-based GUI Agents via Code Generation", "publication": null, "published_at": "2024-12-24", "record_sha256": "2be9b459db2bbb516879d70bf401c10096a5d28f730b6cc2cbb463e1115c48b4", "relations": [], "resource_id": "arxiv:2412.18116", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.18116v3", "topics": ["mobile-navigation", "tool-integration"], "updated_at": "2025-05-06"}, "text": "AutoDroid-V2: Boosting SLM-based GUI Agents via Code Generation\nResource type: paper\nReview status: curated\nScope: core\n\nRecasts mobile UI automation as on-device code generation, using app-specific documentation and synthetic tasks to guide smaller models.\nPlatforms: android\nTopics: mobile-navigation, tool-integration\nPublished: 2024-12-24\nAuthors: Hao Wen, Shizuo Tian, Borislav Pavlov, Wenjie Du, Yixuan Li, Ge Chang, Shanhui Zhao, Jiacheng Liu, Yunxin Liu, Ya-Qin Zhang, Yuanchun Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.18116v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-18116/"}
{"id": "arxiv:2412.18426", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e509f015068bcf1294a97236ff3bbde5e3ceb167f3cb032fcfd84fc4a5b01d0", "submitted_at": "2024-12-24T13:41:47Z", "updated_at": "2024-12-24T13:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.18426v1"}, "authors": ["Kangjia Zhao", "Jiahui Song", "Leigang Sha", "Haozhan Shen", "Zhi Chen", "Tiancheng Zhao", "Xiubo Liang", "Jianwei Yin"], "canonical_source_url": "https://arxiv.org/abs/2412.18426", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-18426/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e509f015068bcf1294a97236ff3bbde5e3ceb167f3cb032fcfd84fc4a5b01d0", "submitted_at": "2024-12-24T13:41:47Z", "updated_at": "2024-12-24T13:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2412.18426v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2412.18426"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.18426"}], "name": "GUI Testing Arena: A Unified Benchmark for Advancing Autonomous GUI Testing Agent", "publication": null, "published_at": "2024-12-24", "record_sha256": "402ec559e726a69d2438fb6f366dd721c571b9aa442c247c23557d88e5fb3385", "relations": [], "resource_id": "arxiv:2412.18426", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.18426v1", "topics": ["evaluation", "automation", "mobile-navigation"], "updated_at": "2024-12-24"}, "text": "GUI Testing Arena: A Unified Benchmark for Advancing Autonomous GUI Testing Agent\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates GUI testing through test-intent generation, execution, and defect detection using real, modified, and synthetic mobile application data.\nPlatforms: unspecified\nTopics: evaluation, automation, mobile-navigation\nPublished: 2024-12-24\nAuthors: Kangjia Zhao, Jiahui Song, Leigang Sha, Haozhan Shen, Zhi Chen, Tiancheng Zhao, Xiubo Liang, Jianwei Yin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.18426v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-18426/"}
{"id": "arxiv:2412.19723", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40b5e3090c25f80d8d786f623812459b329acc2772f945419a585ffb27c54db3", "submitted_at": "2024-12-27T16:21:58Z", "updated_at": "2025-06-27T08:25:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.19723v3"}, "authors": ["Qiushi Sun", "Kanzhi Cheng", "Zichen Ding", "Chuanyang Jin", "Yian Wang", "Fangzhi Xu", "Zhenyu Wu", "Chengyou Jia", "Liheng Chen", "Zhoumianze Liu", "Ben Kao", "Guohao Li", "Junxian He", "Yu Qiao", "Zhiyong Wu"], "canonical_source_url": "https://arxiv.org/abs/2412.19723", "citation_url": "https://awesomecua.github.io/resources/arxiv-2412-19723/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40b5e3090c25f80d8d786f623812459b329acc2772f945419a585ffb27c54db3", "submitted_at": "2024-12-27T16:21:58Z", "updated_at": "2025-06-27T08:25:48Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2412.19723v3"}, "publication": {"authors": ["Qiushi Sun", "Kanzhi Cheng", "Zichen Ding", "Chuanyang Jin", "Yian Wang", "Fangzhi Xu", "Zhenyu Wu", "Chengyou Jia", "Liheng Chen", "Zhoumianze Liu", "Ben Kao", "Guohao Li", "Junxian He", "Yu Qiao", "Zhiyong Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.277", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.277/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5555-5579", "pdf_url": "https://aclanthology.org/2025.acl-long.277.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.277/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:57Z", "role": "primary", "sha256": "ea538c35ce448bce6e03932b4a3e20815503c461907234b3107b3f8ccbc5481d", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.277/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2412.19723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2412.19723"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.277/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.277.pdf"}], "name": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis", "publication": {"authors": ["Qiushi Sun", "Kanzhi Cheng", "Zichen Ding", "Chuanyang Jin", "Yian Wang", "Fangzhi Xu", "Zhenyu Wu", "Chengyou Jia", "Liheng Chen", "Zhoumianze Liu", "Ben Kao", "Guohao Li", "Junxian He", "Yu Qiao", "Zhiyong Wu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.277", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.277/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5555-5579", "pdf_url": "https://aclanthology.org/2025.acl-long.277.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.277/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2024-12-27", "record_sha256": "0bcbb519c9e404ca70a3e7e296555b9f685295bef63edb78cf20277ce4cf40a5", "relations": [], "resource_id": "arxiv:2412.19723", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2412.19723v3", "topics": ["data-generation"], "updated_at": "2025-06-27"}, "text": "OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds GUI trajectories by exploring environments first and deriving tasks retrospectively, with reward-model filtering of generated trajectories.\nPlatforms: cross-platform\nTopics: data-generation\nPublished: 2024-12-27\nAuthors: Qiushi Sun, Kanzhi Cheng, Zichen Ding, Chuanyang Jin, Yian Wang, Fangzhi Xu, Zhenyu Wu, Chengyou Jia, Liheng Chen, Zhoumianze Liu, Ben Kao, Guohao Li, Junxian He, Yu Qiao, Zhiyong Wu\nRecorded arXiv version: v3\n\nPublication edition: OS-Genesis: Automating GUI Agent Trajectory Construction via Reverse Task Synthesis\nPublication authors: Qiushi Sun, Kanzhi Cheng, Zichen Ding, Chuanyang Jin, Yian Wang, Fangzhi Xu, Zhenyu Wu, Chengyou Jia, Liheng Chen, Zhoumianze Liu, Ben Kao, Guohao Li, Junxian He, Yu Qiao, Zhiyong Wu\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 5555-5579\nPublication doi: 10.18653/v1/2025.acl-long.277\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.277/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2412.19723v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2412-19723/"}
{"id": "arxiv:2501.01149", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ac7100a0f3cc3f004672eb6923b1f0ecb43c3d98da6959f0ff7514d0fb2f442", "submitted_at": "2025-01-02T09:03:56Z", "updated_at": "2026-01-12T07:01:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.01149v3"}, "authors": ["Yuxiang Chai", "Shunye Tang", "Han Xiao", "Weifeng Lin", "Hanhao Li", "Jiayu Zhang", "Liang Liu", "Pengxiang Zhao", "Guangyi Liu", "Guozhi Wang", "Shuai Ren", "Rongduo Han", "Haining Zhang", "Siyuan Huang", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2501.01149", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-01149/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ac7100a0f3cc3f004672eb6923b1f0ecb43c3d98da6959f0ff7514d0fb2f442", "submitted_at": "2025-01-02T09:03:56Z", "updated_at": "2026-01-12T07:01:09Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.01149v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2501.01149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.01149"}], "name": "A3: Android Agent Arena for Mobile GUI Agents with Essential-State Procedural Evaluation", "publication": null, "published_at": "2025-01-02", "record_sha256": "ac65245285b829efeda0e9c7a2fe972bd54e648905f0173fb03329378fcdf597", "relations": [], "resource_id": "arxiv:2501.01149", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.01149v3", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2026-01-12"}, "text": "A3: Android Agent Arena for Mobile GUI Agents with Essential-State Procedural Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Android Agent Arena with online application tasks and procedural evaluation based on reaching essential intermediate states.\nPlatforms: android\nTopics: evaluation, mobile-navigation\nPublished: 2025-01-02\nAuthors: Yuxiang Chai, Shunye Tang, Han Xiao, Weifeng Lin, Hanhao Li, Jiayu Zhang, Liang Liu, Pengxiang Zhao, Guangyi Liu, Guozhi Wang, Shuai Ren, Rongduo Han, Haining Zhang, Siyuan Huang, Hongsheng Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.01149v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-01149/"}
{"id": "arxiv:2501.04575", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18d5b2e34f18c24e5d60d5564c9b08eced39c1b17bf5dac8d997640786688fba", "submitted_at": "2025-01-08T15:45:21Z", "updated_at": "2025-01-08T15:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.04575v1"}, "authors": ["Yuhang Liu", "Pengxiang Li", "Zishu Wei", "Congkai Xie", "Xueyu Hu", "Xinchen Xu", "Shengyu Zhang", "Xiaotian Han", "Hongxia Yang", "Fei Wu"], "canonical_source_url": "https://arxiv.org/abs/2501.04575", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-04575/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "18d5b2e34f18c24e5d60d5564c9b08eced39c1b17bf5dac8d997640786688fba", "submitted_at": "2025-01-08T15:45:21Z", "updated_at": "2025-01-08T15:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.04575v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2501.04575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.04575"}], "name": "InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection", "publication": null, "published_at": "2025-01-08", "record_sha256": "f9c0287f2e1db02775c4d8e254a136510bf1ba63719ec1bd6f679127fe0eedd6", "relations": [], "resource_id": "arxiv:2501.04575", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.04575v1", "topics": ["grounding", "planning"], "updated_at": "2025-01-08"}, "text": "InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection\nResource type: paper\nReview status: curated\nScope: core\n\nUses staged supervised training to combine GUI understanding and grounding with hierarchical reasoning and expectation-based reflection.\nPlatforms: cross-platform\nTopics: grounding, planning\nPublished: 2025-01-08\nAuthors: Yuhang Liu, Pengxiang Li, Zishu Wei, Congkai Xie, Xueyu Hu, Xinchen Xu, Shengyu Zhang, Xiaotian Han, Hongxia Yang, Fei Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.04575v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-04575/"}
{"id": "arxiv:2501.07572", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cf550b1e1d5b671b3a4c834787cb67c32c01f508494c5610dac91e95bd92ba", "submitted_at": "2025-01-13T18:58:07Z", "updated_at": "2025-08-10T05:59:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.07572v3"}, "authors": ["Jialong Wu", "Wenbiao Yin", "Yong Jiang", "Zhenglin Wang", "Zekun Xi", "Runnan Fang", "Linhai Zhang", "Yulan He", "Deyu Zhou", "Pengjun Xie", "Fei Huang"], "canonical_source_url": "https://arxiv.org/abs/2501.07572", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-07572/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cf550b1e1d5b671b3a4c834787cb67c32c01f508494c5610dac91e95bd92ba", "submitted_at": "2025-01-13T18:58:07Z", "updated_at": "2025-08-10T05:59:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.07572v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2501.07572"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.07572"}], "name": "WebWalker: Benchmarking LLMs in Web Traversal", "publication": null, "published_at": "2025-01-13", "record_sha256": "fe277d71eca31d337757ab075d1fc871edb2cc47860cbeb4bc52acc936d22d4a", "relations": [], "resource_id": "arxiv:2501.07572", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.07572v3", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-08-10"}, "text": "WebWalker: Benchmarking LLMs in Web Traversal\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebWalkerQA and an exploration-and-critique framework for gathering information by traversing multiple levels of a website.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2025-01-13\nAuthors: Jialong Wu, Wenbiao Yin, Yong Jiang, Zhenglin Wang, Zekun Xi, Runnan Fang, Linhai Zhang, Yulan He, Deyu Zhou, Pengjun Xie, Fei Huang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.07572v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-07572/"}
{"id": "arxiv:2501.11425", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "692a6984b464367c99beb9c5fbd4bbd8378e79f4a4b749503c8fbb72dd80b16f", "submitted_at": "2025-01-20T11:46:04Z", "updated_at": "2025-03-24T10:18:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.11425v3"}, "authors": ["Siyu Yuan", "Zehui Chen", "Zhiheng Xi", "Junjie Ye", "Zhengyin Du", "Jiecao Chen"], "canonical_source_url": "https://arxiv.org/abs/2501.11425", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-11425/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "692a6984b464367c99beb9c5fbd4bbd8378e79f4a4b749503c8fbb72dd80b16f", "submitted_at": "2025-01-20T11:46:04Z", "updated_at": "2025-03-24T10:18:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.11425v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "c35562ab132e13d69155842ec63a2f4189fe8703822e5620df2592a154d2d7ed", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2501.11425"}], "identifiers": {"arxiv": "2501.11425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.11425"}], "name": "Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training", "publication": null, "published_at": "2025-01-20", "record_sha256": "4a59d244108a2b47c80260778000143e7bd6478a509b872a63eca1a71479b43f", "relations": [], "resource_id": "arxiv:2501.11425", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2501.11425v3", "topics": ["data-generation", "planning", "robustness"], "updated_at": "2025-03-24"}, "text": "Agent-R: Training Language Model Agents to Reflect via Iterative Self-Training\nResource type: paper\nReview status: curated\nScope: adjacent\n\nBuilds self-critique training examples from searched recovery paths to improve agents' ability to detect and correct errors during interaction.\nPlatforms: web\nTopics: data-generation, planning, robustness\nPublished: 2025-01-20\nAuthors: Siyu Yuan, Zehui Chen, Zhiheng Xi, Junjie Ye, Zhengyin Du, Jiecao Chen\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.\n\nPrimary source: https://arxiv.org/abs/2501.11425v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-11425/"}
{"id": "arxiv:2501.11733", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cf28f084cac7fc38e5c2dc4654115356d2d57f16009aad8925e071b8005be9c", "submitted_at": "2025-01-20T20:35:46Z", "updated_at": "2025-01-28T16:58:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.11733v2"}, "authors": ["Zhenhailong Wang", "Haiyang Xu", "Junyang Wang", "Xi Zhang", "Ming Yan", "Ji Zhang", "Fei Huang", "Heng Ji"], "canonical_source_url": "https://arxiv.org/abs/2501.11733", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-11733/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5cf28f084cac7fc38e5c2dc4654115356d2d57f16009aad8925e071b8005be9c", "submitted_at": "2025-01-20T20:35:46Z", "updated_at": "2025-01-28T16:58:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.11733v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2501.11733"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.11733"}], "name": "Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Tasks", "publication": null, "published_at": "2025-01-20", "record_sha256": "75315d2a976a1701478588334312518f4fbbe3eda028299cabce29cc1fc389d6", "relations": [], "resource_id": "arxiv:2501.11733", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.11733v2", "topics": ["mobile-navigation", "memory", "long-horizon"], "updated_at": "2025-01-28"}, "text": "Mobile-Agent-E: Self-Evolving Mobile Assistant for Complex Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nUses hierarchical mobile agents and persistent tips and reusable action shortcuts to improve complex tasks from accumulated experience.\nPlatforms: unspecified\nTopics: mobile-navigation, memory, long-horizon\nPublished: 2025-01-20\nAuthors: Zhenhailong Wang, Haiyang Xu, Junyang Wang, Xi Zhang, Ming Yan, Ji Zhang, Fei Huang, Heng Ji\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.11733v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-11733/"}
{"id": "arxiv:2501.12326", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19aa758e66411a31dd1fea6f07c83d1c828bd4aad86f1564a9d7f3670c3f79f", "submitted_at": "2025-01-21T17:48:10Z", "updated_at": "2025-01-21T17:48:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.12326v1"}, "authors": ["Yujia Qin", "Yining Ye", "Junjie Fang", "Haoming Wang", "Shihao Liang", "Shizuo Tian", "Junda Zhang", "Jiahao Li", "Yunxin Li", "Shijue Huang", "Wanjun Zhong", "Kuanye Li", "Jiale Yang", "Yu Miao", "Woyu Lin", "Longxiang Liu", "Xu Jiang", "Qianli Ma", "Jingyu Li", "Xiaojun Xiao", "Kai Cai", "Chuang Li", "Yaowei Zheng", "Chaolin Jin", "Chen Li", "Xiao Zhou", "Minchao Wang", "Haoli Chen", "Zhaojian Li", "Haihua Yang", "Haifeng Liu", "Feng Lin", "Tao Peng", "Xin Liu", "Guang Shi"], "canonical_source_url": "https://arxiv.org/abs/2501.12326", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-12326/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d19aa758e66411a31dd1fea6f07c83d1c828bd4aad86f1564a9d7f3670c3f79f", "submitted_at": "2025-01-21T17:48:10Z", "updated_at": "2025-01-21T17:48:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2501.12326v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2501.12326"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.12326"}], "name": "UI-TARS: Pioneering Automated GUI Interaction with Native Agents", "publication": null, "published_at": "2025-01-21", "record_sha256": "1210398f0cae14ede5984319d7e3a6191abdc25c9aabfd33a77986404a08de07", "relations": [], "resource_id": "arxiv:2501.12326", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.12326v1", "topics": ["grounding", "planning", "data-generation"], "updated_at": "2025-01-21"}, "text": "UI-TARS: Pioneering Automated GUI Interaction with Native Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies a screenshot-driven native GUI model with unified actions, deliberate reasoning, and iterative training on interaction traces.\nPlatforms: cross-platform\nTopics: grounding, planning, data-generation\nPublished: 2025-01-21\nAuthors: Yujia Qin, Yining Ye, Junjie Fang, Haoming Wang, Shihao Liang, Shizuo Tian, Junda Zhang, Jiahao Li, Yunxin Li, Shijue Huang, Wanjun Zhong, Kuanye Li, Jiale Yang, Yu Miao, Woyu Lin, Longxiang Liu, Xu Jiang, Qianli Ma, Jingyu Li, Xiaojun Xiao, Kai Cai, Chuang Li, Yaowei Zheng, Chaolin Jin, Chen Li, Xiao Zhou, Minchao Wang, Haoli Chen, Zhaojian Li, Haihua Yang, Haifeng Liu, Feng Lin, Tao Peng, Xin Liu, Guang Shi\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.12326v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-12326/"}
{"id": "arxiv:2501.13896", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0804055deb6c3513990ed8d596486400f1cb791338b7866d9e949186a34eb9c5", "submitted_at": "2025-01-23T18:16:21Z", "updated_at": "2025-01-27T18:58:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.13896v2"}, "authors": ["Yue Fan", "Handong Zhao", "Ruiyi Zhang", "Yu Shen", "Xin Eric Wang", "Gang Wu"], "canonical_source_url": "https://arxiv.org/abs/2501.13896", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-13896/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0804055deb6c3513990ed8d596486400f1cb791338b7866d9e949186a34eb9c5", "submitted_at": "2025-01-23T18:16:21Z", "updated_at": "2025-01-27T18:58:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2501.13896v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2501.13896"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.13896"}], "name": "GUI-Bee: Align GUI Action Grounding to Novel Environments via Autonomous Exploration", "publication": null, "published_at": "2025-01-23", "record_sha256": "f17ca5023d0aef5470ca3f74d0ffd63d9d6c8f055bf28f779417f966d354ab3c", "relations": [], "resource_id": "arxiv:2501.13896", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.13896v2", "topics": ["grounding", "data-generation"], "updated_at": "2025-01-27"}, "text": "GUI-Bee: Align GUI Action Grounding to Novel Environments via Autonomous Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nCollects environment-specific interaction data through autonomous exploration to adapt grounding models to previously unseen GUIs.\nPlatforms: cross-platform\nTopics: grounding, data-generation\nPublished: 2025-01-23\nAuthors: Yue Fan, Handong Zhao, Ruiyi Zhang, Yu Shen, Xin Eric Wang, Gang Wu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.13896v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-13896/"}
{"id": "arxiv:2501.16150", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b36fe818cb176c434b953fe8d2c97b67176b58cc6f8426b629231875de3c0896", "submitted_at": "2025-01-27T15:44:02Z", "updated_at": "2026-03-27T07:41:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.16150v3"}, "authors": ["Pascal J. Sager", "Benjamin Meyer", "Peng Yan", "Rebekka von Wartburg-Kottler", "Layan Etaiwi", "Aref Enayati", "Gabriel Nobel", "Ahmed Abdulkadir", "Benjamin F. Grewe", "Thilo Stadelmann"], "canonical_source_url": "https://arxiv.org/abs/2501.16150", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-16150/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b36fe818cb176c434b953fe8d2c97b67176b58cc6f8426b629231875de3c0896", "submitted_at": "2025-01-27T15:44:02Z", "updated_at": "2026-03-27T07:41:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2501.16150v3"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2501.16150"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.16150"}], "name": "A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions", "publication": null, "published_at": "2025-01-27", "record_sha256": "47f4f60743e58a3b652194d8b8fa67afbb725ae4e6ebae4db8fa3176cd6f3d32", "relations": [], "resource_id": "arxiv:2501.16150", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.16150v3", "topics": ["survey"], "updated_at": "2026-03-27"}, "text": "A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions\nResource type: survey\nReview status: curated\nScope: core\n\nReviews 87 computer-use agents and 33 datasets through a taxonomy of operating domain, observation and action modalities, and agent design.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2025-01-27\nAuthors: Pascal J. Sager, Benjamin Meyer, Peng Yan, Rebekka von Wartburg-Kottler, Layan Etaiwi, Aref Enayati, Gabriel Nobel, Ahmed Abdulkadir, Benjamin F. Grewe, Thilo Stadelmann\nRecorded arXiv version: v3\n\nPrimary source: https://arxiv.org/abs/2501.16150v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-16150/"}
{"id": "arxiv:2501.16609", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b14b88273749ac54ebbe3694cbe4196640da8f81ba2faf77b0e6bbd79a060f6", "submitted_at": "2025-01-28T00:56:53Z", "updated_at": "2026-02-27T05:35:06Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2501.16609v4"}, "authors": ["Faria Huq", "Zora Zhiruo Wang", "Frank F. Xu", "Tianyue Ou", "Shuyan Zhou", "Jeffrey P. Bigham", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2501.16609", "citation_url": "https://awesomecua.github.io/resources/arxiv-2501-16609/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b14b88273749ac54ebbe3694cbe4196640da8f81ba2faf77b0e6bbd79a060f6", "submitted_at": "2025-01-28T00:56:53Z", "updated_at": "2026-02-27T05:35:06Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2501.16609v4"}, "publication": {"authors": ["Faria Huq", "Zora Zhiruo Wang", "Frank F. Xu", "Tianyue Ou", "Shuyan Zhou", "Jeffrey P. Bigham", "Graham Neubig"], "date_basis": "The ACL Anthology citation_publication_date is 2025/4, and its visible Month/Year fields state April 2025. Only month precision is supplied; no day is inferred from conference dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.17", "event_dates": null, "evidence_urls": ["https://arxiv.org/abs/2501.16609v4", "https://aclanthology.org/2025.naacl-demo.17/"], "identity_basis": "The retained arXiv v4 abstract page explicitly links this DOI. The ACL publication has the same complete title and the same seven authors in the same order, and describes the same CowPilot human-agent navigation framework. This attaches the proceedings edition to the existing arXiv identity.", "issue": null, "notes": "The April 2025 proceedings edition is distinct from the retained arXiv v4, last revised on 2026-02-27. The publisher abstract has minor wording and typesetting differences from the retained arXiv abstract; neither its text nor its date replaces the original arXiv snapshot or annotation. Exact publication day, numeric volume, issue and series are not established.", "pages": "163-172", "pdf_url": "https://aclanthology.org/2025.naacl-demo.17.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.17/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:57Z", "role": "primary", "sha256": "8c39a7773a9f2b58950aa22d6924b00ed32ab5b2b42302590dcec88da255a71f", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes"], "url": "https://arxiv.org/abs/2501.16609v4"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:44:12Z", "role": "primary", "sha256": "410da7f4b7a79e56efc5f88014a078da0feb8d2ae3ef3d36a04dce5be29d4250", "supports": ["details.publication.type", "details.publication.title", "details.publication.authors", "details.publication.date_parts", "details.publication.date_basis", "details.publication.url", "details.publication.pdf_url", "details.publication.doi", "details.publication.venue", "details.publication.publisher", "details.publication.pages", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.naacl-demo.17/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:57Z", "role": "primary", "sha256": "8c39a7773a9f2b58950aa22d6924b00ed32ab5b2b42302590dcec88da255a71f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2501.16609v4"}], "identifiers": {"arxiv": "2501.16609"}, "kind": "paper", "last_verified_at": "2026-10-17T11:19:57Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2501.16609"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.17/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.17.pdf"}], "name": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation", "publication": {"authors": ["Faria Huq", "Zora Zhiruo Wang", "Frank F. Xu", "Tianyue Ou", "Shuyan Zhou", "Jeffrey P. Bigham", "Graham Neubig"], "date_basis": "The ACL Anthology citation_publication_date is 2025/4, and its visible Month/Year fields state April 2025. Only month precision is supplied; no day is inferred from conference dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.17", "event_dates": null, "evidence_urls": ["https://arxiv.org/abs/2501.16609v4", "https://aclanthology.org/2025.naacl-demo.17/"], "identity_basis": "The retained arXiv v4 abstract page explicitly links this DOI. The ACL publication has the same complete title and the same seven authors in the same order, and describes the same CowPilot human-agent navigation framework. This attaches the proceedings edition to the existing arXiv identity.", "issue": null, "notes": "The April 2025 proceedings edition is distinct from the retained arXiv v4, last revised on 2026-02-27. The publisher abstract has minor wording and typesetting differences from the retained arXiv abstract; neither its text nor its date replaces the original arXiv snapshot or annotation. Exact publication day, numeric volume, issue and series are not established.", "pages": "163-172", "pdf_url": "https://aclanthology.org/2025.naacl-demo.17.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.17/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}, "published_at": "2025-01-28", "record_sha256": "0aeaa020233a142296002926ededdde08a8428d1719caebd72a180c72d5556a6", "relations": [], "resource_id": "arxiv:2501.16609", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2501.16609v4", "topics": ["web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2026-02-27"}, "text": "CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CowPilot, a web-navigation framework where agents suggest actions and users can pause, reject, override, or resume them. It evaluates autonomous and collaborative use through task success and human effort.\nPlatforms: web\nTopics: web-navigation, human-in-the-loop, evaluation\nPublished: 2025-01-28\nAuthors: Faria Huq, Zora Zhiruo Wang, Frank F. Xu, Tianyue Ou, Shuyan Zhou, Jeffrey P. Bigham, Graham Neubig\nRecorded arXiv version: v4\n\nPublication edition: CowPilot: A Framework for Autonomous and Human-Agent Collaborative Web Navigation\nPublication authors: Faria Huq, Zora Zhiruo Wang, Frank F. Xu, Tianyue Ou, Shuyan Zhou, Jeffrey P. Bigham, Graham Neubig\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 163-172\nPublication doi: 10.18653/v1/2025.naacl-demo.17\nPublication date basis: The ACL Anthology citation_publication_date is 2025/4, and its visible Month/Year fields state April 2025. Only month precision is supplied; no day is inferred from conference dates.\nPublication notes: The April 2025 proceedings edition is distinct from the retained arXiv v4, last revised on 2026-02-27. The publisher abstract has minor wording and typesetting differences from the retained arXiv abstract; neither its text nor its date replaces the original arXiv snapshot or annotation. Exact publication day, numeric volume, issue and series are not established.\nPublication source: https://aclanthology.org/2025.naacl-demo.17/\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2501.16609v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2501-16609/"}
{"id": "arxiv:2502.01977", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b76ab1c6c44db9f86b2bfc11097c54e5c3115c0c498a58d896bef043d2113c8", "submitted_at": "2025-02-04T03:39:59Z", "updated_at": "2025-06-07T05:23:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.01977v2"}, "authors": ["Hongxin Li", "Jingfan Chen", "Jingran Su", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2502.01977", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-01977/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b76ab1c6c44db9f86b2bfc11097c54e5c3115c0c498a58d896bef043d2113c8", "submitted_at": "2025-02-04T03:39:59Z", "updated_at": "2025-06-07T05:23:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.01977v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.01977"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.01977"}], "name": "AutoGUI: Scaling GUI Grounding with Automatic Functionality Annotations from LLMs", "publication": null, "published_at": "2025-02-04", "record_sha256": "1669da65656f17179f2e744491f23ead3901bacb846c2b445b875eb6cc390052", "relations": [], "resource_id": "arxiv:2502.01977", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.01977v2", "topics": ["grounding", "data-generation"], "updated_at": "2025-06-07"}, "text": "AutoGUI: Scaling GUI Grounding with Automatic Functionality Annotations from LLMs\nResource type: paper\nReview status: curated\nScope: core\n\nInfers element functionality from UI changes before and after interaction to produce verified annotations for grounding-model training.\nPlatforms: cross-platform\nTopics: grounding, data-generation\nPublished: 2025-02-04\nAuthors: Hongxin Li, Jingfan Chen, Jingran Su, Yuntao Chen, Qing Li, Zhaoxiang Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.01977v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-01977/"}
{"id": "arxiv:2502.02955", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3e0b19f259bcdb57b0dea92578d0ee1c7c229ea11e622c56e808f17ca2568cc", "submitted_at": "2025-02-05T07:35:23Z", "updated_at": "2025-02-05T07:35:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.02955v1"}, "authors": ["Qinzhuo Wu", "Wei Liu", "Jian Luan", "Bin Wang"], "canonical_source_url": "https://arxiv.org/abs/2502.02955", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-02955/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3e0b19f259bcdb57b0dea92578d0ee1c7c229ea11e622c56e808f17ca2568cc", "submitted_at": "2025-02-05T07:35:23Z", "updated_at": "2025-02-05T07:35:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.02955v1"}, "publication": {"authors": ["Qinzhuo Wu", "Wei Liu", "Jian Luan", "Bin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.244", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.244/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "4760-4775", "pdf_url": "https://aclanthology.org/2025.naacl-long.244.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.244/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "bb33865bafca385d3d1ce446d5cd85b6c1bf41d567509ac5ea539d4cc9c64799", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-long.244/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2502.02955"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.02955"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-long.244/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-long.244.pdf"}], "name": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation", "publication": {"authors": ["Qinzhuo Wu", "Wei Liu", "Jian Luan", "Bin Wang"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-long.244", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-long.244/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "4760-4775", "pdf_url": "https://aclanthology.org/2025.naacl-long.244.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-long.244/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)", "volume": null}, "published_at": "2025-02-05", "record_sha256": "96c87e089cf709a94d4b1bc486bb895cdf8d37fe93e17528a405cf2e4c313b45", "relations": [], "resource_id": "arxiv:2502.02955", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.02955v1", "topics": ["mobile-navigation", "planning", "data-generation"], "updated_at": "2025-02-05"}, "text": "ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation\nResource type: paper\nReview status: curated\nScope: core\n\nSplits mobile tasks into reaching the required page and operating on it, using subtask data and preferred GUI trajectories for training.\nPlatforms: unspecified\nTopics: mobile-navigation, planning, data-generation\nPublished: 2025-02-05\nAuthors: Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang\nRecorded arXiv version: v1\n\nPublication edition: ReachAgent: Enhancing Mobile Agent via Page Reaching and Operation\nPublication authors: Qinzhuo Wu, Wei Liu, Jian Luan, Bin Wang\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 4760-4775\nPublication doi: 10.18653/v1/2025.naacl-long.244\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.naacl-long.244/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.02955v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-02955/"}
{"id": "arxiv:2502.02982", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecab608111104ac044bb7d28996604e9b87bdad4543c250c57bf432f883b96d5", "submitted_at": "2025-02-05T08:26:17Z", "updated_at": "2025-05-20T07:03:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.02982v2"}, "authors": ["Wenhao Wang", "Mengying Yuan", "Zijie Yu", "Guangyi Liu", "Rui Ye", "Tian Jin", "Siheng Chen", "Yanfeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2502.02982", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-02982/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecab608111104ac044bb7d28996604e9b87bdad4543c250c57bf432f883b96d5", "submitted_at": "2025-02-05T08:26:17Z", "updated_at": "2025-05-20T07:03:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.02982v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.02982"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.02982"}], "name": "MobileA3gent: Training Mobile GUI Agents Using Decentralized Self-Sourced Data from Diverse Users", "publication": null, "published_at": "2025-02-05", "record_sha256": "b7810c30981422b5197ebca75142dbe32e0a7d3a9a93706ceba48b245644d2d0", "relations": [], "resource_id": "arxiv:2502.02982", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.02982v2", "topics": ["mobile-navigation", "data-generation"], "updated_at": "2025-05-20"}, "text": "MobileA3gent: Training Mobile GUI Agents Using Decentralized Self-Sourced Data from Diverse Users\nResource type: paper\nReview status: curated\nScope: core\n\nStudies automatic annotation and federated training of mobile GUI agents from decentralized user interaction data with heterogeneous distributions.\nPlatforms: unspecified\nTopics: mobile-navigation, data-generation\nPublished: 2025-02-05\nAuthors: Wenhao Wang, Mengying Yuan, Zijie Yu, Guangyi Liu, Rui Ye, Tian Jin, Siheng Chen, Yanfeng Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.02982v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-02982/"}
{"id": "arxiv:2502.06395", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44ec9907c8d536488e59dea307aa05bd8adcadf96cb4f3ba25a355ffe4cf8ff7", "submitted_at": "2025-02-10T12:32:21Z", "updated_at": "2025-02-10T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.06395v1"}, "authors": ["Georgios Papoudakis", "Thomas Coste", "Zhihao Wu", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2502.06395", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-06395/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44ec9907c8d536488e59dea307aa05bd8adcadf96cb4f3ba25a355ffe4cf8ff7", "submitted_at": "2025-02-10T12:32:21Z", "updated_at": "2025-02-10T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.06395v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2502.06395"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.06395"}], "name": "AppVLM: A Lightweight Vision Language Model for Online App Control", "publication": null, "published_at": "2025-02-10", "record_sha256": "f8dbff0f7341467b216edccffcc41f18956767cc3040e919dc9cdcc112b22d3d", "relations": [], "resource_id": "arxiv:2502.06395", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.06395v1", "topics": ["mobile-navigation", "imitation-learning", "data-generation", "efficiency"], "updated_at": "2025-02-10"}, "text": "AppVLM: A Lightweight Vision Language Model for Online App Control\nResource type: paper\nReview status: curated\nScope: core\n\nAppVLM first learns app actions from AndroidControl and then refines its policy through data collected in AndroidWorld. The study compares offline action prediction and online task completion for a lightweight vision-language smartphone agent.\nPlatforms: android\nTopics: mobile-navigation, imitation-learning, data-generation, efficiency\nPublished: 2025-02-10\nAuthors: Georgios Papoudakis, Thomas Coste, Zhihao Wu, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.06395v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-06395/"}
{"id": "arxiv:2502.07942", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0535de4778b6a31adb64f833733ebf94acf54ba872d98afd40b3e1012fb737fd", "submitted_at": "2025-02-11T20:41:49Z", "updated_at": "2025-03-06T19:40:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07942v2"}, "authors": ["Ruichen Zhang", "Mufan Qiu", "Zhen Tan", "Mohan Zhang", "Vincent Lu", "Jie Peng", "Kaidi Xu", "Leandro Z. Agudelo", "Peter Qian", "Tianlong Chen"], "canonical_source_url": "https://arxiv.org/abs/2502.07942", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-07942/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0535de4778b6a31adb64f833733ebf94acf54ba872d98afd40b3e1012fb737fd", "submitted_at": "2025-02-11T20:41:49Z", "updated_at": "2025-03-06T19:40:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.07942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.07942"}], "name": "Symbiotic Cooperation for Web Agents: Harnessing Complementary Strengths of Large and Small LLMs", "publication": null, "published_at": "2025-02-11", "record_sha256": "8f272e28a30265bb32d0e794a34b77deefce01f04b5852684635864e53b13207", "relations": [], "resource_id": "arxiv:2502.07942", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.07942v2", "topics": ["data-generation", "web-navigation", "efficiency"], "updated_at": "2025-03-06"}, "text": "Symbiotic Cooperation for Web Agents: Harnessing Complementary Strengths of Large and Small LLMs\nResource type: paper\nReview status: curated\nScope: core\n\nCouples large- and small-model web-agent exploration with trajectory synthesis and distillation to improve both model types iteratively.\nPlatforms: web\nTopics: data-generation, web-navigation, efficiency\nPublished: 2025-02-11\nAuthors: Ruichen Zhang, Mufan Qiu, Zhen Tan, Mohan Zhang, Vincent Lu, Jie Peng, Kaidi Xu, Leandro Z. Agudelo, Peter Qian, Tianlong Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.07942v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-07942/"}
{"id": "arxiv:2502.07949", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "186b7fd8e79db290ea481645d8adab3989a12f787a13ecb11270f4756f5af42a", "submitted_at": "2025-02-11T20:57:46Z", "updated_at": "2025-05-20T19:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07949v2"}, "authors": ["Qingyuan Wu", "Jianheng Liu", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2502.07949", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-07949/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "186b7fd8e79db290ea481645d8adab3989a12f787a13ecb11270f4756f5af42a", "submitted_at": "2025-02-11T20:57:46Z", "updated_at": "2025-05-20T19:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.07949v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2502.07949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.07949"}], "name": "Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning", "publication": null, "published_at": "2025-02-11", "record_sha256": "7cd25a2e8b12d6cf15533b9d3c5a50426f307ddf37c9722d403bf549aa64c343", "relations": [], "resource_id": "arxiv:2502.07949", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.07949v2", "topics": ["reinforcement-learning", "planning"], "updated_at": "2025-05-20"}, "text": "Advancing Autonomous VLM Agents via Variational Subgoal-Conditioned Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUses variational subgoal-conditioned reinforcement learning to improve learning efficiency on sparse-reward mobile and web-control tasks.\nPlatforms: web, cross-platform\nTopics: reinforcement-learning, planning\nPublished: 2025-02-11\nAuthors: Qingyuan Wu, Jianheng Liu, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.07949v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-07949/"}
{"id": "arxiv:2502.08047", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68e6ecfcdff0163debe96f6440fb33529356bf9fd2c8743b5953ae52ae71c9ef", "submitted_at": "2025-02-12T01:06:10Z", "updated_at": "2026-05-23T20:19:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2502.08047v5"}, "authors": ["Henry Hengyuan Zhao", "Kaiming Yang", "Wendi Yu", "Difei Gao", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2502.08047", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-08047/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68e6ecfcdff0163debe96f6440fb33529356bf9fd2c8743b5953ae52ae71c9ef", "submitted_at": "2025-02-12T01:06:10Z", "updated_at": "2026-05-23T20:19:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2502.08047v5"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2502.08047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.08047"}], "name": "WorldGUI: An Interactive Benchmark for Desktop GUI Automation from Any Starting Point", "publication": null, "published_at": "2025-02-12", "record_sha256": "ee819c71fdc89eca5d54eaf8ccbbeae6245064f4a223d65f6865d9a235db74a8", "relations": [], "resource_id": "arxiv:2502.08047", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.08047v5", "topics": ["evaluation", "robustness", "planning"], "updated_at": "2026-05-23"}, "text": "WorldGUI: An Interactive Benchmark for Desktop GUI Automation from Any Starting Point\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces desktop and web tasks with varied initial states to test planning recovery and adaptability beyond canonical starting configurations.\nPlatforms: desktop, web\nTopics: evaluation, robustness, planning\nPublished: 2025-02-12\nAuthors: Henry Hengyuan Zhao, Kaiming Yang, Wendi Yu, Difei Gao, Mike Zheng Shou\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.08047v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-08047/"}
{"id": "arxiv:2502.08226", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a60d581e4318db66495a0333b864757ec896576c720e40da4a881f8740b9d9cf", "submitted_at": "2025-02-12T09:12:30Z", "updated_at": "2025-02-14T06:23:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.08226v2"}, "authors": ["Kunal Singh", "Shreyas Singh", "Mukund Khanna"], "canonical_source_url": "https://arxiv.org/abs/2502.08226", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-08226/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a60d581e4318db66495a0333b864757ec896576c720e40da4a881f8740b9d9cf", "submitted_at": "2025-02-12T09:12:30Z", "updated_at": "2025-02-14T06:23:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.08226v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.08226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.08226"}], "name": "TRISHUL: Towards Region Identification and Screen Hierarchy Understanding for Large VLM based GUI Agents", "publication": null, "published_at": "2025-02-12", "record_sha256": "ce9da8ae0c4ad9e949231880cc65df5fed064fa0070b181e7b047eb6c149f601", "relations": [], "resource_id": "arxiv:2502.08226", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.08226v2", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-02-14"}, "text": "TRISHUL: Towards Region Identification and Screen Hierarchy Understanding for Large VLM based GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines hierarchical screen parsing with spatially enriched element descriptions to support both locating GUI targets and describing indicated elements.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\nPublished: 2025-02-12\nAuthors: Kunal Singh, Shreyas Singh, Mukund Khanna\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.08226v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-08226/"}
{"id": "arxiv:2502.11357", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9f459a02b7a464d9e1e49579943d1b93d46308c1b00e795c903d2fa1502a2b2", "submitted_at": "2025-02-17T02:13:48Z", "updated_at": "2025-05-30T02:30:57Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2502.11357v4"}, "authors": ["Vardaan Pahuja", "Yadong Lu", "Corby Rosset", "Boyu Gou", "Arindam Mitra", "Spencer Whitehead", "Yu Su", "Ahmed Awadallah"], "canonical_source_url": "https://arxiv.org/abs/2502.11357", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-11357/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d9f459a02b7a464d9e1e49579943d1b93d46308c1b00e795c903d2fa1502a2b2", "submitted_at": "2025-02-17T02:13:48Z", "updated_at": "2025-05-30T02:30:57Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2502.11357v4"}, "publication": {"authors": ["Vardaan Pahuja", "Yadong Lu", "Corby Rosset", "Boyu Gou", "Arindam Mitra", "Spencer Whitehead", "Yu Su", "Ahmed Awadallah"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.326", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.326/", "https://aclanthology.org/2025.findings-acl.326.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. Landing-page citation metadata gives Ahmed Hassan for the final author, while the supplied BibTeX expands the name to Ahmed Hassan Awadallah; the PDF uses Ahmed Awadallah.", "pages": "6300-6323", "pdf_url": "https://aclanthology.org/2025.findings-acl.326.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.326/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "8ec40afe322b7c72c9f6d6db617fa2d4fae559cf6cd4e2fcc0ff5a9cc9485e05", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.326/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:14:02.000475+00:00", "role": "primary", "sha256": "c1b232e161839514713789c408ae64f66684b49770033c7f3b04bf1d1a232157", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes"], "url": "https://aclanthology.org/2025.findings-acl.326.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.11357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.11357"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.326/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.326.pdf"}], "name": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents", "publication": {"authors": ["Vardaan Pahuja", "Yadong Lu", "Corby Rosset", "Boyu Gou", "Arindam Mitra", "Spencer Whitehead", "Yu Su", "Ahmed Awadallah"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.326", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.326/", "https://aclanthology.org/2025.findings-acl.326.pdf"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity. The normalized publisher abstract matches the retained arXiv abstract SHA-256. The publisher PDF title and complete first-page author list were visually checked.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. Landing-page citation metadata gives Ahmed Hassan for the final author, while the supplied BibTeX expands the name to Ahmed Hassan Awadallah; the PDF uses Ahmed Awadallah.", "pages": "6300-6323", "pdf_url": "https://aclanthology.org/2025.findings-acl.326.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.326/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2025-02-17", "record_sha256": "19ff24dca7a38f27055f62ab4db75d601905014b4364da643855f16f8f3af82c", "relations": [], "resource_id": "arxiv:2502.11357", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.11357v4", "topics": ["data-generation", "web-navigation"], "updated_at": "2025-05-30"}, "text": "Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSynthesizes multimodal web trajectories through exploration and refinement, then trains a browser agent to study the effect of scaling interaction data.\nPlatforms: web\nTopics: data-generation, web-navigation\nPublished: 2025-02-17\nAuthors: Vardaan Pahuja, Yadong Lu, Corby Rosset, Boyu Gou, Arindam Mitra, Spencer Whitehead, Yu Su, Ahmed Awadallah\nRecorded arXiv version: v4\n\nPublication edition: Explorer: Scaling Exploration-driven Web Trajectory Synthesis for Multimodal Web Agents\nPublication authors: Vardaan Pahuja, Yadong Lu, Corby Rosset, Boyu Gou, Arindam Mitra, Spencer Whitehead, Yu Su, Ahmed Awadallah\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 6300-6323\nPublication doi: 10.18653/v1/2025.findings-acl.326\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint. Authors follow the publisher PDF, page 1. Landing-page citation metadata gives Ahmed Hassan for the final author, while the supplied BibTeX expands the name to Ahmed Hassan Awadallah; the PDF uses Ahmed Awadallah.\nPublication source: https://aclanthology.org/2025.findings-acl.326/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.11357v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-11357/"}
{"id": "arxiv:2502.13053", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5af8d3a8b4db150bab0630f6040a0f3bced612f68b11d31985053f1f25c0f6f6", "submitted_at": "2025-02-18T17:01:28Z", "updated_at": "2025-08-06T07:04:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.13053v3"}, "authors": ["Yurun Chen", "Xavier Hu", "Keting Yin", "Juncheng Li", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2502.13053", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-13053/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5af8d3a8b4db150bab0630f6040a0f3bced612f68b11d31985053f1f25c0f6f6", "submitted_at": "2025-02-18T17:01:28Z", "updated_at": "2025-08-06T07:04:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.13053v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2502.13053"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.13053"}], "name": "Evaluating the Robustness of Multimodal Agents Against Active Environmental Injection Attacks", "publication": null, "published_at": "2025-02-18", "record_sha256": "0e4cd3e7f64d1d6e59dc7ec6880a2b8b02423b0fd24e8f0edcf9aa1ae614d453", "relations": [], "resource_id": "arxiv:2502.13053", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.13053v3", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2025-08-06"}, "text": "Evaluating the Robustness of Multimodal Agents Against Active Environmental Injection Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nExamines active environmental injection against mobile agents, testing whether adversarial interface content redirects decisions during task execution.\nPlatforms: android\nTopics: safety-security, robustness, evaluation\nPublished: 2025-02-18\nAuthors: Yurun Chen, Xavier Hu, Keting Yin, Juncheng Li, Shengyu Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.13053v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-13053/"}
{"id": "arxiv:2502.14282", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75b417f4248397b73934f8b8e45199006e4db6e2ee69b62e0846621411b2dd7", "submitted_at": "2025-02-20T05:41:55Z", "updated_at": "2025-02-21T02:54:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.14282v2"}, "authors": ["Haowei Liu", "Xi Zhang", "Haiyang Xu", "Yuyang Wanyan", "Junyang Wang", "Ming Yan", "Ji Zhang", "Chunfeng Yuan", "Changsheng Xu", "Weiming Hu", "Fei Huang"], "canonical_source_url": "https://arxiv.org/abs/2502.14282", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-14282/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75b417f4248397b73934f8b8e45199006e4db6e2ee69b62e0846621411b2dd7", "submitted_at": "2025-02-20T05:41:55Z", "updated_at": "2025-02-21T02:54:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.14282v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2502.14282"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.14282"}], "name": "PC-Agent: A Hierarchical Multi-Agent Collaboration Framework for Complex Task Automation on PC", "publication": null, "published_at": "2025-02-20", "record_sha256": "8b6303558cc05421efba4779a138b9a251b3bd0d037442ab4f35c7d1d590ebbb", "relations": [], "resource_id": "arxiv:2502.14282", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.14282v2", "topics": ["planning", "cross-app"], "updated_at": "2025-02-21"}, "text": "PC-Agent: A Hierarchical Multi-Agent Collaboration Framework for Complex Task Automation on PC\nResource type: paper\nReview status: curated\nScope: core\n\nUses active perception and hierarchical agents for instruction decomposition, progress tracking, action selection, and reflection in PC workflows.\nPlatforms: desktop\nTopics: planning, cross-app\nPublished: 2025-02-20\nAuthors: Haowei Liu, Xi Zhang, Haiyang Xu, Yuyang Wanyan, Junyang Wang, Ming Yan, Ji Zhang, Chunfeng Yuan, Changsheng Xu, Weiming Hu, Fei Huang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.14282v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-14282/"}
{"id": "arxiv:2502.15760", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99492dab54d9816104d47fca525277b033cadb7a53727dae06ed1a3221fcb416", "submitted_at": "2025-02-13T18:55:14Z", "updated_at": "2025-02-13T18:55:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.15760v1"}, "authors": ["Hao Bai", "Yifei Zhou", "Li Erran Li", "Sergey Levine", "Aviral Kumar"], "canonical_source_url": "https://arxiv.org/abs/2502.15760", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-15760/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99492dab54d9816104d47fca525277b033cadb7a53727dae06ed1a3221fcb416", "submitted_at": "2025-02-13T18:55:14Z", "updated_at": "2025-02-13T18:55:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.15760v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2502.15760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.15760"}], "name": "Digi-Q: Learning Q-Value Functions for Training Device-Control Agents", "publication": null, "published_at": "2025-02-13", "record_sha256": "be71d6c3ecc167b0d01bef57bec445659ebe8285fd35ebf6f35749d6e41d9c52", "relations": [], "resource_id": "arxiv:2502.15760", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.15760v1", "topics": ["reinforcement-learning", "efficiency"], "updated_at": "2025-02-13"}, "text": "Digi-Q: Learning Q-Value Functions for Training Device-Control Agents\nResource type: paper\nReview status: curated\nScope: core\n\nLearns action-value functions from offline device-control experience and uses them to rank candidate actions for policy improvement.\nPlatforms: android\nTopics: reinforcement-learning, efficiency\nPublished: 2025-02-13\nAuthors: Hao Bai, Yifei Zhou, Li Erran Li, Sergey Levine, Aviral Kumar\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.15760v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-15760/"}
{"id": "arxiv:2502.17110", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d723cb764deb78b194ff0199c7f1b921a12b6a41655c07b06ae88c3d0af58358", "submitted_at": "2025-02-24T12:51:23Z", "updated_at": "2025-06-03T08:37:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.17110v3"}, "authors": ["Junyang Wang", "Haiyang Xu", "Xi Zhang", "Ming Yan", "Ji Zhang", "Fei Huang", "Jitao Sang"], "canonical_source_url": "https://arxiv.org/abs/2502.17110", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-17110/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d723cb764deb78b194ff0199c7f1b921a12b6a41655c07b06ae88c3d0af58358", "submitted_at": "2025-02-24T12:51:23Z", "updated_at": "2025-06-03T08:37:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.17110v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2502.17110"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.17110"}], "name": "Mobile-Agent-V: A Video-Guided Approach for Effortless and Efficient Operational Knowledge Injection in Mobile Automation", "publication": null, "published_at": "2025-02-24", "record_sha256": "bbf1b06ffae37600e6934cafea369d6052acdcc26ea02b4f58d6f2c290f96faf", "relations": [], "resource_id": "arxiv:2502.17110", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.17110v3", "topics": ["mobile-navigation", "video", "memory"], "updated_at": "2025-06-03"}, "text": "Mobile-Agent-V: A Video-Guided Approach for Effortless and Efficient Operational Knowledge Injection in Mobile Automation\nResource type: paper\nReview status: curated\nScope: core\n\nExtracts operational guidance from video for mobile task execution and introduces Mobile-Knowledge to evaluate the contribution of external knowledge.\nPlatforms: unspecified\nTopics: mobile-navigation, video, memory\nPublished: 2025-02-24\nAuthors: Junyang Wang, Haiyang Xu, Xi Zhang, Ming Yan, Ji Zhang, Fei Huang, Jitao Sang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The authors identify arXiv:2505.13887 as an accidental duplicate and direct future updates to this record.\n\nPrimary source: https://arxiv.org/abs/2502.17110v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-17110/"}
{"id": "arxiv:2502.17903", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3903e97e6db6bd470699f0d7b2785e53e15291234f1e5f717bae8d2ccd851ca4", "submitted_at": "2025-02-25T06:58:40Z", "updated_at": "2025-02-25T06:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.17903v1"}, "authors": ["Lars Krupp", "Daniel Geißler", "Paul Lukowicz", "Jakob Karolus"], "canonical_source_url": "https://arxiv.org/abs/2502.17903", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-17903/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3903e97e6db6bd470699f0d7b2785e53e15291234f1e5f717bae8d2ccd851ca4", "submitted_at": "2025-02-25T06:58:40Z", "updated_at": "2025-02-25T06:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.17903v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.17903"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.17903"}], "name": "Towards Sustainable Web Agents: A Plea for Transparency and Dedicated Metrics for Energy Consumption", "publication": null, "published_at": "2025-02-25", "record_sha256": "e2b5f118989553ba7bf3d7735bc6803cc224ae0cda66ef36bbaa812bc6d382af", "relations": [], "resource_id": "arxiv:2502.17903", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.17903v1", "topics": ["evaluation", "efficiency"], "updated_at": "2025-02-25"}, "text": "Towards Sustainable Web Agents: A Plea for Transparency and Dedicated Metrics for Energy Consumption\nResource type: paper\nReview status: curated\nScope: core\n\nExamines energy and carbon-accounting limitations in web-agent evaluation and argues for transparent, dedicated sustainability measurements.\nPlatforms: web\nTopics: evaluation, efficiency\nPublished: 2025-02-25\nAuthors: Lars Krupp, Daniel Geißler, Paul Lukowicz, Jakob Karolus\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.17903v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-17903/"}
{"id": "arxiv:2502.18525", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e4091fc740249b6406dd9e96f2692d7fac9323759ec6fad6ba5a05f3e445844", "submitted_at": "2025-02-24T18:41:33Z", "updated_at": "2025-10-03T02:09:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.18525v2"}, "authors": ["Pranjal Aggarwal", "Sean Welleck"], "canonical_source_url": "https://arxiv.org/abs/2502.18525", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-18525/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e4091fc740249b6406dd9e96f2692d7fac9323759ec6fad6ba5a05f3e445844", "submitted_at": "2025-02-24T18:41:33Z", "updated_at": "2025-10-03T02:09:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2502.18525v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.18525"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.18525"}], "name": "Programming with Pixels: Can Computer-Use Agents do Software Engineering?", "publication": null, "published_at": "2025-02-24", "record_sha256": "96903e55d127010856fdfaa271fbc7a74685c9662aa45a44ef3c5a013e7e372b", "relations": [], "resource_id": "arxiv:2502.18525", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.18525v2", "topics": ["evaluation", "tool-integration", "automation"], "updated_at": "2025-10-03"}, "text": "Programming with Pixels: Can Computer-Use Agents do Software Engineering?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a visual IDE environment and benchmark for software-engineering tasks, comparing screenshot-based control with access to editing and shell tools.\nPlatforms: desktop\nTopics: evaluation, tool-integration, automation\nPublished: 2025-02-24\nAuthors: Pranjal Aggarwal, Sean Welleck\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.18525v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-18525/"}
{"id": "arxiv:2502.18906", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d69b006a298a591aacfe7f29df63f508113c0ec824ede82a0bb5bbb9cb66266", "submitted_at": "2025-02-26T07:52:02Z", "updated_at": "2025-02-26T07:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.18906v1"}, "authors": ["Jiani Zheng", "Lu Wang", "Fangkai Yang", "Chaoyun Zhang", "Lingrui Mei", "Wenjie Yin", "Qingwei Lin", "Dongmei Zhang", "Saravan Rajmohan", "Qi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2502.18906", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-18906/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d69b006a298a591aacfe7f29df63f508113c0ec824ede82a0bb5bbb9cb66266", "submitted_at": "2025-02-26T07:52:02Z", "updated_at": "2025-02-26T07:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2502.18906v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2502.18906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.18906"}], "name": "VEM: Environment-Free Exploration for Training GUI Agent with Value Environment Model", "publication": null, "published_at": "2025-02-26", "record_sha256": "59c9ee1adbfdda24551cd397b65c54e1c6bffe32c147208725453827d404264a", "relations": [], "resource_id": "arxiv:2502.18906", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.18906v1", "topics": ["reinforcement-learning", "world-models"], "updated_at": "2025-02-26"}, "text": "VEM: Environment-Free Exploration for Training GUI Agent with Value Environment Model\nResource type: paper\nReview status: curated\nScope: core\n\nTrains GUI agents with a frozen value model learned from offline interactions, guiding action exploration without predicting complete next observations.\nPlatforms: android\nTopics: reinforcement-learning, world-models\nPublished: 2025-02-26\nAuthors: Jiani Zheng, Lu Wang, Fangkai Yang, Chaoyun Zhang, Lingrui Mei, Wenjie Yin, Qingwei Lin, Dongmei Zhang, Saravan Rajmohan, Qi Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.18906v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-18906/"}
{"id": "arxiv:2502.20383", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27eacc2668411e2b7f79423940aa12293fa57aa81390c64587d3074c141dcb4f", "submitted_at": "2025-02-27T18:56:26Z", "updated_at": "2025-09-21T22:29:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.20383v3"}, "authors": ["Jeffrey Yang Fan Chiang", "Seungjae Lee", "Jia-Bin Huang", "Furong Huang", "Yizheng Chen"], "canonical_source_url": "https://arxiv.org/abs/2502.20383", "citation_url": "https://awesomecua.github.io/resources/arxiv-2502-20383/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27eacc2668411e2b7f79423940aa12293fa57aa81390c64587d3074c141dcb4f", "submitted_at": "2025-02-27T18:56:26Z", "updated_at": "2025-09-21T22:29:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2502.20383v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:44Z", "role": "primary", "sha256": "9ca057ee6538002c739e58a7b3f86fc51a434d085a33e5590f0720e4f27c6b24", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2501.06497%2C2501.06994%2C2501.07834%2C2501.07927%2C2501.11233%2C2501.15570%2C2501.16403%2C2501.16609%2C2501.17391%2C2502.00372%2C2502.00392%2C2502.00964%2C2502.00988%2C2502.01562%2C2502.01839%2C2502.03723%2C2502.05174%2C2502.05227%2C2502.08489%2C2502.11221%2C2502.12130%2C2502.14496%2C2502.16587%2C2502.16786%2C2502.16796%2C2502.18356%2C2502.18512%2C2502.20383%2C2502.20390%2C2502.20582%2C2503.06378%2C2503.08026%2C2503.10613%2C2503.10809%2C2503.11579%2C2503.12228%2C2503.12461%2C2503.13843%2C2503.14495%2C2503.15426%2C2503.16036%2C2503.16248%2C2503.16975%2C2503.17671%2C2503.18102%2C2503.21889%2C2503.21949%2C2503.23803%2C2504.01328%2C2504.01377&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:04Z", "role": "primary", "sha256": "0118540c95d316ed30aabf783c072d413677aaf89539316e08e1d81ddab2c5e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2502.20383v3"}], "identifiers": {"arxiv": "2502.20383"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2502.20383"}], "name": "Why Are Web AI Agents More Vulnerable Than Standalone LLMs? A Security Analysis", "publication": null, "published_at": "2025-02-27", "record_sha256": "1066da7e9849c99a7c876262348951a1b4aa0ab95f75525876e8397815c42bc3", "relations": [], "resource_id": "arxiv:2502.20383", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2502.20383v3", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-09-21"}, "text": "Why Are Web AI Agents More Vulnerable Than Standalone LLMs? A Security Analysis\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes web-agent vulnerabilities at the component level, examining goal placement in prompts, multistep action generation, and observations as sources of increased susceptibility.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2025-02-27\nAuthors: Jeffrey Yang Fan Chiang, Seungjae Lee, Jia-Bin Huang, Furong Huang, Yizheng Chen\nRecorded arXiv version: v3\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2502.20383v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2502-20383/"}
{"id": "arxiv:2503.00401", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53781edbb107437b258d0c0e298f80ec14f8f32b51bc0dd11c0417473eb2c606", "submitted_at": "2025-03-01T08:29:59Z", "updated_at": "2025-03-04T12:04:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.00401v2"}, "authors": ["Zongru Wu", "Pengzhou Cheng", "Zheng Wu", "Tianjie Ju", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2503.00401", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-00401/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53781edbb107437b258d0c0e298f80ec14f8f32b51bc0dd11c0417473eb2c606", "submitted_at": "2025-03-01T08:29:59Z", "updated_at": "2025-03-04T12:04:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.00401v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.00401"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.00401"}], "name": "Smoothing Grounding and Reasoning for MLLM-Powered GUI Agents with Query-Oriented Pivot Tasks", "publication": null, "published_at": "2025-03-01", "record_sha256": "5ebb108f107860fffbe73189eb66dd0a04693993ff816a7e3ec62e855b62345b", "relations": [], "resource_id": "arxiv:2503.00401", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.00401v2", "topics": ["grounding", "planning"], "updated_at": "2025-03-04"}, "text": "Smoothing Grounding and Reasoning for MLLM-Powered GUI Agents with Query-Oriented Pivot Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nUses inference of possible user queries from screenshots and target coordinates as a training bridge between GUI grounding and action reasoning.\nPlatforms: cross-platform\nTopics: grounding, planning\nPublished: 2025-03-01\nAuthors: Zongru Wu, Pengzhou Cheng, Zheng Wu, Tianjie Ju, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.00401v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-00401/"}
{"id": "arxiv:2503.01861", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2248d423c954ee6e406e3f46f6da43fabd6e832ffacc05cfa32083f2ff0c2700", "submitted_at": "2025-02-24T09:31:56Z", "updated_at": "2025-07-09T13:52:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.01861v3"}, "authors": ["Sami Marreed", "Alon Oved", "Avi Yaeli", "Segev Shlomov", "Ido Levy", "Offer Akrabi", "Aviad Sela", "Asaf Adi", "Nir Mashkif"], "canonical_source_url": "https://arxiv.org/abs/2503.01861", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-01861/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2248d423c954ee6e406e3f46f6da43fabd6e832ffacc05cfa32083f2ff0c2700", "submitted_at": "2025-02-24T09:31:56Z", "updated_at": "2025-07-09T13:52:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.01861v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.01861"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.01861"}], "name": "Towards Enterprise-Ready Computer Using Generalist Agent", "publication": null, "published_at": "2025-02-24", "record_sha256": "80b6415f022fd85d643ee2650c6aa9bb7ccaf6a123252e417efccfa77ba2b6c2", "relations": [], "resource_id": "arxiv:2503.01861", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.01861v3", "topics": ["automation", "evaluation", "workflow"], "updated_at": "2025-07-09"}, "text": "Towards Enterprise-Ready Computer Using Generalist Agent\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes the CUGA enterprise-agent development process, emphasizing iterative evaluation, failure analysis, and refinement on software-task benchmarks.\nPlatforms: web\nTopics: automation, evaluation, workflow\nPublished: 2025-02-24\nAuthors: Sami Marreed, Alon Oved, Avi Yaeli, Segev Shlomov, Ido Levy, Offer Akrabi, Aviad Sela, Asaf Adi, Nir Mashkif\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.01861v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-01861/"}
{"id": "arxiv:2503.02197", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a58e33cd2b60a8c734968dd92f031b9aa4811932b1a666d947b20ec4c6035a7", "submitted_at": "2025-03-04T02:14:55Z", "updated_at": "2025-06-05T01:42:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02197v2"}, "authors": ["Zhixun Chen", "Ming Li", "Yuxuan Huang", "Yali Du", "Meng Fang", "Tianyi Zhou"], "canonical_source_url": "https://arxiv.org/abs/2503.02197", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-02197/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a58e33cd2b60a8c734968dd92f031b9aa4811932b1a666d947b20ec4c6035a7", "submitted_at": "2025-03-04T02:14:55Z", "updated_at": "2025-06-05T01:42:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02197v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "9048f54fbe64ee62117e37e3c9ad5b157fc70e83304d9ccb782f8782e487dd4a", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2503.02197"}], "identifiers": {"arxiv": "2503.02197"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02197"}], "name": "ATLaS: Agent Tuning via Learning Critical Steps", "publication": null, "published_at": "2025-03-04", "record_sha256": "460063a1595768af7f1c8f311e2b1d5bcf9ff0b61292d572cd52719c3ecb6d04", "relations": [], "resource_id": "arxiv:2503.02197", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2503.02197v2", "topics": ["data-generation", "planning", "efficiency"], "updated_at": "2025-06-05"}, "text": "ATLaS: Agent Tuning via Learning Critical Steps\nResource type: paper\nReview status: curated\nScope: adjacent\n\nSelects critical steps from expert interaction trajectories for targeted agent fine-tuning, studying efficiency and generalization across tasks.\nPlatforms: web\nTopics: data-generation, planning, efficiency\nPublished: 2025-03-04\nAuthors: Zhixun Chen, Ming Li, Yuxuan Huang, Yali Du, Meng Fang, Tianyi Zhou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.\n\nPrimary source: https://arxiv.org/abs/2503.02197v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-02197/"}
{"id": "arxiv:2503.02268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d396e4e249fc2aa1f4d53eadac38c3c814aae3f0208921b3388d5b009e6037b3", "submitted_at": "2025-03-04T04:34:09Z", "updated_at": "2025-04-15T02:32:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.02268v3"}, "authors": ["Wenjia Jiang", "Yangyang Zhuang", "Chenxi Song", "Xu Yang", "Joey Tianyi Zhou", "Chi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2503.02268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-02268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d396e4e249fc2aa1f4d53eadac38c3c814aae3f0208921b3388d5b009e6037b3", "submitted_at": "2025-03-04T04:34:09Z", "updated_at": "2025-04-15T02:32:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.02268v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.02268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02268"}], "name": "AppAgentX: Evolving GUI Agents as Proficient Smartphone Users", "publication": null, "published_at": "2025-03-04", "record_sha256": "800880517e522df440bb37ecbb951c417322b5722fb5516ad090837e06f4e9a2", "relations": [], "resource_id": "arxiv:2503.02268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.02268v3", "topics": ["memory", "efficiency"], "updated_at": "2025-04-15"}, "text": "AppAgentX: Evolving GUI Agents as Proficient Smartphone Users\nResource type: paper\nReview status: curated\nScope: core\n\nRecords interaction history and converts repeated action sequences into reusable high-level shortcuts for mobile GUI tasks.\nPlatforms: android\nTopics: memory, efficiency\nPublished: 2025-03-04\nAuthors: Wenjia Jiang, Yangyang Zhuang, Chenxi Song, Xu Yang, Joey Tianyi Zhou, Chi Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.02268v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-02268/"}
{"id": "arxiv:2503.02403", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0c393d109ab0f8da2bf25bb39581236babd50f7554f2ec99679a99fb7215171", "submitted_at": "2025-03-04T08:44:30Z", "updated_at": "2025-09-24T05:11:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02403v2"}, "authors": ["Jiahui Sun", "Zhichao Hua", "Yubin Xia"], "canonical_source_url": "https://arxiv.org/abs/2503.02403", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-02403/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0c393d109ab0f8da2bf25bb39581236babd50f7554f2ec99679a99fb7215171", "submitted_at": "2025-03-04T08:44:30Z", "updated_at": "2025-09-24T05:11:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02403v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2503.02403"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02403"}], "name": "AutoEval: A Practical Framework for Autonomous Evaluation of Mobile Agents", "publication": null, "published_at": "2025-03-04", "record_sha256": "ff05f6fc25947a7087db8c08b1ff1dec74e141bcaadb23d934eadd1115648362", "relations": [], "resource_id": "arxiv:2503.02403", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.02403v2", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-09-24"}, "text": "AutoEval: A Practical Framework for Autonomous Evaluation of Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGenerates reward signals from UI-state changes and uses a judging system to evaluate mobile agents with reduced manual evaluator construction.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation\nPublished: 2025-03-04\nAuthors: Jiahui Sun, Zhichao Hua, Yubin Xia\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.02403v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-02403/"}
{"id": "arxiv:2503.02682", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3395107d8c7db3562bf75a3c22fee55d3da0564bd2f413491bc0d381fc601323", "submitted_at": "2025-03-04T14:54:45Z", "updated_at": "2025-09-10T16:45:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02682v2"}, "authors": ["Weimin Xiong", "Yifan Song", "Qingxiu Dong", "Bingchan Zhao", "Feifan Song", "Xun Wang", "Sujian Li"], "canonical_source_url": "https://arxiv.org/abs/2503.02682", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-02682/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3395107d8c7db3562bf75a3c22fee55d3da0564bd2f413491bc0d381fc601323", "submitted_at": "2025-03-04T14:54:45Z", "updated_at": "2025-09-10T16:45:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02682v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:06:43Z", "role": "primary", "sha256": "e634a3e576c94f1f7f5409d28a5c4c80c46141f1763c81337bb45f19793815f6", "supports": ["environments", "scope", "notes"], "url": "https://arxiv.org/html/2503.02682"}], "identifiers": {"arxiv": "2503.02682"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02682"}], "name": "MPO: Boosting LLM Agents with Meta Plan Optimization", "publication": null, "published_at": "2025-03-04", "record_sha256": "63703ae4781d1e59e331994bd5536b6e7e5e2438f5133546c2eac8ffb44b1fba", "relations": [], "resource_id": "arxiv:2503.02682", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2503.02682v2", "topics": ["planning", "memory"], "updated_at": "2025-09-10"}, "text": "MPO: Boosting LLM Agents with Meta Plan Optimization\nResource type: paper\nReview status: curated\nScope: adjacent\n\nRefines reusable high-level meta plans with execution feedback to guide language-agent planning across interactive environments.\nPlatforms: web\nTopics: planning, memory\nPublished: 2025-03-04\nAuthors: Weimin Xiong, Yifan Song, Qingxiu Dong, Bingchan Zhao, Feifan Song, Xun Wang, Sujian Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The full paper was also inspected to confirm WebShop evaluation; the broader agent-learning method is classified as adjacent.\n\nPrimary source: https://arxiv.org/abs/2503.02682v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-02682/"}
{"id": "arxiv:2503.02950", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df682f883bed5e0e9917074dfa130839fabaa318c55714f26943d944774854fe", "submitted_at": "2025-03-04T19:13:10Z", "updated_at": "2025-05-06T06:42:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02950v2"}, "authors": ["Danqing Zhang", "Balaji Rama", "Jingyi Ni", "Shiying He", "Fu Zhao", "Kunyu Chen", "Arnold Chen", "Junyu Cao"], "canonical_source_url": "https://arxiv.org/abs/2503.02950", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-02950/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df682f883bed5e0e9917074dfa130839fabaa318c55714f26943d944774854fe", "submitted_at": "2025-03-04T19:13:10Z", "updated_at": "2025-05-06T06:42:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.02950v2"}, "publication": {"authors": ["Danqing Zhang", "Balaji Rama", "Jingyi Ni", "Shiying He", "Fu Zhao", "Kunyu Chen", "Arnold Chen", "Junyu Cao"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.36", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.36/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "449-455", "pdf_url": "https://aclanthology.org/2025.naacl-demo.36.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.36/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:58Z", "role": "primary", "sha256": "9b130228c049f324632f571dd9c9e74c9a293ee240351c7a6fd4fef3359afc12", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.naacl-demo.36/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.02950"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.02950"}, {"rel": "publication", "url": "https://aclanthology.org/2025.naacl-demo.36/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.naacl-demo.36.pdf"}], "name": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications", "publication": {"authors": ["Danqing Zhang", "Balaji Rama", "Jingyi Ni", "Shiying He", "Fu Zhao", "Kunyu Chen", "Arnold Chen", "Junyu Cao"], "date_basis": "Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 4], "doi": "10.18653/v1/2025.naacl-demo.36", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.naacl-demo.36/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "449-455", "pdf_url": "https://aclanthology.org/2025.naacl-demo.36.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications", "type": "conference-paper", "url": "https://aclanthology.org/2025.naacl-demo.36/", "venue": "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)", "volume": null}, "published_at": "2025-03-04", "record_sha256": "9e93cbae059137d18830c9cc720bb1a5ddf4a5d2c8507142301a29359db90089", "relations": [], "resource_id": "arxiv:2503.02950", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.02950v2", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2025-05-06"}, "text": "LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a modular visual web-agent stack with planning and memory components, exposed through a remote-browser application and Chrome extension.\nPlatforms: web\nTopics: web-navigation, tool-integration, planning\nPublished: 2025-03-04\nAuthors: Danqing Zhang, Balaji Rama, Jingyi Ni, Shiying He, Fu Zhao, Kunyu Chen, Arnold Chen, Junyu Cao\nRecorded arXiv version: v2\n\nPublication edition: LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications\nPublication authors: Danqing Zhang, Balaji Rama, Jingyi Ni, Shiying He, Fu Zhao, Kunyu Chen, Arnold Chen, Junyu Cao\nPublication edition date (recorded precision): 2025-04\nPublication venue: Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (System Demonstrations)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 449-455\nPublication doi: 10.18653/v1/2025.naacl-demo.36\nPublication date basis: Explicit publisher citation_publication_date: 2025/4. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.naacl-demo.36/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.02950v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-02950/"}
{"id": "arxiv:2503.03196", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cacb33df18242eb3f47f6c89c56e41a146dd110e4b4dc5340062b9b981250", "submitted_at": "2025-03-05T05:30:22Z", "updated_at": "2025-04-16T15:34:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.03196v2"}, "authors": ["Zhiyuan Huang", "Ziming Cheng", "Junting Pan", "Zhaohui Hou", "Mingjie Zhan"], "canonical_source_url": "https://arxiv.org/abs/2503.03196", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-03196/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cacb33df18242eb3f47f6c89c56e41a146dd110e4b4dc5340062b9b981250", "submitted_at": "2025-03-05T05:30:22Z", "updated_at": "2025-04-16T15:34:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.03196v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.03196"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.03196"}], "name": "SpiritSight Agent: Advanced GUI Agent with One Look", "publication": null, "published_at": "2025-03-05", "record_sha256": "25df5f51f64f4787e10e9cc00191a04842569a67ed9e2c4beee13ef6d1202515", "relations": [], "resource_id": "arxiv:2503.03196", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.03196v2", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": "2025-04-16"}, "text": "SpiritSight Agent: Advanced GUI Agent with One Look\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SpiritSight and GUI-Lasagne data, combining multilevel GUI annotations with visual block parsing for high-resolution interface navigation.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding, data-generation\nPublished: 2025-03-05\nAuthors: Zhiyuan Huang, Ziming Cheng, Junting Pan, Zhaohui Hou, Mingjie Zhan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.03196v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-03196/"}
{"id": "arxiv:2503.03743", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "380d00c46e5dd221b7c481a68b4fbdc00826e64e4b70a98f06a56f8c12b8f5e4", "submitted_at": "2025-03-05T18:56:16Z", "updated_at": "2025-03-05T18:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.03743v1"}, "authors": ["Yuqi Zhou", "Shuai Wang", "Sunhao Dai", "Qinglin Jia", "Zhaocheng Du", "Zhenhua Dong", "Jun Xu"], "canonical_source_url": "https://arxiv.org/abs/2503.03743", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-03743/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "380d00c46e5dd221b7c481a68b4fbdc00826e64e4b70a98f06a56f8c12b8f5e4", "submitted_at": "2025-03-05T18:56:16Z", "updated_at": "2025-03-05T18:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.03743v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2503.03743"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.03743"}], "name": "CHOP: Mobile Operating Assistant with Constrained High-frequency Optimized Subtask Planning", "publication": null, "published_at": "2025-03-05", "record_sha256": "cc5b6c0897510198778f6d092ca5fa500943cddff31f12840634f4fc8ef5ec54", "relations": [], "resource_id": "arxiv:2503.03743", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.03743v1", "topics": ["mobile-navigation", "planning"], "updated_at": "2025-03-05"}, "text": "CHOP: Mobile Operating Assistant with Constrained High-frequency Optimized Subtask Planning\nResource type: paper\nReview status: curated\nScope: core\n\nUses human-planned subtasks to constrain and refine the intermediate plans connecting user goals with executable mobile operations.\nPlatforms: unspecified\nTopics: mobile-navigation, planning\nPublished: 2025-03-05\nAuthors: Yuqi Zhou, Shuai Wang, Sunhao Dai, Qinglin Jia, Zhaocheng Du, Zhenhua Dong, Jun Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.03743v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-03743/"}
{"id": "arxiv:2503.04730", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc499e361f0bc46c3c5ee5a8abebd911396e383a004655de56db1ec7cec8e231", "submitted_at": "2025-01-27T08:29:17Z", "updated_at": "2025-01-27T08:29:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04730v1"}, "authors": ["Zheng Hui", "Yinheng Li", "Dan zhao", "Tianyi Chen", "Colby Banbury", "Kazuhito Koishida"], "canonical_source_url": "https://arxiv.org/abs/2503.04730", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-04730/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc499e361f0bc46c3c5ee5a8abebd911396e383a004655de56db1ec7cec8e231", "submitted_at": "2025-01-27T08:29:17Z", "updated_at": "2025-01-27T08:29:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04730v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.04730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.04730"}], "name": "WinClick: GUI Grounding with Multimodal Large Language Models", "publication": null, "published_at": "2025-01-27", "record_sha256": "71308d808831a201fce8c487e21a74ddbe10e1cb35d865c8563044db6560dd47", "relations": [], "resource_id": "arxiv:2503.04730", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.04730v1", "topics": ["grounding", "evaluation"], "updated_at": "2025-01-27"}, "text": "WinClick: GUI Grounding with Multimodal Large Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a screenshot-based Windows GUI agent and the WinSpot benchmark for evaluating localization of actionable interface regions.\nPlatforms: desktop, windows\nTopics: grounding, evaluation\nPublished: 2025-01-27\nAuthors: Zheng Hui, Yinheng Li, Dan zhao, Tianyi Chen, Colby Banbury, Kazuhito Koishida\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.04730v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-04730/"}
{"id": "arxiv:2503.04957", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4588022dd47624058db8a16d7f7763c591241298e94594b56bfeb4ab2e01737", "submitted_at": "2025-03-06T20:43:14Z", "updated_at": "2025-03-06T20:43:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04957v1"}, "authors": ["Ada Defne Tur", "Nicholas Meade", "Xing Han Lù", "Alejandra Zambrano", "Arkil Patel", "Esin Durmus", "Spandana Gella", "Karolina Stańczak", "Siva Reddy"], "canonical_source_url": "https://arxiv.org/abs/2503.04957", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-04957/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4588022dd47624058db8a16d7f7763c591241298e94594b56bfeb4ab2e01737", "submitted_at": "2025-03-06T20:43:14Z", "updated_at": "2025-03-06T20:43:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.04957v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.04957"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.04957"}], "name": "SafeArena: Evaluating the Safety of Autonomous Web Agents", "publication": null, "published_at": "2025-03-06", "record_sha256": "823a9e37264bce388ef53ae8cf4eb6995a13291e6ff7fcab9d312bb6bdb2cc54", "relations": [], "resource_id": "arxiv:2503.04957", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.04957v1", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-06"}, "text": "SafeArena: Evaluating the Safety of Autonomous Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPairs benign and harmful web tasks to assess deliberate misuse risks and categorize agent behavior by risk level.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-03-06\nAuthors: Ada Defne Tur, Nicholas Meade, Xing Han Lù, Alejandra Zambrano, Arkil Patel, Esin Durmus, Spandana Gella, Karolina Stańczak, Siva Reddy\nRecorded arXiv version: v1\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.04957v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-04957/"}
{"id": "arxiv:2503.05143", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b408c850ae83e603a1b3d8303164b9f2d7be51580be99027275b0734ea8f56f2", "submitted_at": "2025-03-07T04:52:20Z", "updated_at": "2025-03-07T04:52:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.05143v1"}, "authors": ["Wenhao Wang", "Zijie Yu", "Rui Ye", "Jianqing Zhang", "Siheng Chen", "Yanfeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2503.05143", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-05143/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b408c850ae83e603a1b3d8303164b9f2d7be51580be99027275b0734ea8f56f2", "submitted_at": "2025-03-07T04:52:20Z", "updated_at": "2025-03-07T04:52:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.05143v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2503.05143"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.05143"}], "name": "FedMABench: Benchmarking Mobile Agents on Decentralized Heterogeneous User Data", "publication": null, "published_at": "2025-03-07", "record_sha256": "ca8c80ac5afddad11c6e45e13dcdedbedfa11c75055c136815e76d6064e9778f", "relations": [], "resource_id": "arxiv:2503.05143", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.05143v1", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2025-03-07"}, "text": "FedMABench: Benchmarking Mobile Agents on Decentralized Heterogeneous User Data\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a benchmark for federated mobile-agent training under heterogeneous user and application distributions, combining datasets, algorithms, and base models.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation\nPublished: 2025-03-07\nAuthors: Wenhao Wang, Zijie Yu, Rui Ye, Jianqing Zhang, Siheng Chen, Yanfeng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.05143v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-05143/"}
{"id": "arxiv:2503.06470", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebc104ee6a868092ae0215f84a3c5c71894bfb321671476a74b794e1645210c7", "submitted_at": "2025-03-09T06:14:17Z", "updated_at": "2025-03-09T06:14:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.06470v1"}, "authors": ["Fei Tang", "Yongliang Shen", "Hang Zhang", "Siqi Chen", "Guiyang Hou", "Wenqi Zhang", "Wenqiao Zhang", "Kaitao Song", "Weiming Lu", "Yueting Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2503.06470", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-06470/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebc104ee6a868092ae0215f84a3c5c71894bfb321671476a74b794e1645210c7", "submitted_at": "2025-03-09T06:14:17Z", "updated_at": "2025-03-09T06:14:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.06470v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.06470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.06470"}], "name": "Think Twice, Click Once: Enhancing GUI Grounding via Fast and Slow Systems", "publication": null, "published_at": "2025-03-09", "record_sha256": "0f8083553591b42f985f6b89e62cd2f5301c32b17bfcced3c492db8a9f4068f4", "relations": [], "resource_id": "arxiv:2503.06470", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.06470v1", "topics": ["grounding", "efficiency", "planning"], "updated_at": "2025-03-09"}, "text": "Think Twice, Click Once: Enhancing GUI Grounding via Fast and Slow Systems\nResource type: paper\nReview status: curated\nScope: core\n\nCombines rapid target prediction with deliberate interface analysis, adaptively selecting the grounding process according to task complexity.\nPlatforms: cross-platform\nTopics: grounding, efficiency, planning\nPublished: 2025-03-09\nAuthors: Fei Tang, Yongliang Shen, Hang Zhang, Siqi Chen, Guiyang Hou, Wenqi Zhang, Wenqiao Zhang, Kaitao Song, Weiming Lu, Yueting Zhuang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.06470v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-06470/"}
{"id": "arxiv:2503.07919", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1cd6327559d11982fb4a97e739cb6b5234c509859692084ed78130241f6d252", "submitted_at": "2025-03-10T23:50:30Z", "updated_at": "2025-07-24T17:45:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.07919v3"}, "authors": ["Yixiao Song", "Katherine Thai", "Chau Minh Pham", "Yapei Chang", "Mazin Nadaf", "Mohit Iyyer"], "canonical_source_url": "https://arxiv.org/abs/2503.07919", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-07919/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1cd6327559d11982fb4a97e739cb6b5234c509859692084ed78130241f6d252", "submitted_at": "2025-03-10T23:50:30Z", "updated_at": "2025-07-24T17:45:05Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.07919v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.07919"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.07919"}], "name": "BEARCUBS: A benchmark for computer-using web agents", "publication": null, "published_at": "2025-03-10", "record_sha256": "c1e36cb0c653498bdf5d1b5422e498aa8a396a345446b5170add78e76282a5e0", "relations": [], "resource_id": "arxiv:2503.07919", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.07919v3", "topics": ["evaluation", "web-navigation", "video"], "updated_at": "2025-07-24"}, "text": "BEARCUBS: A benchmark for computer-using web agents\nResource type: paper\nReview status: curated\nScope: core\n\nBEARCUBS provides 111 live-web information-seeking questions with short answers and human-validated browsing trajectories. Its tasks require multimodal computer interaction, including video understanding and navigation of interactive 3D content.\nPlatforms: web\nTopics: evaluation, web-navigation, video\nPublished: 2025-03-10\nAuthors: Yixiao Song, Katherine Thai, Chau Minh Pham, Yapei Chang, Mazin Nadaf, Mohit Iyyer\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.07919v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-07919/"}
{"id": "arxiv:2503.09241", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d4423ee1e7047a6228f5b29684618fc4108af0694484a7ceddf6c3b53fbb06c", "submitted_at": "2025-03-12T10:38:15Z", "updated_at": "2025-03-12T10:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09241v1"}, "authors": ["Pei Yang", "Hai Ci", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2503.09241", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-09241/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d4423ee1e7047a6228f5b29684618fc4108af0694484a7ceddf6c3b53fbb06c", "submitted_at": "2025-03-12T10:38:15Z", "updated_at": "2025-03-12T10:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09241v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2503.09241"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09241"}], "name": "In-Context Defense in Computer Agents: An Empirical Study", "publication": null, "published_at": "2025-03-12", "record_sha256": "b92b3c8fbb2b2836eae7f03cdccfd8c8087d275b6f73662cdae37fe09e18b2a7", "relations": [], "resource_id": "arxiv:2503.09241", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.09241v1", "topics": ["safety-security", "evaluation"], "updated_at": "2025-03-12"}, "text": "In-Context Defense in Computer Agents: An Empirical Study\nResource type: paper\nReview status: curated\nScope: core\n\nStudies in-context defenses for computer agents exposed to misleading pop-ups and other environmental instructions. It examines defensive reasoning before action planning and how a small set of examples changes resistance to these attacks.\nPlatforms: unspecified\nTopics: safety-security, evaluation\nPublished: 2025-03-12\nAuthors: Pei Yang, Hai Ci, Mike Zheng Shou\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.09241v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-09241/"}
{"id": "arxiv:2503.09263", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ea50d595c5d82b4a5d538b4b20725fa1f5afe5b29b03c022b726c307c8cca6e", "submitted_at": "2025-03-12T11:03:27Z", "updated_at": "2025-03-12T11:03:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09263v1"}, "authors": ["Di Zhao", "Longhui Ma", "Siwei Wang", "Miao Wang", "Zhao Lv"], "canonical_source_url": "https://arxiv.org/abs/2503.09263", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-09263/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ea50d595c5d82b4a5d538b4b20725fa1f5afe5b29b03c022b726c307c8cca6e", "submitted_at": "2025-03-12T11:03:27Z", "updated_at": "2025-03-12T11:03:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.09263v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2503.09263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09263"}], "name": "COLA: A Scalable Multi-Agent Framework For Windows UI Task Automation", "publication": null, "published_at": "2025-03-12", "record_sha256": "bad51141f2f622f1df1de0db6cfe4010a511edd27cae23318bb4e9dbd1c8d49f", "relations": [], "resource_id": "arxiv:2503.09263", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.09263v1", "topics": ["planning", "memory", "human-in-the-loop"], "updated_at": "2025-03-12"}, "text": "COLA: A Scalable Multi-Agent Framework For Windows UI Task Automation\nResource type: paper\nReview status: curated\nScope: core\n\nCoordinates specialized Windows UI agents with dynamic task scheduling, shared memory, and human-triggered rollback for recovering from errors.\nPlatforms: desktop, windows\nTopics: planning, memory, human-in-the-loop\nPublished: 2025-03-12\nAuthors: Di Zhao, Longhui Ma, Siwei Wang, Miao Wang, Zhao Lv\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.09263v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-09263/"}
{"id": "arxiv:2503.09572", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b03d84f29bb524d7950039624e0d460d8381f2ba63d70c69f5f958013730db9", "submitted_at": "2025-03-12T17:40:52Z", "updated_at": "2025-04-22T17:56:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09572v3"}, "authors": ["Lutfi Eren Erdogan", "Nicholas Lee", "Sehoon Kim", "Suhong Moon", "Hiroki Furuta", "Gopala Anumanchipalli", "Kurt Keutzer", "Amir Gholami"], "canonical_source_url": "https://arxiv.org/abs/2503.09572", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-09572/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b03d84f29bb524d7950039624e0d460d8381f2ba63d70c69f5f958013730db9", "submitted_at": "2025-03-12T17:40:52Z", "updated_at": "2025-04-22T17:56:22Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09572v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2503.09572"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09572"}], "name": "Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks", "publication": null, "published_at": "2025-03-12", "record_sha256": "01522df78e9b02c99651ce0e523e5e7ffe62da4ebf46a2eae9a221c333a0b7f3", "relations": [], "resource_id": "arxiv:2503.09572", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.09572v3", "topics": ["planning", "long-horizon", "data-generation"], "updated_at": "2025-04-22"}, "text": "Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates planning from execution and creates synthetic plan annotations for training agents on extended web-navigation tasks.\nPlatforms: web\nTopics: planning, long-horizon, data-generation\nPublished: 2025-03-12\nAuthors: Lutfi Eren Erdogan, Nicholas Lee, Sehoon Kim, Suhong Moon, Hiroki Furuta, Gopala Anumanchipalli, Kurt Keutzer, Amir Gholami\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.09572v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-09572/"}
{"id": "arxiv:2503.09780", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbb0726446c9679781b8e5930da0d3f406989a4b47759ce3440c9ad56d5839d8", "submitted_at": "2025-03-12T19:30:31Z", "updated_at": "2025-10-01T23:55:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09780v3"}, "authors": ["Arman Zharmagambetov", "Chuan Guo", "Ivan Evtimov", "Maya Pavlova", "Ruslan Salakhutdinov", "Kamalika Chaudhuri"], "canonical_source_url": "https://arxiv.org/abs/2503.09780", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-09780/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbb0726446c9679781b8e5930da0d3f406989a4b47759ce3440c9ad56d5839d8", "submitted_at": "2025-03-12T19:30:31Z", "updated_at": "2025-10-01T23:55:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.09780v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.09780"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.09780"}], "name": "AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents", "publication": null, "published_at": "2025-03-12", "record_sha256": "2c75f3230f6d389dc5a768ceacb9a64014de5fc6e30f65716bc15e37b784d249", "relations": [], "resource_id": "arxiv:2503.09780", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.09780v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-01"}, "text": "AgentDAM: Privacy Leakage Evaluation for Autonomous Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates whether autonomous web agents minimize use of sensitive personal information when completing tasks that expose unnecessary private data.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-03-12\nAuthors: Arman Zharmagambetov, Chuan Guo, Ivan Evtimov, Maya Pavlova, Ruslan Salakhutdinov, Kamalika Chaudhuri\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.09780v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-09780/"}
{"id": "arxiv:2503.11069", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a533866496d3ebd3e34c04142a702e4a3aaa387c3830e3bb993fdbb0ce8432", "submitted_at": "2025-03-14T04:26:21Z", "updated_at": "2025-06-23T13:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.11069v2"}, "authors": ["Chaoyun Zhang", "Shilin He", "Liqun Li", "Si Qin", "Yu Kang", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2503.11069", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-11069/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a533866496d3ebd3e34c04142a702e4a3aaa387c3830e3bb993fdbb0ce8432", "submitted_at": "2025-03-14T04:26:21Z", "updated_at": "2025-06-23T13:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.11069v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.11069"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.11069"}], "name": "API Agents vs. GUI Agents: Divergence and Convergence", "publication": null, "published_at": "2025-03-14", "record_sha256": "6abe7534353aeb4b4342c7e349f49e6546af654eaad383684eb03e8f506e21a1", "relations": [], "resource_id": "arxiv:2503.11069", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.11069v2", "topics": ["tool-integration", "survey", "automation"], "updated_at": "2025-06-23"}, "text": "API Agents vs. GUI Agents: Divergence and Convergence\nResource type: paper\nReview status: curated\nScope: core\n\nCompares API-driven and GUI-driven agents across architecture, development, and interaction considerations, discussing when hybrid approaches are useful.\nPlatforms: cross-platform\nTopics: tool-integration, survey, automation\nPublished: 2025-03-14\nAuthors: Chaoyun Zhang, Shilin He, Liqun Li, Si Qin, Yu Kang, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.11069v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-11069/"}
{"id": "arxiv:2503.11170", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2e6250e87ea9340914b457c4b0d471450f8cf49d2e2d53abd8e9e76105e908a", "submitted_at": "2025-03-14T08:16:02Z", "updated_at": "2025-03-14T08:16:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.11170v1"}, "authors": ["Yibin Xu", "Liang Yang", "Hao Chen", "Hua Wang", "Zhi Chen", "Yaohua Tang"], "canonical_source_url": "https://arxiv.org/abs/2503.11170", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-11170/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2e6250e87ea9340914b457c4b0d471450f8cf49d2e2d53abd8e9e76105e908a", "submitted_at": "2025-03-14T08:16:02Z", "updated_at": "2025-03-14T08:16:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.11170v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.11170"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.11170"}], "name": "DeskVision: Large Scale Desktop Region Captioning for Advanced GUI Agents", "publication": null, "published_at": "2025-03-14", "record_sha256": "d9a7083e71389d29997ea863a800ca9d95a3040b3ba347b1201121bbe1a721e7", "relations": [], "resource_id": "arxiv:2503.11170", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.11170v1", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-03-14"}, "text": "DeskVision: Large Scale Desktop Region Captioning for Advanced GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AutoCaptioner, DeskVision, and GUIExplorer for generating desktop region descriptions and training interface-understanding models.\nPlatforms: desktop\nTopics: data-generation, grounding, screen-understanding\nPublished: 2025-03-14\nAuthors: Yibin Xu, Liang Yang, Hao Chen, Hua Wang, Zhi Chen, Yaohua Tang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.11170v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-11170/"}
{"id": "arxiv:2503.12532", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e883289bafee01759a0f4f3f7051cbbebdcd8059e7520d68c25ba313b8da69a6", "submitted_at": "2025-03-16T14:53:43Z", "updated_at": "2025-03-24T16:33:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.12532v2"}, "authors": ["Fanbin Lu", "Zhisheng Zhong", "Ziqin Wei", "Shu Liu", "Chi-Wing Fu", "Jiaya Jia"], "canonical_source_url": "https://arxiv.org/abs/2503.12532", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-12532/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e883289bafee01759a0f4f3f7051cbbebdcd8059e7520d68c25ba313b8da69a6", "submitted_at": "2025-03-16T14:53:43Z", "updated_at": "2025-03-24T16:33:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.12532v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.12532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.12532"}], "name": "STEVE: A Step Verification Pipeline for Computer-use Agent Training", "publication": null, "published_at": "2025-03-16", "record_sha256": "3d38364a09a88bf35ad7326a57fcb966cf7bbf7af0c5b3560a5dcd0b6ebb3e78", "relations": [], "resource_id": "arxiv:2503.12532", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.12532v2", "topics": ["data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2025-03-24"}, "text": "STEVE: A Step Verification Pipeline for Computer-use Agent Training\nResource type: paper\nReview status: curated\nScope: core\n\nLabels individual computer-use actions by comparing surrounding screenshots, then trains agents from both successful and unsuccessful steps.\nPlatforms: desktop, windows\nTopics: data-generation, evaluation, reinforcement-learning\nPublished: 2025-03-16\nAuthors: Fanbin Lu, Zhisheng Zhong, Ziqin Wei, Shu Liu, Chi-Wing Fu, Jiaya Jia\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.12532v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-12532/"}
{"id": "arxiv:2503.14021", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b260fd1d7e28c8263710f116638e7aa7b51f46aac15199c8a29092161a13f37d", "submitted_at": "2025-03-18T08:32:22Z", "updated_at": "2025-03-18T08:32:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.14021v1"}, "authors": ["Ziwei Wang", "Weizhi Chen", "Leyang Yang", "Sheng Zhou", "Shengchu Zhao", "Hanbei Zhan", "Jiongchao Jin", "Liangcheng Li", "Zirui Shao", "Jiajun Bu"], "canonical_source_url": "https://arxiv.org/abs/2503.14021", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-14021/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b260fd1d7e28c8263710f116638e7aa7b51f46aac15199c8a29092161a13f37d", "submitted_at": "2025-03-18T08:32:22Z", "updated_at": "2025-03-18T08:32:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.14021v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2503.14021"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.14021"}], "name": "MP-GUI: Modality Perception with MLLMs for GUI Understanding", "publication": null, "published_at": "2025-03-18", "record_sha256": "75bafc3f2b339d92744a01e4fd88280ec4673a95ca184da5320c30527bc053f4", "relations": [], "resource_id": "arxiv:2503.14021", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.14021v1", "topics": ["screen-understanding", "grounding", "data-generation"], "updated_at": "2025-03-18"}, "text": "MP-GUI: Modality Perception with MLLMs for GUI Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nUses specialized graphical, textual, and spatial perception components with a data-collection pipeline to improve GUI understanding.\nPlatforms: cross-platform\nTopics: screen-understanding, grounding, data-generation\nPublished: 2025-03-18\nAuthors: Ziwei Wang, Weizhi Chen, Leyang Yang, Sheng Zhou, Shengchu Zhao, Hanbei Zhan, Jiongchao Jin, Liangcheng Li, Zirui Shao, Jiajun Bu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.14021v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-14021/"}
{"id": "arxiv:2503.15515", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c778da33baa88916fbcb3e02a33f760dccea8b193d25087df9b56cd42c3c248", "submitted_at": "2025-01-31T12:26:27Z", "updated_at": "2025-01-31T12:26:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.15515v1"}, "authors": ["Piero A. Bonatti", "John Domingue", "Anna Lisa Gentile", "Andreas Harth", "Olaf Hartig", "Aidan Hogan", "Katja Hose", "Ernesto Jimenez-Ruiz", "Deborah L. McGuinness", "Chang Sun", "Ruben Verborgh", "Jesse Wright"], "canonical_source_url": "https://arxiv.org/abs/2503.15515", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-15515/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c778da33baa88916fbcb3e02a33f760dccea8b193d25087df9b56cd42c3c248", "submitted_at": "2025-01-31T12:26:27Z", "updated_at": "2025-01-31T12:26:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.15515v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.15515"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15515"}], "name": "Towards Computer-Using Personal Agents", "publication": null, "published_at": "2025-01-31", "record_sha256": "43ed9d0adf255c5a4985e12fdea7fe61cda808d2870f31bc5a0c7a4d3abd9934", "relations": [], "resource_id": "arxiv:2503.15515", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.15515v1", "topics": ["safety-security", "memory", "human-in-the-loop"], "updated_at": "2025-01-31"}, "text": "Towards Computer-Using Personal Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProposes computer-using personal agents connected to user-controlled data repositories, discussing privacy, interoperability, and collaborative tasks.\nPlatforms: web, cross-platform\nTopics: safety-security, memory, human-in-the-loop\nPublished: 2025-01-31\nAuthors: Piero A. Bonatti, John Domingue, Anna Lisa Gentile, Andreas Harth, Olaf Hartig, Aidan Hogan, Katja Hose, Ernesto Jimenez-Ruiz, Deborah L. McGuinness, Chang Sun, Ruben Verborgh, Jesse Wright\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.15515v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-15515/"}
{"id": "arxiv:2503.15661", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5abf3c7b08a4331d4f634d8cc63a7f8ff40713f7344231887dac1b5e061a8eae", "submitted_at": "2025-03-19T19:26:17Z", "updated_at": "2025-05-06T17:43:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.15661v2"}, "authors": ["Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Juan A. Rodriguez", "Montek Kalsi", "Rabiul Awal", "Nicolas Chapados", "M. Tamer Özsu", "Aishwarya Agrawal", "David Vazquez", "Christopher Pal", "Perouz Taslakian", "Spandana Gella", "Sai Rajeswar"], "canonical_source_url": "https://arxiv.org/abs/2503.15661", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-15661/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5abf3c7b08a4331d4f634d8cc63a7f8ff40713f7344231887dac1b5e061a8eae", "submitted_at": "2025-03-19T19:26:17Z", "updated_at": "2025-05-06T17:43:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.15661v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {"arxiv": "2503.15661"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:38Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15661"}], "name": "UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction", "publication": null, "published_at": "2025-03-19", "record_sha256": "e71d5d84fe59d9ea7193065b87eecae3726e8c00daa59ad5b1b797081dee6b82", "relations": [], "resource_id": "arxiv:2503.15661", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.15661v2", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2025-05-06"}, "text": "UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nProvides annotated desktop demonstrations for evaluating element grounding, layout grounding, and action prediction.\nPlatforms: desktop\nTopics: grounding, evaluation, screen-understanding\nPublished: 2025-03-19\nAuthors: Shravan Nayak, Xiangru Jian, Kevin Qinghong Lin, Juan A. Rodriguez, Montek Kalsi, Rabiul Awal, Nicolas Chapados, M. Tamer Özsu, Aishwarya Agrawal, David Vazquez, Christopher Pal, Perouz Taslakian, Spandana Gella, Sai Rajeswar\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.15661v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-15661/"}
{"id": "arxiv:2503.15937", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6449576cb8ba45569b57844d87f22d01645b3639a222c648d914dc38e3421553", "submitted_at": "2025-03-20T08:25:00Z", "updated_at": "2026-02-21T01:36:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.15937v5"}, "authors": ["Gaole Dai", "Shiqi Jiang", "Ting Cao", "Yuanchun Li", "Yuqing Yang", "Rui Tan", "Mo Li", "Lili Qiu"], "canonical_source_url": "https://arxiv.org/abs/2503.15937", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-15937/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6449576cb8ba45569b57844d87f22d01645b3639a222c648d914dc38e3421553", "submitted_at": "2025-03-20T08:25:00Z", "updated_at": "2026-02-21T01:36:43Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.15937v5"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.15937"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.15937"}], "name": "Advancing Mobile GUI Agents: A Verifier-Driven Approach to Practical Deployment", "publication": null, "published_at": "2025-03-20", "record_sha256": "5e5277369b9dc34a76603af3447580a46567915b8c669adf9a126a08557fab84", "relations": [], "resource_id": "arxiv:2503.15937", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.15937v5", "topics": ["mobile-navigation", "efficiency", "planning"], "updated_at": "2026-02-21"}, "text": "Advancing Mobile GUI Agents: A Verifier-Driven Approach to Practical Deployment\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces V-Droid, which verifies candidate actions in a discrete action space rather than directly generating each mobile operation.\nPlatforms: android\nTopics: mobile-navigation, efficiency, planning\nPublished: 2025-03-20\nAuthors: Gaole Dai, Shiqi Jiang, Ting Cao, Yuanchun Li, Yuqing Yang, Rui Tan, Mo Li, Lili Qiu\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.15937v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-15937/"}
{"id": "arxiv:2503.16465", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6201528b13c6a1aae8dcd745f4f5d995e79982ef7c54a1962e2f9035233293", "submitted_at": "2025-02-26T12:31:16Z", "updated_at": "2025-07-14T15:56:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.16465v3"}, "authors": ["Pengzhou Cheng", "Zheng Wu", "Zongru Wu", "Aston Zhang", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2503.16465", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-16465/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6201528b13c6a1aae8dcd745f4f5d995e79982ef7c54a1962e2f9035233293", "submitted_at": "2025-02-26T12:31:16Z", "updated_at": "2025-07-14T15:56:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.16465v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.16465"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.16465"}], "name": "OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents", "publication": null, "published_at": "2025-02-26", "record_sha256": "5356a6152cb6e330437eeeed54fd9b45f4df912a541590998bbba04afb494009", "relations": [], "resource_id": "arxiv:2503.16465", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.16465v3", "topics": ["human-in-the-loop", "safety-security", "mobile-navigation"], "updated_at": "2025-07-14"}, "text": "OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses step-level confidence estimates to decide when a GUI agent should act autonomously or request human intervention in complex situations.\nPlatforms: android\nTopics: human-in-the-loop, safety-security, mobile-navigation\nPublished: 2025-02-26\nAuthors: Pengzhou Cheng, Zheng Wu, Zongru Wu, Aston Zhang, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.16465v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-16465/"}
{"id": "arxiv:2503.16788", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e450216f8f802d8f3d16c5886097cef801d051951af0c57fdf195cc176ac744f", "submitted_at": "2025-03-21T01:52:43Z", "updated_at": "2025-03-21T01:52:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.16788v1"}, "authors": ["Li Zhang", "Longxi Gao", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2503.16788", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-16788/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e450216f8f802d8f3d16c5886097cef801d051951af0c57fdf195cc176ac744f", "submitted_at": "2025-03-21T01:52:43Z", "updated_at": "2025-03-21T01:52:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.16788v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.16788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.16788"}], "name": "Does Chain-of-Thought Reasoning Help Mobile GUI Agent? An Empirical Study", "publication": null, "published_at": "2025-03-21", "record_sha256": "b50902c319c77f6d82848a38eb98542e8529594dd05ce6173a0384bcda658845", "relations": [], "resource_id": "arxiv:2503.16788", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.16788v1", "topics": ["evaluation", "mobile-navigation", "planning"], "updated_at": "2025-03-21"}, "text": "Does Chain-of-Thought Reasoning Help Mobile GUI Agent? An Empirical Study\nResource type: paper\nReview status: curated\nScope: core\n\nThis study compares matched reasoning and non-reasoning vision-language models on static mobile GUI tasks and interactive AndroidWorld navigation. It finds that chain-of-thought effects vary across tasks and model families.\nPlatforms: android\nTopics: evaluation, mobile-navigation, planning\nPublished: 2025-03-21\nAuthors: Li Zhang, Longxi Gao, Mengwei Xu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.16788v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-16788/"}
{"id": "arxiv:2503.17709", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80bb507bf3cdccf6f023d0c334899d9d7099c362e9f37c314b15ff9388bae65c", "submitted_at": "2025-03-22T09:30:37Z", "updated_at": "2025-03-22T09:30:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.17709v1"}, "authors": ["Yuchen Sun", "Shanhui Zhao", "Tao Yu", "Hao Wen", "Samith Va", "Mengwei Xu", "Yuanchun Li", "Chongyang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2503.17709", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-17709/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80bb507bf3cdccf6f023d0c334899d9d7099c362e9f37c314b15ff9388bae65c", "submitted_at": "2025-03-22T09:30:37Z", "updated_at": "2025-03-22T09:30:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.17709v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.17709"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.17709"}], "name": "GUI-Xplore: Empowering Generalizable GUI Agents with One Exploration", "publication": null, "published_at": "2025-03-22", "record_sha256": "b8361678893e17e6a31e887790d014b0c32976fa9367c6315cbd81b478635e62", "relations": [], "resource_id": "arxiv:2503.17709", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.17709v1", "topics": ["data-generation", "planning", "video"], "updated_at": "2025-03-22"}, "text": "GUI-Xplore: Empowering Generalizable GUI Agents with One Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-Xplore exploration videos and downstream tasks, with an agent that combines action-aware modeling and graph-based environment reasoning.\nPlatforms: cross-platform\nTopics: data-generation, planning, video\nPublished: 2025-03-22\nAuthors: Yuchen Sun, Shanhui Zhao, Tao Yu, Hao Wen, Samith Va, Mengwei Xu, Yuanchun Li, Chongyang Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.17709v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-17709/"}
{"id": "arxiv:2503.18492", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10e7a9a4471555c6e4001cd10c1e4b375b503adfe6ea48d5808d1d7721dfc162", "submitted_at": "2025-03-24T09:46:05Z", "updated_at": "2025-09-11T04:15:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.18492v2"}, "authors": ["Jungjae Lee", "Dongjae Lee", "Chihun Choi", "Youngmin Im", "Jaeyoung Wi", "Kihong Heo", "Sangeun Oh", "Sunjae Lee", "Insik Shin"], "canonical_source_url": "https://arxiv.org/abs/2503.18492", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-18492/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10e7a9a4471555c6e4001cd10c1e4b375b503adfe6ea48d5808d1d7721dfc162", "submitted_at": "2025-03-24T09:46:05Z", "updated_at": "2025-09-11T04:15:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.18492v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.18492"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.18492"}], "name": "VeriSafe Agent: Safeguarding Mobile GUI Agent via Logic-based Action Verification", "publication": null, "published_at": "2025-03-24", "record_sha256": "7a2eac9f74845ea3da4edb40f64a791a03fd7705d0a79c8e9090a819fb1d14bf", "relations": [], "resource_id": "arxiv:2503.18492", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.18492v2", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2025-09-11"}, "text": "VeriSafe Agent: Safeguarding Mobile GUI Agent via Logic-based Action Verification\nResource type: paper\nReview status: curated\nScope: core\n\nTranslates user instructions into formal specifications and applies runtime logic-based checks to proposed mobile GUI actions.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation\nPublished: 2025-03-24\nAuthors: Jungjae Lee, Dongjae Lee, Chihun Choi, Youngmin Im, Jaeyoung Wi, Kihong Heo, Sangeun Oh, Sunjae Lee, Insik Shin\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.18492v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-18492/"}
{"id": "arxiv:2503.19537", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a49d662eddf84690395b025e50f3378c3be2a3fc8abe02bd7202c6ff879bbe3", "submitted_at": "2025-03-25T10:46:08Z", "updated_at": "2025-03-25T10:46:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.19537v1"}, "authors": ["Noam Kahlon", "Guy Rom", "Anatoly Efros", "Filippo Galgani", "Omri Berkovitch", "Sapir Caduri", "William E. Bishop", "Oriana Riva", "Ido Dagan"], "canonical_source_url": "https://arxiv.org/abs/2503.19537", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-19537/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a49d662eddf84690395b025e50f3378c3be2a3fc8abe02bd7202c6ff879bbe3", "submitted_at": "2025-03-25T10:46:08Z", "updated_at": "2025-03-25T10:46:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.19537v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2503.19537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.19537"}], "name": "Agent-Initiated Interaction in Phone UI Automation", "publication": null, "published_at": "2025-03-25", "record_sha256": "32557f9cd750ab1aeb686d350e8c96441753a85a5a46c60d565c68be8d2f0b60", "relations": [], "resource_id": "arxiv:2503.19537", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.19537v1", "topics": ["human-in-the-loop", "mobile-navigation", "evaluation"], "updated_at": "2025-03-25"}, "text": "Agent-Initiated Interaction in Phone UI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nDefines when phone agents should request user interaction and introduces AndroidInteraction annotations for deciding whether and how to engage the user.\nPlatforms: android\nTopics: human-in-the-loop, mobile-navigation, evaluation\nPublished: 2025-03-25\nAuthors: Noam Kahlon, Guy Rom, Anatoly Efros, Filippo Galgani, Omri Berkovitch, Sapir Caduri, William E. Bishop, Oriana Riva, Ido Dagan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.19537v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-19537/"}
{"id": "arxiv:2503.20279", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e90cbc547703a2673cfb30bbe9079496ce4c076b51f39e0557f5db9048947c8", "submitted_at": "2025-03-26T07:08:15Z", "updated_at": "2025-06-08T05:39:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.20279v3"}, "authors": ["Sejin Lee", "Jian Kim", "Haon Park", "Ashkan Yousefpour", "Sangyoon Yu", "Min Song"], "canonical_source_url": "https://arxiv.org/abs/2503.20279", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-20279/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e90cbc547703a2673cfb30bbe9079496ce4c076b51f39e0557f5db9048947c8", "submitted_at": "2025-03-26T07:08:15Z", "updated_at": "2025-06-08T05:39:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2503.20279v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2503.20279"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.20279"}], "name": "sudo rm -rf agentic_security", "publication": null, "published_at": "2025-03-26", "record_sha256": "1b022d3ff449c5ecd8a235c800fd9ccb25db016231b08668ae5cb0148858c7a7", "relations": [], "resource_id": "arxiv:2503.20279", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.20279v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-06-08"}, "text": "sudo rm -rf agentic_security\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates security failures of computer-use agents under adversarial requests and screen-context manipulation. The study examines refusal behavior while agents perform tasks in desktop and web environments.\nPlatforms: desktop, web\nTopics: safety-security, evaluation\nPublished: 2025-03-26\nAuthors: Sejin Lee, Jian Kim, Haon Park, Ashkan Yousefpour, Sangyoon Yu, Min Song\nRecorded arXiv version: v3\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.20279v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-20279/"}
{"id": "arxiv:2503.20978", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0888d0da1c5da9cf9956fd8b145a09f463aa432a60b142c4aea8d20b4aa84c4d", "submitted_at": "2025-03-26T20:41:24Z", "updated_at": "2025-03-26T20:41:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.20978v1"}, "authors": ["Yiqiao Jin", "Stefano Petrangeli", "Yu Shen", "Gang Wu"], "canonical_source_url": "https://arxiv.org/abs/2503.20978", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-20978/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0888d0da1c5da9cf9956fd8b145a09f463aa432a60b142c4aea8d20b4aa84c4d", "submitted_at": "2025-03-26T20:41:24Z", "updated_at": "2025-03-26T20:41:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2503.20978v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2503.20978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.20978"}], "name": "ScreenLLM: Stateful Screen Schema for Efficient Action Understanding and Prediction", "publication": null, "published_at": "2025-03-26", "record_sha256": "00b1f114e39d6e7425b3e22d564fc25fc7d72f08fec4e921bfe0af310add8357", "relations": [], "resource_id": "arxiv:2503.20978", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.20978v1", "topics": ["screen-understanding", "memory"], "updated_at": "2025-03-26"}, "text": "ScreenLLM: Stateful Screen Schema for Efficient Action Understanding and Prediction\nResource type: paper\nReview status: curated\nScope: core\n\nRepresents interface interactions with a stateful screen schema to model user behavior and support action understanding and prediction.\nPlatforms: cross-platform\nTopics: screen-understanding, memory\nPublished: 2025-03-26\nAuthors: Yiqiao Jin, Stefano Petrangeli, Yu Shen, Gang Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.20978v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-20978/"}
{"id": "arxiv:2503.21620", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b794a2f59cbe26a889b242ff47a7cd982f6367f759ddabcf99d96f31e5c122da", "submitted_at": "2025-03-27T15:39:30Z", "updated_at": "2025-05-24T08:46:08Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.21620v5"}, "authors": ["Zhengxi Lu", "Yuxiang Chai", "Yaxuan Guo", "Xi Yin", "Liang Liu", "Hao Wang", "Han Xiao", "Shuai Ren", "Guanjing Xiong", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2503.21620", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-21620/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b794a2f59cbe26a889b242ff47a7cd982f6367f759ddabcf99d96f31e5c122da", "submitted_at": "2025-03-27T15:39:30Z", "updated_at": "2025-05-24T08:46:08Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2503.21620v5"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.21620"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.21620"}], "name": "UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning", "publication": null, "published_at": "2025-03-27", "record_sha256": "d85bccbd9442b8c957fbf2c4b65df52d9f5ec95de4a07951ed55824b497b37e8", "relations": [], "resource_id": "arxiv:2503.21620", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.21620v5", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-05-24"}, "text": "UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUses rule-based action rewards and policy optimization to train GUI action prediction from a compact set of mobile tasks.\nPlatforms: cross-platform\nTopics: reinforcement-learning, grounding\nPublished: 2025-03-27\nAuthors: Zhengxi Lu, Yuxiang Chai, Yaxuan Guo, Xi Yin, Liang Liu, Hao Wang, Han Xiao, Shuai Ren, Guanjing Xiong, Hongsheng Li\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.21620v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-21620/"}
{"id": "arxiv:2503.22738", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362fe117797181225641b71b23df95d82c2f90ebf2bbfa1bebea56427e502091", "submitted_at": "2025-03-26T17:58:40Z", "updated_at": "2025-11-27T22:01:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.22738v2"}, "authors": ["Zhaorun Chen", "Mintong Kang", "Bo Li"], "canonical_source_url": "https://arxiv.org/abs/2503.22738", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-22738/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "362fe117797181225641b71b23df95d82c2f90ebf2bbfa1bebea56427e502091", "submitted_at": "2025-03-26T17:58:40Z", "updated_at": "2025-11-27T22:01:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.22738v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2503.22738"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.22738"}], "name": "ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning", "publication": null, "published_at": "2025-03-26", "record_sha256": "6fa6a1f1e4c728e90514e6210a30ed6804ff080caf2535be9ca1cf7e04c883a1", "relations": [], "resource_id": "arxiv:2503.22738", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.22738v2", "topics": ["safety-security"], "updated_at": "2025-11-27"}, "text": "ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning\nResource type: paper\nReview status: curated\nScope: core\n\nProposes a guardrail agent that reasons over explicit safety policies and checks action trajectories, with a web-agent safety benchmark.\nPlatforms: web\nTopics: safety-security\nPublished: 2025-03-26\nAuthors: Zhaorun Chen, Mintong Kang, Bo Li\nRecorded arXiv version: v2\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.22738v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-22738/"}
{"id": "arxiv:2503.23350", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b214bd8e5a34c0e8ca16a3be4fcf638a59a17ca524276dbef3c6b75f08398bb", "submitted_at": "2025-03-30T08:15:44Z", "updated_at": "2025-08-05T11:47:41Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2503.23350v4"}, "authors": ["Liangbo Ning", "Ziran Liang", "Zhuohang Jiang", "Haohao Qu", "Yujuan Ding", "Wenqi Fan", "Xiao-yong Wei", "Shanru Lin", "Hui Liu", "Philip S. Yu", "Qing Li"], "canonical_source_url": "https://arxiv.org/abs/2503.23350", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-23350/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4b214bd8e5a34c0e8ca16a3be4fcf638a59a17ca524276dbef3c6b75f08398bb", "submitted_at": "2025-03-30T08:15:44Z", "updated_at": "2025-08-05T11:47:41Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2503.23350v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2503.23350"}, "kind": "survey", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.23350"}], "name": "A Survey of WebAgents: Towards Next-Generation AI Agents for Web Automation with Large Foundation Models", "publication": null, "published_at": "2025-03-30", "record_sha256": "7c17451d2a680c52689b6960ddfd8927615ba86a8b36bb4009c40f4c60735b13", "relations": [], "resource_id": "arxiv:2503.23350", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.23350v4", "topics": ["survey", "web-navigation", "safety-security"], "updated_at": "2025-08-05"}, "text": "A Survey of WebAgents: Towards Next-Generation AI Agents for Web Automation with Large Foundation Models\nResource type: survey\nReview status: curated\nScope: core\n\nSurveys foundation-model web agents through their architectures, training methods and trustworthiness. It organizes research on completing browser tasks from user instructions and identifies directions for further development.\nPlatforms: web\nTopics: survey, web-navigation, safety-security\nPublished: 2025-03-30\nAuthors: Liangbo Ning, Ziran Liang, Zhuohang Jiang, Haohao Qu, Yujuan Ding, Wenqi Fan, Xiao-yong Wei, Shanru Lin, Hui Liu, Philip S. Yu, Qing Li\nRecorded arXiv version: v4\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.23350v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-23350/"}
{"id": "arxiv:2503.23434", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a6811c6d0c3b30d4e6af2f7b0beef8aa8f1ad99ceb325182305e571fe5c398f", "submitted_at": "2025-03-30T13:26:00Z", "updated_at": "2026-02-24T05:24:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.23434v2"}, "authors": ["Yucheng Shi", "Wenhao Yu", "Jingyuan Huang", "Wenlin Yao", "Wenhu Chen", "Ninghao Liu"], "canonical_source_url": "https://arxiv.org/abs/2503.23434", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-23434/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a6811c6d0c3b30d4e6af2f7b0beef8aa8f1ad99ceb325182305e571fe5c398f", "submitted_at": "2025-03-30T13:26:00Z", "updated_at": "2026-02-24T05:24:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.23434v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2503.23434"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.23434"}], "name": "Towards Trustworthy GUI Agents: A Survey", "publication": null, "published_at": "2025-03-30", "record_sha256": "2ad9215c8a39a9c088cfc9d255d4e2a3695d5b1812f22e55cc81893a65191600", "relations": [], "resource_id": "arxiv:2503.23434", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.23434v2", "topics": ["survey", "safety-security"], "updated_at": "2026-02-24"}, "text": "Towards Trustworthy GUI Agents: A Survey\nResource type: survey\nReview status: curated\nScope: core\n\nFrames GUI-agent trustworthiness as an execution gap and reviews failures, attacks, and defenses across perception, reasoning, and interaction.\nPlatforms: cross-platform\nTopics: survey, safety-security\nPublished: 2025-03-30\nAuthors: Yucheng Shi, Wenhao Yu, Jingyuan Huang, Wenlin Yao, Wenhu Chen, Ninghao Liu\nRecorded arXiv version: v2\n\nPrimary source: https://arxiv.org/abs/2503.23434v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-23434/"}
{"id": "arxiv:2503.24180", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bb09b54379d2fbc5b489f24a27f6311b91bb0821eefe5187aa160ffb00639d9", "submitted_at": "2025-03-31T14:56:24Z", "updated_at": "2025-08-24T08:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.24180v2"}, "authors": ["Ziming Cheng", "Zhiyuan Huang", "Junting Pan", "Zhaohui Hou", "Mingjie Zhan"], "canonical_source_url": "https://arxiv.org/abs/2503.24180", "citation_url": "https://awesomecua.github.io/resources/arxiv-2503-24180/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bb09b54379d2fbc5b489f24a27f6311b91bb0821eefe5187aa160ffb00639d9", "submitted_at": "2025-03-31T14:56:24Z", "updated_at": "2025-08-24T08:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2503.24180v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2503.24180"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2503.24180"}], "name": "Navi-plus: Managing Ambiguous GUI Navigation Tasks with Follow-up Questions", "publication": null, "published_at": "2025-03-31", "record_sha256": "ea3e352bdd9ab924deeee3a4e35b2588b710389ee74e04955d25bd433d6e53b6", "relations": [], "resource_id": "arxiv:2503.24180", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2503.24180v2", "topics": ["human-in-the-loop", "evaluation", "planning"], "updated_at": "2025-08-24"}, "text": "Navi-plus: Managing Ambiguous GUI Navigation Tasks with Follow-up Questions\nResource type: paper\nReview status: curated\nScope: core\n\nNavi-plus evaluates GUI agents that ask follow-up questions to complete ambiguous user instructions. It provides GUI clarification question-answer pairs and Dual-Stream Trajectory Evaluation for this interactive setting.\nPlatforms: unspecified\nTopics: human-in-the-loop, evaluation, planning\nPublished: 2025-03-31\nAuthors: Ziming Cheng, Zhiyuan Huang, Junting Pan, Zhaohui Hou, Mingjie Zhan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2503.24180v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2503-24180/"}
{"id": "arxiv:2504.00906", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "11958f69503db514a91af8721870a4cbc1fa19dabf94e3f4ce4118d2fad9862a", "submitted_at": "2025-04-01T15:40:27Z", "updated_at": "2025-04-01T15:40:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.00906v1"}, "authors": ["Saaket Agashe", "Kyle Wong", "Vincent Tu", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2504.00906", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-00906/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "11958f69503db514a91af8721870a4cbc1fa19dabf94e3f4ce4118d2fad9862a", "submitted_at": "2025-04-01T15:40:27Z", "updated_at": "2025-04-01T15:40:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.00906v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.00906"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.00906"}], "name": "Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents", "publication": null, "published_at": "2025-04-01", "record_sha256": "58827491a5499cbedf58aa1f058aa7db9e171d37f10c78fbc46553439597f240", "relations": [], "resource_id": "arxiv:2504.00906", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.00906v1", "topics": ["grounding", "planning"], "updated_at": "2025-04-01"}, "text": "Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines generalist and specialist models with hierarchical planning and multiple grounding methods for computer-use tasks.\nPlatforms: cross-platform\nTopics: grounding, planning\nPublished: 2025-04-01\nAuthors: Saaket Agashe, Kyle Wong, Vincent Tu, Jiachen Yang, Ang Li, Xin Eric Wang\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.00906v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-00906/"}
{"id": "arxiv:2504.01382", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccad667f7c5bb75e58d75d5eab5cc24c7c3610be9eed1a09079de5e23c85af04", "submitted_at": "2025-04-02T05:51:29Z", "updated_at": "2025-10-08T02:22:45Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.01382v4"}, "authors": ["Tianci Xue", "Weijian Qi", "Tianneng Shi", "Chan Hee Song", "Boyu Gou", "Dawn Song", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2504.01382", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-01382/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccad667f7c5bb75e58d75d5eab5cc24c7c3610be9eed1a09079de5e23c85af04", "submitted_at": "2025-04-02T05:51:29Z", "updated_at": "2025-10-08T02:22:45Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.01382v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.01382"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.01382"}], "name": "An Illusion of Progress? Assessing the Current State of Web Agents", "publication": null, "published_at": "2025-04-02", "record_sha256": "861940f4b25a4547e10508d8f2a7b0a3b58e2ac62b50d85d80409705345d57b1", "relations": [], "resource_id": "arxiv:2504.01382", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.01382v4", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-10-08"}, "text": "An Illusion of Progress? Assessing the Current State of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Online-Mind2Web, with 300 tasks across 136 live websites, and examines how benchmark design affects assessments of web agents. It also develops an LLM judge for task completion and compares its judgments with human evaluation.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2025-04-02\nAuthors: Tianci Xue, Weijian Qi, Tianneng Shi, Chan Hee Song, Boyu Gou, Dawn Song, Huan Sun, Yu Su\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.01382v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-01382/"}
{"id": "arxiv:2504.01495", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76d0ffb55478570499c8cddf92a27d346cf11749f8ae5f8e128c87653b2bdbdc", "submitted_at": "2025-04-02T08:48:01Z", "updated_at": "2025-04-02T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.01495v1"}, "authors": ["Antoine Chevrot", "Alexandre Vernotte", "Jean-Rémy Falleri", "Xavier Blanc", "Bruno Legeard"], "canonical_source_url": "https://arxiv.org/abs/2504.01495", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-01495/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76d0ffb55478570499c8cddf92a27d346cf11749f8ae5f8e128c87653b2bdbdc", "submitted_at": "2025-04-02T08:48:01Z", "updated_at": "2025-04-02T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.01495v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.01495"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.01495"}], "name": "Are Autonomous Web Agents Good Testers?", "publication": null, "published_at": "2025-04-02", "record_sha256": "4346c71da8314c41ed01cc819374072ef492f2c7aa09d5bfe270fe4f16b97b61", "relations": [], "resource_id": "arxiv:2504.01495", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.01495v1", "topics": ["automation", "evaluation", "web-navigation"], "updated_at": "2025-04-02"}, "text": "Are Autonomous Web Agents Good Testers?\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts web agents to execute natural-language test steps, verify assertions, and issue test verdicts. The study introduces 113 passing and failing test cases across three offline web applications and compares two agent implementations with qualitative failure analysis.\nPlatforms: web\nTopics: automation, evaluation, web-navigation\nPublished: 2025-04-02\nAuthors: Antoine Chevrot, Alexandre Vernotte, Jean-Rémy Falleri, Xavier Blanc, Bruno Legeard\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.01495v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-01495/"}
{"id": "arxiv:2504.02357", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a646bb2136cdaf47a0d8b3f485d3e1973fc15ac4f844e1beddedf6021b56ffe4", "submitted_at": "2025-04-03T07:45:09Z", "updated_at": "2025-04-03T07:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.02357v1"}, "authors": ["Xiaolei Li", "Jialun Cao", "Yepang Liu", "Shing-Chi Cheung", "Hailong Wang"], "canonical_source_url": "https://arxiv.org/abs/2504.02357", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-02357/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a646bb2136cdaf47a0d8b3f485d3e1973fc15ac4f844e1beddedf6021b56ffe4", "submitted_at": "2025-04-03T07:45:09Z", "updated_at": "2025-04-03T07:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.02357v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:12:09Z", "role": "primary", "sha256": "054a5a305d86c7fb9107d467d194fc4a64aa4f5af23d214bea7f52c2f2c9fd35", "supports": ["environments"], "url": "https://arxiv.org/html/2504.02357"}], "identifiers": {"arxiv": "2504.02357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.02357"}], "name": "ReuseDroid: A VLM-empowered Android UI Test Migrator Boosted by Active Feedback", "publication": null, "published_at": "2025-04-03", "record_sha256": "827fe8c81335c93e9fc6da88fd0b4ddc73a88485a44761ee886fb1b8b94bad37", "relations": [], "resource_id": "arxiv:2504.02357", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.02357v1", "topics": ["automation", "evaluation", "cross-app"], "updated_at": "2025-04-03"}, "text": "ReuseDroid: A VLM-empowered Android UI Test Migrator Boosted by Active Feedback\nResource type: paper\nReview status: curated\nScope: core\n\nReuseDroid migrates Android GUI tests between functionally similar apps using VLM agents that separate shared test intent from app-specific interaction flows. It introduces LinPro, a dataset of 578 migration tasks across 39 apps, to evaluate test transfer.\nPlatforms: android\nTopics: automation, evaluation, cross-app\nPublished: 2025-04-03\nAuthors: Xiaolei Li, Jialun Cao, Yepang Liu, Shing-Chi Cheung, Hailong Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The platform or GUI evaluation setting was additionally checked in the primary full text.\n\nPrimary source: https://arxiv.org/abs/2504.02357v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-02357/"}
{"id": "arxiv:2504.04716", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ededfa864150f1b7b4ba53b265b5cbaa60bcf42690a8e22f5c13f915b6cd0108", "submitted_at": "2025-04-07T03:58:45Z", "updated_at": "2025-04-07T03:58:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.04716v1"}, "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "canonical_source_url": "https://arxiv.org/abs/2504.04716", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-04716/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ededfa864150f1b7b4ba53b265b5cbaa60bcf42690a8e22f5c13f915b6cd0108", "submitted_at": "2025-04-07T03:58:45Z", "updated_at": "2025-04-07T03:58:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.04716v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.04716"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.04716"}], "name": "On the Robustness of GUI Grounding Models Against Image Attacks", "publication": null, "published_at": "2025-04-07", "record_sha256": "dd2906c7dac54d780aecb4502c2887d7706ece69970442b5d1f35a4a58818f0b", "relations": [], "resource_id": "arxiv:2504.04716", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.04716v1", "topics": ["grounding", "robustness", "safety-security", "evaluation"], "updated_at": "2025-04-07"}, "text": "On the Robustness of GUI Grounding Models Against Image Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates GUI grounding models under natural image noise, untargeted adversarial perturbations, and targeted image attacks. Experiments cover mobile, desktop, and web interfaces and analyze sensitivity to image resolution and adversarial changes.\nPlatforms: web, desktop, cross-platform\nTopics: grounding, robustness, safety-security, evaluation\nPublished: 2025-04-07\nAuthors: Haoren Zhao, Tianyi Chen, Zhen Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.04716v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-04716/"}
{"id": "arxiv:2504.07079", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f0489433a41b49ab8ab42eea57503685ba60beadabea7822134e96d865a7435", "submitted_at": "2025-04-09T17:51:50Z", "updated_at": "2025-04-09T17:51:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07079v1"}, "authors": ["Boyuan Zheng", "Michael Y. Fatemi", "Xiaolong Jin", "Zora Zhiruo Wang", "Apurva Gandhi", "Yueqi Song", "Yu Gu", "Jayanth Srinivasa", "Gaowen Liu", "Graham Neubig", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2504.07079", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-07079/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f0489433a41b49ab8ab42eea57503685ba60beadabea7822134e96d865a7435", "submitted_at": "2025-04-09T17:51:50Z", "updated_at": "2025-04-09T17:51:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07079v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.07079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.07079"}], "name": "SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skills", "publication": null, "published_at": "2025-04-09", "record_sha256": "e991ba4becc579ef033a92f3fc857ffa243337eeb429d850269b3a495b562724", "relations": [], "resource_id": "arxiv:2504.07079", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.07079v1", "topics": ["web-navigation", "tool-integration", "planning"], "updated_at": "2025-04-09"}, "text": "SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skills\nResource type: paper\nReview status: curated\nScope: core\n\nSkillWeaver explores websites, practices interactions, and converts the resulting experience into a reusable library of callable skill APIs. Its WebArena and live-website experiments examine skill refinement, composition, and transfer from stronger agents to weaker ones.\nPlatforms: web\nTopics: web-navigation, tool-integration, planning\nPublished: 2025-04-09\nAuthors: Boyuan Zheng, Michael Y. Fatemi, Xiaolong Jin, Zora Zhiruo Wang, Apurva Gandhi, Yueqi Song, Yu Gu, Jayanth Srinivasa, Gaowen Liu, Graham Neubig, Yu Su\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.07079v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-07079/"}
{"id": "arxiv:2504.07981", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "679bfcac736544e9a241c579adfda0cacbc40838452ea56ad631f8086868415a", "submitted_at": "2025-04-04T14:25:17Z", "updated_at": "2025-04-04T14:25:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07981v1"}, "authors": ["Kaixin Li", "Ziyang Meng", "Hongzhan Lin", "Ziyang Luo", "Yuchen Tian", "Jing Ma", "Zhiyong Huang", "Tat-Seng Chua"], "canonical_source_url": "https://arxiv.org/abs/2504.07981", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-07981/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "679bfcac736544e9a241c579adfda0cacbc40838452ea56ad631f8086868415a", "submitted_at": "2025-04-04T14:25:17Z", "updated_at": "2025-04-04T14:25:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.07981v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.07981"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.07981"}], "name": "ScreenSpot-Pro: GUI Grounding for Professional High-Resolution Computer Use", "publication": null, "published_at": "2025-04-04", "record_sha256": "513393931f48b096a9448e0bc09584b79bc2b4bf0bbdb6cd421baee3fc8d38d0", "relations": [], "resource_id": "arxiv:2504.07981", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.07981v1", "topics": ["grounding", "evaluation"], "updated_at": "2025-04-04"}, "text": "ScreenSpot-Pro: GUI Grounding for Professional High-Resolution Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates GUI target localization in high-resolution professional software and studies cascaded visual search to narrow the target region.\nPlatforms: desktop\nTopics: grounding, evaluation\nPublished: 2025-04-04\nAuthors: Kaixin Li, Ziyang Meng, Hongzhan Lin, Ziyang Luo, Yuchen Tian, Jing Ma, Zhiyong Huang, Tat-Seng Chua\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.07981v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-07981/"}
{"id": "arxiv:2504.08942", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53cba694577319ac1992c853be79735d654d9d32c12edcc543e6738ce010050d", "submitted_at": "2025-04-11T19:49:22Z", "updated_at": "2025-10-06T15:46:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.08942v2"}, "authors": ["Xing Han Lù", "Amirhossein Kazemnejad", "Nicholas Meade", "Arkil Patel", "Dongchan Shin", "Alejandra Zambrano", "Karolina Stańczak", "Peter Shaw", "Christopher J. Pal", "Siva Reddy"], "canonical_source_url": "https://arxiv.org/abs/2504.08942", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-08942/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53cba694577319ac1992c853be79735d654d9d32c12edcc543e6738ce010050d", "submitted_at": "2025-04-11T19:49:22Z", "updated_at": "2025-10-06T15:46:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.08942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.08942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.08942"}], "name": "AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories", "publication": null, "published_at": "2025-04-11", "record_sha256": "b80f457cfcac301287ec496363a8773d44269609ab1908d301d1ea6eff06ff2b", "relations": [], "resource_id": "arxiv:2504.08942", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.08942v2", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-10-06"}, "text": "AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories\nResource type: paper\nReview status: curated\nScope: core\n\nAgentRewardBench evaluates automated web-agent judges against expert reviews of 1,302 trajectories from five benchmarks. The annotations cover completion, side effects, and repetition, enabling comparison of LLM judges and rule-based task evaluators.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2025-04-11\nAuthors: Xing Han Lù, Amirhossein Kazemnejad, Nicholas Meade, Arkil Patel, Dongchan Shin, Alejandra Zambrano, Karolina Stańczak, Peter Shaw, Christopher J. Pal, Siva Reddy\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.08942v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-08942/"}
{"id": "arxiv:2504.09352", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf23700bf7626733b0016b773852455c503617adc81025fc9e097879b5534cca", "submitted_at": "2025-04-12T22:02:29Z", "updated_at": "2025-04-12T22:02:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.09352v1"}, "authors": ["Iason Chaimalas", "Arnas Vyšniauskas", "Gabriel Brostow"], "canonical_source_url": "https://arxiv.org/abs/2504.09352", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-09352/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf23700bf7626733b0016b773852455c503617adc81025fc9e097879b5534cca", "submitted_at": "2025-04-12T22:02:29Z", "updated_at": "2025-04-12T22:02:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.09352v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:42Z", "role": "primary", "sha256": "d1c1bc3fee56d124427abd2aabbcf876418633a8adf7a03f4ef0317cba6a42ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2008.05132%2C2010.04295%2C2101.04893%2C2112.05692%2C2306.01242%2C2311.03079%2C2401.05459%2C2402.03746%2C2402.15506%2C2403.20329%2C2404.06395%2C2404.10179%2C2405.10292%2C2406.09187%2C2406.14596%2C2408.03326%2C2409.12089%2C2501.10893%2C2504.09352&max_results=19"}], "identifiers": {"arxiv": "2504.09352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:42:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.09352"}], "name": "Explorer: Robust Collection of Interactable GUI Elements", "publication": null, "published_at": "2025-04-12", "record_sha256": "7d2904c34ceb6bf2f9051f43998a98bd99f6ac587af42ea70b652aab6b059ef4", "relations": [], "resource_id": "arxiv:2504.09352", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.09352v1", "topics": ["data-generation", "recording", "grounding", "planning"], "updated_at": "2025-04-12"}, "text": "Explorer: Robust Collection of Interactable GUI Elements\nResource type: paper\nReview status: curated\nScope: core\n\nExplorer collects interactable GUI elements from a live target application and records user sessions for application-specific model training. It links repeated interface states into a navigation map that supports path planning and voice-directed interaction.\nPlatforms: android, web\nTopics: data-generation, recording, grounding, planning\nPublished: 2025-04-12\nAuthors: Iason Chaimalas, Arnas Vyšniauskas, Gabriel Brostow\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.09352v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-09352/"}
{"id": "arxiv:2504.10127", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171517999164a3581f71cf363ec6eed357b28ca86908cb97e8498c6958762a5f", "submitted_at": "2025-04-14T11:35:02Z", "updated_at": "2025-04-15T17:13:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10127v2"}, "authors": ["Junlei Zhang", "Zichen Ding", "Chang Ma", "Zijie Chen", "Qiushi Sun", "Zhenzhong Lan", "Junxian He"], "canonical_source_url": "https://arxiv.org/abs/2504.10127", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-10127/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171517999164a3581f71cf363ec6eed357b28ca86908cb97e8498c6958762a5f", "submitted_at": "2025-04-14T11:35:02Z", "updated_at": "2025-04-15T17:13:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10127v2"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.10127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10127"}], "name": "Breaking the Data Barrier -- Building GUI Agents Through Task Generalization", "publication": null, "published_at": "2025-04-14", "record_sha256": "07ddea79984076adadabc3f52aa5344951819d252230cd6379d35e6118e91703", "relations": [], "resource_id": "arxiv:2504.10127", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.10127v2", "topics": ["data-generation", "planning", "evaluation"], "updated_at": "2025-04-15"}, "text": "Breaking the Data Barrier -- Building GUI Agents Through Task Generalization\nResource type: paper\nReview status: curated\nScope: core\n\nStudies whether mid-training on GUI perception, multimodal reasoning, and text reasoning transfers to GUI-agent planning. Experiments compare 11 task sources and curated mixtures using WebArena and AndroidWorld to investigate training-data choices when GUI trajectories are scarce.\nPlatforms: web, android\nTopics: data-generation, planning, evaluation\nPublished: 2025-04-14\nAuthors: Junlei Zhang, Zichen Ding, Chang Ma, Zijie Chen, Qiushi Sun, Zhenzhong Lan, Junxian He\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.10127v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-10127/"}
{"id": "arxiv:2504.10445", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268adf347e1f224f434e6dba838873c99a4050686cae1495643d9638d840d809", "submitted_at": "2025-04-14T17:36:46Z", "updated_at": "2025-12-01T15:22:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10445v2"}, "authors": ["Suyu Ye", "Haojun Shi", "Darren Shih", "Hyokun Yun", "Tanya Roosta", "Tianmin Shu"], "canonical_source_url": "https://arxiv.org/abs/2504.10445", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-10445/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268adf347e1f224f434e6dba838873c99a4050686cae1495643d9638d840d809", "submitted_at": "2025-04-14T17:36:46Z", "updated_at": "2025-12-01T15:22:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.10445v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2504.10445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10445"}], "name": "RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users", "publication": null, "published_at": "2025-04-14", "record_sha256": "4a430f0fc9a84ee0b5ccc22b2930ede1356aab333b31149073d5e3d960339289", "relations": [], "resource_id": "arxiv:2504.10445", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.10445v2", "topics": ["evaluation", "long-horizon", "human-in-the-loop", "grounding"], "updated_at": "2025-12-01"}, "text": "RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users\nResource type: paper\nReview status: curated\nScope: core\n\nRealWebAssist evaluates web assistance through sequences of instructions collected from real users across multiple websites. It tests intent interpretation, user-specific routines, long-term context, and visual grounding when instructions are ambiguous or evolve during interaction.\nPlatforms: web\nTopics: evaluation, long-horizon, human-in-the-loop, grounding\nPublished: 2025-04-14\nAuthors: Suyu Ye, Haojun Shi, Darren Shih, Hyokun Yun, Tanya Roosta, Tianmin Shu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.10445v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-10445/"}
{"id": "arxiv:2504.10458", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b875617b795e0f8d7d6348e22ec85ccc9ab4dd931e66850182348927b1e5209f", "submitted_at": "2025-04-14T17:45:54Z", "updated_at": "2025-10-01T04:55:20Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.10458v4"}, "authors": ["Run Luo", "Lu Wang", "Wanwei He", "Longze Chen", "Jiaming Li", "Xiaobo Xia"], "canonical_source_url": "https://arxiv.org/abs/2504.10458", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-10458/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b875617b795e0f8d7d6348e22ec85ccc9ab4dd931e66850182348927b1e5209f", "submitted_at": "2025-04-14T17:45:54Z", "updated_at": "2025-10-01T04:55:20Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.10458v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.10458"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.10458"}], "name": "GUI-R1 : A Generalist R1-Style Vision-Language Action Model For GUI Agents", "publication": null, "published_at": "2025-04-14", "record_sha256": "9f6c2f66a1ba13cb9358fd604ef4dce55eff827feea7a48128e3f4983a369e79", "relations": [], "resource_id": "arxiv:2504.10458", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.10458v4", "topics": ["reinforcement-learning", "grounding"], "updated_at": "2025-10-01"}, "text": "GUI-R1 : A Generalist R1-Style Vision-Language Action Model For GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nApplies reinforcement fine-tuning with unified action rules and curated cross-platform examples to vision-language GUI agents.\nPlatforms: cross-platform\nTopics: reinforcement-learning, grounding\nPublished: 2025-04-14\nAuthors: Run Luo, Lu Wang, Wanwei He, Longze Chen, Jiaming Li, Xiaobo Xia\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.10458v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-10458/"}
{"id": "arxiv:2504.11257", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "363e3c3e882fa93a19df74bbf77aa98525cf4639c8f62100c3ddc3e0a438997f", "submitted_at": "2025-04-15T14:56:21Z", "updated_at": "2025-07-30T13:57:49Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.11257v4"}, "authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2504.11257", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-11257/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "363e3c3e882fa93a19df74bbf77aa98525cf4639c8f62100c3ddc3e0a438997f", "submitted_at": "2025-04-15T14:56:21Z", "updated_at": "2025-07-30T13:57:49Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.11257v4"}, "publication": {"authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.809", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.809/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "15668-15684", "pdf_url": "https://aclanthology.org/2025.findings-acl.809.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.809/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "f7bcea2d3656ebc3ac305d65271000e411b735eb10556d23ee0d6157593a109f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.11257"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11257"}, {"rel": "publication", "url": "https://aclanthology.org/2025.findings-acl.809/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.findings-acl.809.pdf"}], "name": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "publication": {"authors": ["Xinyi Liu", "Xiaoyi Zhang", "Ziyun Zhang", "Yan Lu"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.findings-acl.809", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.findings-acl.809/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "15668-15684", "pdf_url": "https://aclanthology.org/2025.findings-acl.809.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis", "type": "conference-paper", "url": "https://aclanthology.org/2025.findings-acl.809/", "venue": "Findings of the Association for Computational Linguistics: ACL 2025", "volume": null}, "published_at": "2025-04-15", "record_sha256": "2a21253c87ee5807238158bef551098600e88d332222c7782734193def869d05", "relations": [], "resource_id": "arxiv:2504.11257", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.11257v4", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-07-30"}, "text": "UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nUI-E2I-Synth uses GPT-4o to synthesize GUI grounding instructions with varied complexity, addressing small elements, imbalanced element types, and implicit requests. The accompanying UI-I2E-Bench evaluates instruction-to-element localization across these annotation challenges.\nPlatforms: unspecified\nTopics: data-generation, grounding, evaluation\nPublished: 2025-04-15\nAuthors: Xinyi Liu, Xiaoyi Zhang, Ziyun Zhang, Yan Lu\nRecorded arXiv version: v4\n\nPublication edition: UI-E2I-Synth: Advancing GUI Grounding with Large-Scale Instruction Synthesis\nPublication authors: Xinyi Liu, Xiaoyi Zhang, Ziyun Zhang, Yan Lu\nPublication edition date (recorded precision): 2025-07\nPublication venue: Findings of the Association for Computational Linguistics: ACL 2025\nPublication publisher: Association for Computational Linguistics\nPublication pages: 15668-15684\nPublication doi: 10.18653/v1/2025.findings-acl.809\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.findings-acl.809/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.11257v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-11257/"}
{"id": "arxiv:2504.11281", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48acf3279b854d45ec076e630e5a84601879fc9ff894c5d368adb59d03bd3de6", "submitted_at": "2025-04-15T15:21:09Z", "updated_at": "2025-04-15T15:21:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11281v1"}, "authors": ["Chaoran Chen", "Zhiping Zhang", "Bingcan Guo", "Shang Ma", "Ibrahim Khalilov", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "canonical_source_url": "https://arxiv.org/abs/2504.11281", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-11281/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "48acf3279b854d45ec076e630e5a84601879fc9ff894c5d368adb59d03bd3de6", "submitted_at": "2025-04-15T15:21:09Z", "updated_at": "2025-04-15T15:21:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11281v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.11281"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11281"}], "name": "The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injections", "publication": null, "published_at": "2025-04-15", "record_sha256": "4fd1c274efe4c6e9f02407ffdd27d27c80e06688dce7f941bfcc9cacfdbe4131", "relations": [], "resource_id": "arxiv:2504.11281", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.11281v1", "topics": ["safety-security", "human-in-the-loop", "evaluation"], "updated_at": "2025-04-15"}, "text": "The Obvious Invisible Threat: LLM-Powered GUI Agents' Vulnerability to Fine-Print Injections\nResource type: paper\nReview status: curated\nScope: core\n\nCharacterizes six forms of GUI-content injection and studies their effects on GUI agents and human users using adversarial webpages. The analysis examines privacy failures and the limits of human oversight.\nPlatforms: web\nTopics: safety-security, human-in-the-loop, evaluation\nPublished: 2025-04-15\nAuthors: Chaoran Chen, Zhiping Zhang, Bingcan Guo, Shang Ma, Ibrahim Khalilov, Simret A Gebreegziabher, Yanfang Ye, Ziang Xiao, Yaxing Yao, Tianshi Li, Toby Jia-Jun Li\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.11281v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-11281/"}
{"id": "arxiv:2504.11675", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a25cb3ca680732a08cf2c461a65caa0c6a70404224bdfd919ba5bec7760ba72", "submitted_at": "2025-04-16T00:19:31Z", "updated_at": "2025-04-16T00:19:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11675v1"}, "authors": ["Biniam Fisseha Demissie", "Yan Naing Tun", "Lwin Khin Shar", "Mariano Ceccato"], "canonical_source_url": "https://arxiv.org/abs/2504.11675", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-11675/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a25cb3ca680732a08cf2c461a65caa0c6a70404224bdfd919ba5bec7760ba72", "submitted_at": "2025-04-16T00:19:31Z", "updated_at": "2025-04-16T00:19:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.11675v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2504.11675"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11675"}], "name": "VLM-Fuzz: Vision Language Model Assisted Recursive Depth-first Search Exploration for Effective UI Testing of Android Apps", "publication": null, "published_at": "2025-04-16", "record_sha256": "164d63cc2577f38360d871323559f2df244011ef87ef132a266aade683bc3305", "relations": [], "resource_id": "arxiv:2504.11675", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.11675v1", "topics": ["automation", "evaluation", "mobile-navigation"], "updated_at": "2025-04-16"}, "text": "VLM-Fuzz: Vision Language Model Assisted Recursive Depth-first Search Exploration for Effective UI Testing of Android Apps\nResource type: paper\nReview status: curated\nScope: core\n\nVLM-Fuzz combines depth-first UI exploration with Android manifest analysis, runtime widget hierarchies, and on-demand visual reasoning to generate app test inputs. It evaluates coverage on 59 benchmark apps and uses additional Google Play apps to study crash discovery.\nPlatforms: android\nTopics: automation, evaluation, mobile-navigation\nPublished: 2025-04-16\nAuthors: Biniam Fisseha Demissie, Yan Naing Tun, Lwin Khin Shar, Mariano Ceccato\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.11675v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-11675/"}
{"id": "arxiv:2504.11788", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76f84d5375ab7321512a1f4f8724c011cf180bf5a7a0d532ff3218cc9d7a6e75", "submitted_at": "2025-04-16T05:41:20Z", "updated_at": "2026-01-15T02:00:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.11788v3"}, "authors": ["Zhisong Zhang", "Tianqing Fang", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Haitao Mi", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2504.11788", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-11788/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76f84d5375ab7321512a1f4f8724c011cf180bf5a7a0d532ff3218cc9d7a6e75", "submitted_at": "2025-04-16T05:41:20Z", "updated_at": "2026-01-15T02:00:45Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.11788v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.11788"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.11788"}], "name": "WebRollback: Enhancing Web Agents with Explicit Rollback Mechanisms", "publication": null, "published_at": "2025-04-16", "record_sha256": "9152ec78924bef7ca2545a7afe50f749601ab7b22035198eabab190297352e11", "relations": [], "resource_id": "arxiv:2504.11788", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.11788v3", "topics": ["web-navigation", "planning", "robustness"], "updated_at": "2026-01-15"}, "text": "WebRollback: Enhancing Web Agents with Explicit Rollback Mechanisms\nResource type: paper\nReview status: curated\nScope: core\n\nWebRollback gives a web agent an explicit action for returning to an earlier state in its navigation trajectory. It studies this recovery mechanism on two live web-navigation benchmarks under both zero-shot and fine-tuned settings.\nPlatforms: web\nTopics: web-navigation, planning, robustness\nPublished: 2025-04-16\nAuthors: Zhisong Zhang, Tianqing Fang, Kaixin Ma, Wenhao Yu, Hongming Zhang, Haitao Mi, Dong Yu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.11788v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-11788/"}
{"id": "arxiv:2504.12679", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "346fc4db7f8fbf2dc66c32e36f5cc5ace22228d60bdd4c61b6b7ac65487a4a8c", "submitted_at": "2025-04-17T06:15:56Z", "updated_at": "2025-12-04T15:01:09Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.12679v4"}, "authors": ["Bofei Zhang", "Zirui Shang", "Zhi Gao", "Wang Zhang", "Rui Xie", "Xiaojian Ma", "Tao Yuan", "Xinxiao Wu", "Song-Chun Zhu", "Qing Li"], "canonical_source_url": "https://arxiv.org/abs/2504.12679", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-12679/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "346fc4db7f8fbf2dc66c32e36f5cc5ace22228d60bdd4c61b6b7ac65487a4a8c", "submitted_at": "2025-04-17T06:15:56Z", "updated_at": "2025-12-04T15:01:09Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2504.12679v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.12679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.12679"}], "name": "TongUI: Internet-Scale Trajectories from Multimodal Web Tutorials for Generalized GUI Agents", "publication": null, "published_at": "2025-04-17", "record_sha256": "cdd2b29f6667d3be9dae9de56a7619c40cfa8100f323013cba795a0873eae333", "relations": [], "resource_id": "arxiv:2504.12679", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.12679v4", "topics": ["data-generation", "grounding", "video"], "updated_at": "2025-12-04"}, "text": "TongUI: Internet-Scale Trajectories from Multimodal Web Tutorials for Generalized GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTongUI converts multimodal online tutorials, including videos and articles, into GUI-agent trajectories. Its GUI-Net dataset contains 143,000 trajectories spanning five operating systems and more than 200 applications, and supports fine-tuning for grounding and navigation.\nPlatforms: cross-platform\nTopics: data-generation, grounding, video\nPublished: 2025-04-17\nAuthors: Bofei Zhang, Zirui Shang, Zhi Gao, Wang Zhang, Rui Xie, Xiaojian Ma, Tao Yuan, Xinxiao Wu, Song-Chun Zhu, Qing Li\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.12679v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-12679/"}
{"id": "arxiv:2504.13805", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbe6c4969a9557546752235809d25912bff96a97001fade30102d8d2c751054", "submitted_at": "2025-04-18T17:13:34Z", "updated_at": "2025-04-18T17:13:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.13805v1"}, "authors": ["Guangyi Liu", "Pengxiang Zhao", "Liang Liu", "Zhiming Chen", "Yuxiang Chai", "Shuai Ren", "Hao Wang", "Shibo He", "Wenchao Meng"], "canonical_source_url": "https://arxiv.org/abs/2504.13805", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-13805/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbe6c4969a9557546752235809d25912bff96a97001fade30102d8d2c751054", "submitted_at": "2025-04-18T17:13:34Z", "updated_at": "2025-04-18T17:13:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.13805v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.13805"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13805"}], "name": "LearnAct: Few-Shot Mobile GUI Agent with a Unified Demonstration Benchmark", "publication": null, "published_at": "2025-04-18", "record_sha256": "2be7ce5ec7c9a6ed5ec01f682b5808f1c29ee3103c87747e5cc3b2505c2b3564", "relations": [], "resource_id": "arxiv:2504.13805", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.13805v1", "topics": ["mobile-navigation", "memory", "human-in-the-loop", "evaluation"], "updated_at": "2025-04-18"}, "text": "LearnAct: Few-Shot Mobile GUI Agent with a Unified Demonstration Benchmark\nResource type: paper\nReview status: curated\nScope: core\n\nLearnAct parses human demonstrations into reusable knowledge, retrieves relevant examples, and uses them during mobile GUI task execution. The accompanying LearnGUI benchmark includes 2,252 offline tasks and 101 online tasks for studying adaptation from demonstrations.\nPlatforms: unspecified\nTopics: mobile-navigation, memory, human-in-the-loop, evaluation\nPublished: 2025-04-18\nAuthors: Guangyi Liu, Pengxiang Zhao, Liang Liu, Zhiming Chen, Yuxiang Chai, Shuai Ren, Hao Wang, Shibo He, Wenchao Meng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.13805v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-13805/"}
{"id": "arxiv:2504.13865", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "216d0605205cfd7f17ffd7d7f4c9752577ca427f370df42faf4a2250b80bcc39", "submitted_at": "2025-03-27T17:58:31Z", "updated_at": "2025-06-04T17:29:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13865v2"}, "authors": ["Fei Tang", "Haolei Xu", "Hang Zhang", "Siqi Chen", "Xingyu Wu", "Yongliang Shen", "Wenqi Zhang", "Guiyang Hou", "Zeqi Tan", "Yuchen Yan", "Kaitao Song", "Jian Shao", "Weiming Lu", "Jun Xiao", "Yueting Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2504.13865", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-13865/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "216d0605205cfd7f17ffd7d7f4c9752577ca427f370df42faf4a2250b80bcc39", "submitted_at": "2025-03-27T17:58:31Z", "updated_at": "2025-06-04T17:29:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13865v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.13865"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13865"}], "name": "A Survey on (M)LLM-Based GUI Agents", "publication": null, "published_at": "2025-03-27", "record_sha256": "a5978df66e1a12122bca472c396dc6a3f6db52cbc9c9339313ed738d3602e5d0", "relations": [], "resource_id": "arxiv:2504.13865", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.13865v2", "topics": ["survey"], "updated_at": "2025-06-04"}, "text": "A Survey on (M)LLM-Based GUI Agents\nResource type: survey\nReview status: curated\nScope: core\n\nAnalyzes LLM-based GUI agents through four components (perception, exploration, planning, and interaction) and the methods used to evaluate them.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2025-03-27\nAuthors: Fei Tang, Haolei Xu, Hang Zhang, Siqi Chen, Xingyu Wu, Yongliang Shen, Wenqi Zhang, Guiyang Hou, Zeqi Tan, Yuchen Yan, Kaitao Song, Jian Shao, Weiming Lu, Jun Xiao, Yueting Zhuang\nRecorded arXiv version: v2\n\nPrimary source: https://arxiv.org/abs/2504.13865v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-13865/"}
{"id": "arxiv:2504.13936", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e00237ef710c37c0c9f7768d699f57e9d5dcc190e8de7e20e9ec3b21b82750", "submitted_at": "2025-04-15T14:03:10Z", "updated_at": "2025-05-20T14:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13936v2"}, "authors": ["Dezhao Luo", "Bohan Tang", "Kang Li", "Georgios Papoudakis", "Jifei Song", "Shaogang Gong", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2504.13936", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-13936/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15e00237ef710c37c0c9f7768d699f57e9d5dcc190e8de7e20e9ec3b21b82750", "submitted_at": "2025-04-15T14:03:10Z", "updated_at": "2025-05-20T14:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.13936v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:45:04Z", "role": "primary", "sha256": "bab88dfed88ca440c231aa768530374a0f19498777a70921481a904fddd46bb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2303.17760%2C2310.03720%2C2312.11190%2C2403.02502%2C2406.12639%2C2409.18060%2C2410.00201%2C2501.01702%2C2501.11425%2C2502.02533%2C2502.02955%2C2502.12110%2C2502.15760%2C2502.17110%2C2502.19328%2C2503.02197%2C2503.02403%2C2503.02682%2C2503.06580%2C2503.07365%2C2503.09572%2C2503.14021%2C2503.24180%2C2504.02357%2C2504.11675%2C2504.13936&max_results=26"}], "identifiers": {"arxiv": "2504.13936"}, "kind": "paper", "last_verified_at": "2026-10-16T06:45:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.13936"}], "name": "ViMo: A Generative Visual GUI World Model for App Agents", "publication": null, "published_at": "2025-04-15", "record_sha256": "4ed6c96e523a523a1d3d5f103615e82eb3c8d559d12677c2ceab81760f9ed1f1", "relations": [], "resource_id": "arxiv:2504.13936", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.13936v2", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2025-05-20"}, "text": "ViMo: A Generative Visual GUI World Model for App Agents\nResource type: paper\nReview status: curated\nScope: core\n\nViMo predicts the next mobile-app screenshot after a candidate action by separately modeling graphics and text. A symbolic placeholder representation preserves layout while a second predictor supplies readable text, enabling visual lookahead for GUI-agent planning.\nPlatforms: unspecified\nTopics: world-models, planning, mobile-navigation\nPublished: 2025-04-15\nAuthors: Dezhao Luo, Bohan Tang, Kang Li, Georgios Papoudakis, Jifei Song, Shaogang Gong, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.13936v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-13936/"}
{"id": "arxiv:2504.14064", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5bdf5fefa1ee31126d99de08236991df559896212f60e0b3058fcd8cc8a134b", "submitted_at": "2025-04-18T20:36:10Z", "updated_at": "2025-10-07T13:10:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.14064v3"}, "authors": ["Leo Boisvert", "Mihir Bansal", "Chandra Kiran Reddy Evuru", "Gabriel Huang", "Abhay Puri", "Avinandan Bose", "Maryam Fazel", "Quentin Cappart", "Jason Stanley", "Alexandre Lacoste", "Alexandre Drouin", "Krishnamurthy Dvijotham"], "canonical_source_url": "https://arxiv.org/abs/2504.14064", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-14064/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5bdf5fefa1ee31126d99de08236991df559896212f60e0b3058fcd8cc8a134b", "submitted_at": "2025-04-18T20:36:10Z", "updated_at": "2025-10-07T13:10:30Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.14064v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "identifiers": {"arxiv": "2504.14064"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14064"}], "name": "DoomArena: A framework for Testing AI Agents Against Evolving Security Threats", "publication": null, "published_at": "2025-04-18", "record_sha256": "b43bd758f05657da9ac8a126002114c659152209cabf55b9d6b1d78810c325f7", "relations": [], "resource_id": "arxiv:2504.14064", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.14064v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-07"}, "text": "DoomArena: A framework for Testing AI Agents Against Evolving Security Threats\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a modular security testing framework that plugs into BrowserGym and tool-use benchmarks to evaluate agents under configurable threat models.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-04-18\nAuthors: Leo Boisvert, Mihir Bansal, Chandra Kiran Reddy Evuru, Gabriel Huang, Abhay Puri, Avinandan Bose, Maryam Fazel, Quentin Cappart, Jason Stanley, Alexandre Lacoste, Alexandre Drouin, Krishnamurthy Dvijotham\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.14064v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-14064/"}
{"id": "arxiv:2504.14239", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce03185cc73c2723f444bd9729c67da9aab7d96b5ef099fb78e47f58790bc18", "submitted_at": "2025-04-19T09:25:55Z", "updated_at": "2025-04-19T09:25:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.14239v1"}, "authors": ["Yuhang Liu", "Pengxiang Li", "Congkai Xie", "Xavier Hu", "Xiaotian Han", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "canonical_source_url": "https://arxiv.org/abs/2504.14239", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-14239/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce03185cc73c2723f444bd9729c67da9aab7d96b5ef099fb78e47f58790bc18", "submitted_at": "2025-04-19T09:25:55Z", "updated_at": "2025-04-19T09:25:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.14239v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.14239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14239"}], "name": "InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners", "publication": null, "published_at": "2025-04-19", "record_sha256": "0053e3870756d3f56619eeaa9cbd359d1fb065738ce7a9c7f9578ae32c7d910b", "relations": [], "resource_id": "arxiv:2504.14239", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.14239v1", "topics": ["planning", "reinforcement-learning"], "updated_at": "2025-04-19"}, "text": "InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners\nResource type: paper\nReview status: curated\nScope: core\n\nUses a staged reasoning-focused training framework to move GUI agents from reactive actions toward deliberative planning and recovery.\nPlatforms: cross-platform\nTopics: planning, reinforcement-learning\nPublished: 2025-04-19\nAuthors: Yuhang Liu, Pengxiang Li, Congkai Xie, Xavier Hu, Xiaotian Han, Shengyu Zhang, Hongxia Yang, Fei Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.14239v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-14239/"}
{"id": "arxiv:2504.14603", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc47a80b48a6a5f126e36fa14639d56d8c8025a453fbcee2cbdb73eb67945561", "submitted_at": "2025-04-20T13:04:43Z", "updated_at": "2025-04-25T05:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.14603v2"}, "authors": ["Chaoyun Zhang", "He Huang", "Chiming Ni", "Jian Mu", "Si Qin", "Shilin He", "Lu Wang", "Fangkai Yang", "Pu Zhao", "Chao Du", "Liqun Li", "Yu Kang", "Zhao Jiang", "Suzhen Zheng", "Rujia Wang", "Jiaxu Qian", "Minghua Ma", "Jian-Guang Lou", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2504.14603", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-14603/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc47a80b48a6a5f126e36fa14639d56d8c8025a453fbcee2cbdb73eb67945561", "submitted_at": "2025-04-20T13:04:43Z", "updated_at": "2025-04-25T05:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.14603v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:07Z", "role": "primary", "sha256": "0ff3084c62a2ae0fbda31be5d409fb8ca3a6887da897a238beec93a8d7be7825", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.14603v2"}], "identifiers": {"arxiv": "2504.14603"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.14603"}], "name": "UFO2: The Desktop AgentOS", "publication": null, "published_at": "2025-04-20", "record_sha256": "dcaf9d4302222d4eb4e6742916f98dd776d69dcf4bdd22e6c936ef4c85f2a974", "relations": [], "resource_id": "arxiv:2504.14603", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.14603v2", "topics": ["automation", "cross-app", "tool-integration"], "updated_at": "2025-04-25"}, "text": "UFO2: The Desktop AgentOS\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UFO2, a Windows desktop agent architecture combining a coordinating HostAgent, application-specialized agents, and a shared GUI/API action layer. It integrates UI Automation, visual parsing, speculative planning, and an isolated picture-in-picture desktop.\nPlatforms: desktop, windows\nTopics: automation, cross-app, tool-integration\nPublished: 2025-04-20\nAuthors: Chaoyun Zhang, He Huang, Chiming Ni, Jian Mu, Si Qin, Shilin He, Lu Wang, Fangkai Yang, Pu Zhao, Chao Du, Liqun Li, Yu Kang, Zhao Jiang, Suzhen Zheng, Rujia Wang, Jiaxu Qian, Minghua Ma, Jian-Guang Lou, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.14603v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-14603/"}
{"id": "arxiv:2504.16073", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1353bb4651f449763ac58803b4858edf6e1d5b0ad3ab3e4c333883a1ebbaebf4", "submitted_at": "2025-04-22T17:52:42Z", "updated_at": "2025-04-22T17:52:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.16073v1"}, "authors": ["Zhiyuan Hu", "Shiyun Xiong", "Yifan Zhang", "See-Kiong Ng", "Anh Tuan Luu", "Bo An", "Shuicheng Yan", "Bryan Hooi"], "canonical_source_url": "https://arxiv.org/abs/2504.16073", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-16073/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1353bb4651f449763ac58803b4858edf6e1d5b0ad3ab3e4c333883a1ebbaebf4", "submitted_at": "2025-04-22T17:52:42Z", "updated_at": "2025-04-22T17:52:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2504.16073v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:10Z", "role": "primary", "sha256": "b1cf161fc20efc51363d9007952ce5f917963d4ab2ca32d39099229060b29528", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16073v1"}], "identifiers": {"arxiv": "2504.16073"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16073"}], "name": "Guiding VLM Agents with Process Rewards at Inference Time for GUI Navigation", "publication": null, "published_at": "2025-04-22", "record_sha256": "b43fcba14e5604dc2c6ccc34bbf3afb3245ce3e549405ab65a27907459dff3f8", "relations": [], "resource_id": "arxiv:2504.16073", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.16073v1", "topics": ["reward-models", "self-reflection", "grounding"], "updated_at": "2025-04-22"}, "text": "Guiding VLM Agents with Process Rewards at Inference Time for GUI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nGuides visual GUI agents at inference time with a process-reward model that evaluates intermediate actions. It combines stepwise guidance with trajectory reflection and retry mechanisms in static and interactive navigation tasks.\nPlatforms: unspecified\nTopics: reward-models, self-reflection, grounding\nPublished: 2025-04-22\nAuthors: Zhiyuan Hu, Shiyun Xiong, Yifan Zhang, See-Kiong Ng, Anh Tuan Luu, Bo An, Shuicheng Yan, Bryan Hooi\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.16073v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-16073/"}
{"id": "arxiv:2504.16419", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1ec3311094166baac4de66ad0f083b66585c933f19cb66fc580bf128ad809ca", "submitted_at": "2025-04-23T05:01:25Z", "updated_at": "2025-04-27T06:52:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.16419v2"}, "authors": ["Qi Yang", "Weichen Bi", "Haiyang Shen", "Yaoqi Guo", "Yun Ma"], "canonical_source_url": "https://arxiv.org/abs/2504.16419", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-16419/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1ec3311094166baac4de66ad0f083b66585c933f19cb66fc580bf128ad809ca", "submitted_at": "2025-04-23T05:01:25Z", "updated_at": "2025-04-27T06:52:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.16419v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:13Z", "role": "primary", "sha256": "496c0419e2c8c752fc2c88d015ad2cf13c182aa4501a8305ff65ad4178113843", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2504.16419v2"}], "identifiers": {"arxiv": "2504.16419"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.16419"}], "name": "PixelWeb: The First Web GUI Dataset with Pixel-Wise Labels", "publication": null, "published_at": "2025-04-23", "record_sha256": "147a376e074f8b6c96fa5b142509f390c1bde77c2e35a918e00fb461ca43f794", "relations": [], "resource_id": "arxiv:2504.16419", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.16419v2", "topics": ["data-generation", "grounding", "screen-understanding"], "updated_at": "2025-04-27"}, "text": "PixelWeb: The First Web GUI Dataset with Pixel-Wise Labels\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PixelWeb, a webpage GUI dataset with element images, contours, and pixel masks. Its annotation pipeline combines bitmap channel information with DOM-based visibility and stacking analysis to address occlusion and overlapping elements.\nPlatforms: web\nTopics: data-generation, grounding, screen-understanding\nPublished: 2025-04-23\nAuthors: Qi Yang, Weichen Bi, Haiyang Shen, Yaoqi Guo, Yun Ma\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.16419v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-16419/"}
{"id": "arxiv:2504.17934", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a55e347bc10e5571464924d9301ff2c32d5e9f1f990ca91725ec8c9f6da594", "submitted_at": "2025-04-24T20:51:20Z", "updated_at": "2025-06-05T12:40:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.17934v2"}, "authors": ["Chaoran Chen", "Zhiping Zhang", "Ibrahim Khalilov", "Bingcan Guo", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "canonical_source_url": "https://arxiv.org/abs/2504.17934", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-17934/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a55e347bc10e5571464924d9301ff2c32d5e9f1f990ca91725ec8c9f6da594", "submitted_at": "2025-04-24T20:51:20Z", "updated_at": "2025-06-05T12:40:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.17934v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.17934"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.17934"}], "name": "Toward a Human-Centered Evaluation Framework for Trustworthy LLM-Powered GUI Agents", "publication": null, "published_at": "2025-04-24", "record_sha256": "15e0e1829b1618c358bb98f74c0c4de3a09ce36a94fabc54490ee0748727f733", "relations": [], "resource_id": "arxiv:2504.17934", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.17934v2", "topics": ["evaluation", "safety-security", "human-in-the-loop"], "updated_at": "2025-06-05"}, "text": "Toward a Human-Centered Evaluation Framework for Trustworthy LLM-Powered GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a position paper on privacy and security evaluation for GUI agents, identifying risk categories and challenges in involving human evaluators. It proposes risk-aware assessment and in-context consent as design directions.\nPlatforms: unspecified\nTopics: evaluation, safety-security, human-in-the-loop\nPublished: 2025-04-24\nAuthors: Chaoran Chen, Zhiping Zhang, Ibrahim Khalilov, Bingcan Guo, Simret A Gebreegziabher, Yanfang Ye, Ziang Xiao, Yaxing Yao, Tianshi Li, Toby Jia-Jun Li\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. This is a position paper proposing an evaluation framework; the abstract does not report a new benchmark or empirical validation of the proposed framework.\n\nPrimary source: https://arxiv.org/abs/2504.17934v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-17934/"}
{"id": "arxiv:2504.18575", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a104d9076797555795766749f6b4eeec9428d0efb9340f7e55cffecd5d727b0", "submitted_at": "2025-04-22T17:51:03Z", "updated_at": "2025-05-16T22:42:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.18575v3"}, "authors": ["Ivan Evtimov", "Arman Zharmagambetov", "Aaron Grattafiori", "Chuan Guo", "Kamalika Chaudhuri"], "canonical_source_url": "https://arxiv.org/abs/2504.18575", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-18575/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a104d9076797555795766749f6b4eeec9428d0efb9340f7e55cffecd5d727b0", "submitted_at": "2025-04-22T17:51:03Z", "updated_at": "2025-05-16T22:42:29Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.18575v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.18575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.18575"}], "name": "WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks", "publication": null, "published_at": "2025-04-22", "record_sha256": "54f488b38d11873eee6ff9a19ae6dea91ce72fd5c5b645508366960d12073428", "relations": [], "resource_id": "arxiv:2504.18575", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.18575v3", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-05-16"}, "text": "WASP: Benchmarking Web Agent Security Against Prompt Injection Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WASP, an end-to-end benchmark for prompt-injection security in web agents. Its evaluation separates partial attack progress from completion of the attacker’s full objective.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2025-04-22\nAuthors: Ivan Evtimov, Arman Zharmagambetov, Aaron Grattafiori, Chuan Guo, Kamalika Chaudhuri\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.18575v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-18575/"}
{"id": "arxiv:2504.19838", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9103197fa1274fac79ae9e4472f5fbdc9a210548afb024f1d4259c6bb0131bf2", "submitted_at": "2025-04-28T14:39:25Z", "updated_at": "2025-11-17T15:51:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.19838v3"}, "authors": ["Guangyi Liu", "Pengxiang Zhao", "Yaozhen Liang", "Liang Liu", "Yaxuan Guo", "Han Xiao", "Weifeng Lin", "Yuxiang Chai", "Yue Han", "Shuai Ren", "Hao Wang", "Xiaoyu Liang", "WenHao Wang", "Tianze Wu", "Zhengxi Lu", "Siheng Chen", "LiLinghao", "Hao Wang", "Guanjing Xiong", "Yong Liu", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2504.19838", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-19838/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9103197fa1274fac79ae9e4472f5fbdc9a210548afb024f1d4259c6bb0131bf2", "submitted_at": "2025-04-28T14:39:25Z", "updated_at": "2025-11-17T15:51:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2504.19838v3"}}, "environments": ["android", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.19838"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.19838"}], "name": "LLM-Powered GUI Agents in Phone Automation: Surveying Progress and Prospects", "publication": null, "published_at": "2025-04-28", "record_sha256": "8493b73746a84ffac7edc2ecd0fed49f0f5188bb7f17c8929ab8429064673cd6", "relations": [], "resource_id": "arxiv:2504.19838", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.19838v3", "topics": ["survey"], "updated_at": "2025-11-17"}, "text": "LLM-Powered GUI Agents in Phone Automation: Surveying Progress and Prospects\nResource type: survey\nReview status: curated\nScope: core\n\nReviews LLM-driven phone GUI agents with a taxonomy of agent frameworks, prompt- and training-based modeling, datasets, and benchmarks.\nPlatforms: android, ios\nTopics: survey\nPublished: 2025-04-28\nAuthors: Guangyi Liu, Pengxiang Zhao, Yaozhen Liang, Liang Liu, Yaxuan Guo, Han Xiao, Weifeng Lin, Yuxiang Chai, Yue Han, Shuai Ren, Hao Wang, Xiaoyu Liang, WenHao Wang, Tianze Wu, Zhengxi Lu, Siheng Chen, LiLinghao, Hao Wang, Guanjing Xiong, Yong Liu, Hongsheng Li\nRecorded arXiv version: v3\n\nPrimary source: https://arxiv.org/abs/2504.19838v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-19838/"}
{"id": "arxiv:2504.20464", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5040073e7c278734df219d71d94e236468468621f5971ef3ccbc97b94e1792b9", "submitted_at": "2025-04-29T06:55:15Z", "updated_at": "2025-05-13T01:59:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.20464v2"}, "authors": ["Jiahao Li", "Kaer Huang"], "canonical_source_url": "https://arxiv.org/abs/2504.20464", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-20464/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5040073e7c278734df219d71d94e236468468621f5971ef3ccbc97b94e1792b9", "submitted_at": "2025-04-29T06:55:15Z", "updated_at": "2025-05-13T01:59:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.20464v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.20464"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.20464"}], "name": "A Survey on GUI Agents with Foundation Models Enhanced by Reinforcement Learning", "publication": null, "published_at": "2025-04-29", "record_sha256": "63ee18ff335163a3a5472b403564910e98b125b93cb017094baad0d8c1256196", "relations": [], "resource_id": "arxiv:2504.20464", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.20464v2", "topics": ["survey", "reinforcement-learning"], "updated_at": "2025-05-13"}, "text": "A Survey on GUI Agents with Foundation Models Enhanced by Reinforcement Learning\nResource type: survey\nReview status: curated\nScope: core\n\nFormalizes GUI tasks as Markov decision processes and traces agent training from prompting through supervised fine-tuning to reinforcement learning.\nPlatforms: cross-platform\nTopics: survey, reinforcement-learning\nPublished: 2025-04-29\nAuthors: Jiahao Li, Kaer Huang\nRecorded arXiv version: v2\n\nPrimary source: https://arxiv.org/abs/2504.20464v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-20464/"}
{"id": "arxiv:2504.21024", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86ca58c92ffc2cbbcb0297d8c7618dfab966e36a07b141c068998b0f2934d11e", "submitted_at": "2025-04-23T02:54:31Z", "updated_at": "2025-08-21T05:55:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21024v2"}, "authors": ["Tianqing Fang", "Hongming Zhang", "Zhisong Zhang", "Kaixin Ma", "Wenhao Yu", "Haitao Mi", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2504.21024", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-21024/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86ca58c92ffc2cbbcb0297d8c7618dfab966e36a07b141c068998b0f2934d11e", "submitted_at": "2025-04-23T02:54:31Z", "updated_at": "2025-08-21T05:55:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21024v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2504.21024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.21024"}], "name": "WebEvolver: Enhancing Web Agent Self-Improvement with Coevolving World Model", "publication": null, "published_at": "2025-04-23", "record_sha256": "3aa2f0d83080ded691056ac26c03cdc8fa21b6da6897f3b87617967399a0799e", "relations": [], "resource_id": "arxiv:2504.21024", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2504.21024v2", "topics": ["world-models", "data-generation", "web-navigation", "planning"], "updated_at": "2025-08-21"}, "text": "WebEvolver: Enhancing Web Agent Self-Improvement with Coevolving World Model\nResource type: paper\nReview status: curated\nScope: core\n\nWebEvolver trains a co-evolving world model to predict web observations and generate additional agent-training trajectories. The same model supports look-ahead action selection at inference, with evaluation on live web-task benchmarks.\nPlatforms: web\nTopics: world-models, data-generation, web-navigation, planning\nPublished: 2025-04-23\nAuthors: Tianqing Fang, Hongming Zhang, Zhisong Zhang, Kaixin Ma, Wenhao Yu, Haitao Mi, Dong Yu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.21024v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-21024/"}
{"id": "arxiv:2504.21776", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "351b6d7fbe4e217a5211884832c6ad07d76a0463e6a8c996815999ff3a5f1921", "submitted_at": "2025-04-30T16:25:25Z", "updated_at": "2025-10-13T12:40:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21776v2"}, "authors": ["Xiaoxi Li", "Jiajie Jin", "Guanting Dong", "Hongjin Qian", "Yongkang Wu", "Ji-Rong Wen", "Yutao Zhu", "Zhicheng Dou"], "canonical_source_url": "https://arxiv.org/abs/2504.21776", "citation_url": "https://awesomecua.github.io/resources/arxiv-2504-21776/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "351b6d7fbe4e217a5211884832c6ad07d76a0463e6a8c996815999ff3a5f1921", "submitted_at": "2025-04-30T16:25:25Z", "updated_at": "2025-10-13T12:40:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2504.21776v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:06.851520+00:00", "role": "primary", "sha256": "d06baa20c7775347cbb4492473418d11cbe56941ab90baa16381353ab6c39e07", "supports": ["summary", "scope", "notes"], "url": "https://arxiv.org/html/2504.21776v2"}], "identifiers": {"arxiv": "2504.21776"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2504.21776"}], "name": "WebThinker: Empowering Large Reasoning Models with Deep Research Capability", "publication": null, "published_at": "2025-04-30", "record_sha256": "80a2bd43372a888f77bf92b559a5f6598c47854fecdb46eb8105db571af04197", "relations": [], "resource_id": "arxiv:2504.21776", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2504.21776v2", "topics": ["web-navigation", "planning", "tool-integration"], "updated_at": "2025-10-13"}, "text": "WebThinker: Empowering Large Reasoning Models with Deep Research Capability\nResource type: paper\nReview status: curated\nScope: adjacent\n\nCombines iterative web search and link-following with interleaved reasoning and report drafting for deep research. Its reported web explorer retrieves page content through a crawler rather than controlling a visual browser interface.\nPlatforms: web\nTopics: web-navigation, planning, tool-integration\nPublished: 2025-04-30\nAuthors: Xiaoxi Li, Jiajie Jin, Guanting Dong, Hongjin Qian, Yongkang Wu, Ji-Rong Wen, Yutao Zhu, Zhicheng Dou\nRecorded arXiv version: v2\n\nAdjacent web-research method. The retained abstract and full-text Section 3 and Appendix A describe search APIs and Crawl4AI URL retrieval; no visual GUI execution is inferred. Results were not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2504.21776v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2504-21776/"}
{"id": "arxiv:2505.00416", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e582c92521ec41a0a3daa0ca72b8c00250621289352b28400ee6d311c4a106eb", "submitted_at": "2025-05-01T09:27:13Z", "updated_at": "2025-05-01T09:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.00416v1"}, "authors": ["Jing Huang", "Zhixiong Zeng", "Wenkang Han", "Yufeng Zhong", "Liming Zheng", "Shuai Fu", "Jingyuan Chen", "Lin Ma"], "canonical_source_url": "https://arxiv.org/abs/2505.00416", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-00416/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e582c92521ec41a0a3daa0ca72b8c00250621289352b28400ee6d311c4a106eb", "submitted_at": "2025-05-01T09:27:13Z", "updated_at": "2025-05-01T09:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.00416v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.00416"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00416"}], "name": "ScaleTrack: Scaling and back-tracking Automated GUI Agents", "publication": null, "published_at": "2025-05-01", "record_sha256": "4e8f0e887d727105b69cef3b7b0de10bf93ffaadb1fdb80d4fac838624a4088e", "relations": [], "resource_id": "arxiv:2505.00416", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.00416v1", "topics": ["grounding", "planning", "data-generation", "interaction-history"], "updated_at": "2025-05-01"}, "text": "ScaleTrack: Scaling and back-tracking Automated GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nScaleTrack combines a unified collection of GUI grounding samples with a planning objective that predicts the next action while reconstructing prior actions. This backward prediction links the current screenshot to interaction history during agent training.\nPlatforms: unspecified\nTopics: grounding, planning, data-generation, interaction-history\nPublished: 2025-05-01\nAuthors: Jing Huang, Zhixiong Zeng, Wenkang Han, Yufeng Zhong, Liming Zheng, Shuai Fu, Jingyuan Chen, Lin Ma\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.00416v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-00416/"}
{"id": "arxiv:2505.00684", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2dc8ca397adce1ecf6cf53afa82a8d4f1825bf00277aeb117966f965abfe48", "submitted_at": "2025-05-01T17:45:59Z", "updated_at": "2025-07-14T17:45:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.00684v2"}, "authors": ["Tiange Luo", "Lajanugen Logeswaran", "Justin Johnson", "Honglak Lee"], "canonical_source_url": "https://arxiv.org/abs/2505.00684", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-00684/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2dc8ca397adce1ecf6cf53afa82a8d4f1825bf00277aeb117966f965abfe48", "submitted_at": "2025-05-01T17:45:59Z", "updated_at": "2025-07-14T17:45:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.00684v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.00684"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.00684"}], "name": "Visual Test-time Scaling for GUI Agent Grounding", "publication": null, "published_at": "2025-05-01", "record_sha256": "fe562f421fe7d886b5621eaf41ebf85574d6d5ef1404a7ab413a0b26d6dcb1a4", "relations": [], "resource_id": "arxiv:2505.00684", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.00684v2", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-14"}, "text": "Visual Test-time Scaling for GUI Agent Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nRegionFocus adds visual test-time scaling by zooming into selected screenshot regions and maintaining an image-based map of landmarks and actions. It evaluates this grounding procedure with UI-TARS and Qwen2.5-VL on ScreenSpot-Pro and WebVoyager.\nPlatforms: web\nTopics: grounding, screen-understanding, web-navigation\nPublished: 2025-05-01\nAuthors: Tiange Luo, Lajanugen Logeswaran, Justin Johnson, Honglak Lee\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.00684v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-00684/"}
{"id": "arxiv:2505.03570", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8bb4c6f8080ed1ee91e1089233cae6f231ab3256b38ac7d6403ebea3ae56fe0", "submitted_at": "2025-05-06T14:29:47Z", "updated_at": "2025-05-06T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.03570v1"}, "authors": ["Mariya Davydova", "Daniel Jeffries", "Patrick Barker", "Arturo Márquez Flores", "Sinéad Ryan"], "canonical_source_url": "https://arxiv.org/abs/2505.03570", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-03570/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8bb4c6f8080ed1ee91e1089233cae6f231ab3256b38ac7d6403ebea3ae56fe0", "submitted_at": "2025-05-06T14:29:47Z", "updated_at": "2025-05-06T14:29:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.03570v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "39de8f4a9aee74cc0eae64e24efc08ed070714eaaea4a30dbc20cb08397eb9bd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.03570v1"}], "identifiers": {"arxiv": "2505.03570"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.03570"}], "name": "OSUniverse: Benchmark for Multimodal GUI-navigation AI Agents", "publication": null, "published_at": "2025-05-06", "record_sha256": "8844947ec98f0c792e008cc0b8e6e3d96ab8352daa1c923f7fe9d3f01ae7b0ab", "relations": [], "resource_id": "arxiv:2505.03570", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.03570v1", "topics": ["evaluation", "cross-app", "long-horizon"], "updated_at": "2025-05-06"}, "text": "OSUniverse: Benchmark for Multimodal GUI-navigation AI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OSUniverse, a desktop GUI benchmark spanning precise clicks through multistep and multiapplication tasks. It provides both manual scoring and an automated validation mechanism.\nPlatforms: desktop\nTopics: evaluation, cross-app, long-horizon\nPublished: 2025-05-06\nAuthors: Mariya Davydova, Daniel Jeffries, Patrick Barker, Arturo Márquez Flores, Sinéad Ryan\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.03570v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-03570/"}
{"id": "arxiv:2505.05026", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f65cd2856c554584baab9b1cbf9c5250b5c01d4f31980d69aefab727427ded2", "submitted_at": "2025-05-08T08:00:32Z", "updated_at": "2026-06-04T06:58:13Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.05026v5"}, "authors": ["Jaehyun Jeon", "Min Soo Kim", "Jang Han Yoon", "Sumin Shim", "Yejin Choi", "Hanbin Kim", "Dae Hyun Kim", "Youngjae Yu"], "canonical_source_url": "https://arxiv.org/abs/2505.05026", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-05026/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f65cd2856c554584baab9b1cbf9c5250b5c01d4f31980d69aefab727427ded2", "submitted_at": "2025-05-08T08:00:32Z", "updated_at": "2026-06-04T06:58:13Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.05026v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:21Z", "role": "primary", "sha256": "1220bb1498a7f9c68cff0fbbd8eee209dfed138e771dfd6eb946fc0639c49cd1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.05026v5"}], "identifiers": {"arxiv": "2505.05026"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.05026"}], "name": "Do MLLMs Capture How Interfaces Guide User Behavior? A Benchmark for Multimodal UI/UX Design Understanding", "publication": null, "published_at": "2025-05-08", "record_sha256": "4bab6f32a89be93d6fc280d8a8dcbb6e4438273a09f3216f9e2329916236601f", "relations": [], "resource_id": "arxiv:2505.05026", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2505.05026v5", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": "2026-06-04"}, "text": "Do MLLMs Capture How Interfaces Guide User Behavior? A Benchmark for Multimodal UI/UX Design Understanding\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces WiserUI-Bench, pairing real UI variants from industry A/B tests with observed behavioral winners and expert explanations. It evaluates models’ ability to predict the more effective interface and explain its effect on user behavior.\nPlatforms: unspecified\nTopics: screen-understanding, ui-design, evaluation\nPublished: 2025-05-08\nAuthors: Jaehyun Jeon, Min Soo Kim, Jang Han Yoon, Sumin Shim, Yejin Choi, Hanbin Kim, Dae Hyun Kim, Youngjae Yu\nRecorded arXiv version: v5\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced. This benchmark measures UI/UX design understanding from static image pairs; it does not measure autonomous GUI action execution.\n\nPrimary source: https://arxiv.org/abs/2505.05026v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-05026/"}
{"id": "arxiv:2505.05440", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f5fdf474c1cb0d7e9f42d617de27afae38a9fd8f5e0b07d1c86e082dc872a35", "submitted_at": "2025-05-08T17:31:20Z", "updated_at": "2025-11-17T09:12:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.05440v3"}, "authors": ["Biao Yi", "Xavier Hu", "Yurun Chen", "Shengyu Zhang", "Hongxia Yang", "Fan Wu"], "canonical_source_url": "https://arxiv.org/abs/2505.05440", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-05440/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f5fdf474c1cb0d7e9f42d617de27afae38a9fd8f5e0b07d1c86e082dc872a35", "submitted_at": "2025-05-08T17:31:20Z", "updated_at": "2025-11-17T09:12:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.05440v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2505.05440"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.05440"}], "name": "EcoAgent: An Efficient Device-Cloud Collaborative Multi-Agent Framework for Mobile Automation", "publication": null, "published_at": "2025-05-08", "record_sha256": "df62d26b488027554eb843dc58db948d0e4aed1a9b6def9e6e7feda9c5e56436", "relations": [], "resource_id": "arxiv:2505.05440", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.05440v3", "topics": ["mobile-navigation", "planning", "efficiency"], "updated_at": "2025-11-17"}, "text": "EcoAgent: An Efficient Device-Cloud Collaborative Multi-Agent Framework for Mobile Automation\nResource type: paper\nReview status: curated\nScope: core\n\nEcoAgent splits mobile automation between cloud planning and on-device observation and verification. Compact screen descriptions and feedback reduce communication costs, with evaluation on AndroidWorld.\nPlatforms: android\nTopics: mobile-navigation, planning, efficiency\nPublished: 2025-05-08\nAuthors: Biao Yi, Xavier Hu, Yurun Chen, Shengyu Zhang, Hongxia Yang, Fan Wu\nRecorded arXiv version: v3\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.05440v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-05440/"}
{"id": "arxiv:2505.09875", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63f4a505645fec16eac6b59923a6c1979c53376f19c6a4ea53c9a5d32a45e3d4", "submitted_at": "2025-05-15T00:42:51Z", "updated_at": "2025-05-16T07:01:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.09875v2"}, "authors": ["Shuning Zhang", "Jingruo Chen", "Zhiqi Gao", "Jiajing Gao", "Xin Yi", "Hewu Li"], "canonical_source_url": "https://arxiv.org/abs/2505.09875", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-09875/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63f4a505645fec16eac6b59923a6c1979c53376f19c6a4ea53c9a5d32a45e3d4", "submitted_at": "2025-05-15T00:42:51Z", "updated_at": "2025-05-16T07:01:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.09875v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.09875"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.09875"}], "name": "Characterizing Unintended Consequences in Human-GUI Agent Collaboration for Web Browsing", "publication": null, "published_at": "2025-05-15", "record_sha256": "41627931777c8bd9d277e9f9fa8e07cceff688c8ecdae3246c3a5905d307ecdb", "relations": [], "resource_id": "arxiv:2505.09875", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.09875v2", "topics": ["human-in-the-loop", "evaluation", "safety-security"], "updated_at": "2025-05-16"}, "text": "Characterizing Unintended Consequences in Human-GUI Agent Collaboration for Web Browsing\nResource type: paper\nReview status: curated\nScope: core\n\nStudies unintended consequences of collaborative web browsing with GUI agents through social-media analysis and interviews. It organizes observed failures, user impacts, and mitigation practices around planning, interface interaction, feedback, privacy, and trust.\nPlatforms: web\nTopics: human-in-the-loop, evaluation, safety-security\nPublished: 2025-05-15\nAuthors: Shuning Zhang, Jingruo Chen, Zhiqi Gao, Jiajing Gao, Xin Yi, Hewu Li\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.09875v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-09875/"}
{"id": "arxiv:2505.10593", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08cc5be720ac4e82560ad6b7549f4c38c24603d23d9d051894d1078c247059e2", "submitted_at": "2025-05-15T05:28:35Z", "updated_at": "2025-05-15T05:28:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.10593v1"}, "authors": ["Shanhui Zhao", "Hao Wen", "Wenjie Du", "Cheng Liang", "Yunxin Liu", "Xiaozhou Ye", "Ye Ouyang", "Yuanchun Li"], "canonical_source_url": "https://arxiv.org/abs/2505.10593", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-10593/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08cc5be720ac4e82560ad6b7549f4c38c24603d23d9d051894d1078c247059e2", "submitted_at": "2025-05-15T05:28:35Z", "updated_at": "2025-05-15T05:28:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.10593v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2505.10593"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10593"}], "name": "LLM-Explorer: Towards Efficient and Affordable LLM-based Exploration for Mobile Apps", "publication": null, "published_at": "2025-05-15", "record_sha256": "b550e9044198ee4ac747f3252f4344178e626be4e2eb80e8106c6eac6d90cf50", "relations": [], "resource_id": "arxiv:2505.10593", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.10593v1", "topics": ["mobile-navigation", "efficiency"], "updated_at": "2025-05-15"}, "text": "LLM-Explorer: Towards Efficient and Affordable LLM-based Exploration for Mobile Apps\nResource type: paper\nReview status: curated\nScope: core\n\nLLM-Explorer uses language models to maintain compact knowledge for mobile app exploration, then generates actions without invoking the model at every step. The study compares exploration coverage and cost across twenty apps.\nPlatforms: unspecified\nTopics: mobile-navigation, efficiency\nPublished: 2025-05-15\nAuthors: Shanhui Zhao, Hao Wen, Wenjie Du, Cheng Liang, Yunxin Liu, Xiaozhou Ye, Ye Ouyang, Yuanchun Li\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.10593v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-10593/"}
{"id": "arxiv:2505.10831", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d986ee6aa62b11759c366fa9f6a01a1a0b12eae715970eacd8053dbbb6622fe5", "submitted_at": "2025-05-16T04:00:31Z", "updated_at": "2025-09-21T00:03:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.10831v3"}, "authors": ["Omar Shaikh", "Shardul Sapkota", "Shan Rizvi", "Eric Horvitz", "Joon Sung Park", "Diyi Yang", "Michael S. Bernstein"], "canonical_source_url": "https://arxiv.org/abs/2505.10831", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-10831/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d986ee6aa62b11759c366fa9f6a01a1a0b12eae715970eacd8053dbbb6622fe5", "submitted_at": "2025-05-16T04:00:31Z", "updated_at": "2025-09-21T00:03:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.10831v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.10831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10831"}], "name": "Creating General User Models from Computer Use", "publication": null, "published_at": "2025-05-16", "record_sha256": "dc3b71fe0984dc6147a5ad60384f8198ded6c921d579d6df36a5fcba28ad2aa8", "relations": [], "resource_id": "arxiv:2505.10831", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.10831v3", "topics": ["memory", "human-in-the-loop", "cross-app"], "updated_at": "2025-09-21"}, "text": "Creating General User Models from Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nProposes general user models that turn computer-use observations into confidence-weighted propositions and revise them as new context arrives. Demonstrations use this memory to contextualize assistants, manage notifications, and support proactive actions across apps.\nPlatforms: unspecified\nTopics: memory, human-in-the-loop, cross-app\nPublished: 2025-05-16\nAuthors: Omar Shaikh, Shardul Sapkota, Shan Rizvi, Eric Horvitz, Joon Sung Park, Diyi Yang, Michael S. Bernstein\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.10831v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-10831/"}
{"id": "arxiv:2505.10924", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a0890cd57809ff5e3bede6dc2b851b42fad9307a5bfa0ddd744cc2ac70d0c2", "submitted_at": "2025-05-16T06:56:42Z", "updated_at": "2026-04-29T05:56:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.10924v4"}, "authors": ["Ada Chen", "Yongjiang Wu", "Junyuan Zhang", "Jingyu Xiao", "Shu Yang", "Jen-tse Huang", "Kun Wang", "Wenxuan Wang", "Shuai Wang"], "canonical_source_url": "https://arxiv.org/abs/2505.10924", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-10924/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a0890cd57809ff5e3bede6dc2b851b42fad9307a5bfa0ddd744cc2ac70d0c2", "submitted_at": "2025-05-16T06:56:42Z", "updated_at": "2026-04-29T05:56:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.10924v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.10924"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.10924"}], "name": "A Survey on the Safety and Security Threats of Computer-Using Agents: JARVIS or Ultron?", "publication": null, "published_at": "2025-05-16", "record_sha256": "8a6a39ad469a7785aaeef404f78c2cdf3072778ec7517f72c6d5b74bb8b22b6b", "relations": [], "resource_id": "arxiv:2505.10924", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.10924v4", "topics": ["survey", "safety-security"], "updated_at": "2026-04-29"}, "text": "A Survey on the Safety and Security Threats of Computer-Using Agents: JARVIS or Ultron?\nResource type: survey\nReview status: curated\nScope: core\n\nSystematizes safety and security threats to computer-using agents, with taxonomies of threats and defensive strategies.\nPlatforms: cross-platform\nTopics: survey, safety-security\nPublished: 2025-05-16\nAuthors: Ada Chen, Yongjiang Wu, Junyuan Zhang, Jingyu Xiao, Shu Yang, Jen-tse Huang, Kun Wang, Wenxuan Wang, Shuai Wang\nRecorded arXiv version: v4\n\nPrimary source: https://arxiv.org/abs/2505.10924v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-10924/"}
{"id": "arxiv:2505.11891", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ca1ebfdd5a96e146d37539294462c7fb67bf20deaf77307de37aa3bb1dde082", "submitted_at": "2025-05-17T07:58:34Z", "updated_at": "2026-02-02T06:55:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.11891v3"}, "authors": ["Weikai Xu", "Zhizheng Jiang", "Yuxuan Liu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Yuanchun Li", "Yunxin Liu", "Bin Wang", "Bo An"], "canonical_source_url": "https://arxiv.org/abs/2505.11891", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-11891/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ca1ebfdd5a96e146d37539294462c7fb67bf20deaf77307de37aa3bb1dde082", "submitted_at": "2025-05-17T07:58:34Z", "updated_at": "2026-02-02T06:55:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.11891v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2505.11891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.11891"}], "name": "Mobile-Bench-v2: A More Realistic and Comprehensive Benchmark for VLM-based Mobile Agents", "publication": null, "published_at": "2025-05-17", "record_sha256": "1ce28465b7b0cc1eca2ae30ab5b0b194b5347582fc88e1b689d1f30a8fc99c36", "relations": [], "resource_id": "arxiv:2505.11891", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.11891v3", "topics": ["evaluation", "mobile-navigation", "robustness", "human-in-the-loop"], "updated_at": "2026-02-02"}, "text": "Mobile-Bench-v2: A More Realistic and Comprehensive Benchmark for VLM-based Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMobile-Bench-v2 evaluates mobile GUI agents using multiple acceptable offline action paths, noisy interfaces with ads or pop-ups, and ambiguous instructions with scripted clarification exchanges. These splits separate step-level execution, robustness, and proactive interaction.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, robustness, human-in-the-loop\nPublished: 2025-05-17\nAuthors: Weikai Xu, Zhizheng Jiang, Yuxuan Liu, Pengzhi Gao, Wei Liu, Jian Luan, Yuanchun Li, Yunxin Liu, Bin Wang, Bo An\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.11891v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-11891/"}
{"id": "arxiv:2505.12299", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bc92029e16b5afb4b54837d2ab2304c7b8176548061c9deafdca101754128dc", "submitted_at": "2025-05-18T08:28:05Z", "updated_at": "2026-03-21T09:25:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12299v4"}, "authors": ["Kun Huang", "Weikai Xu", "Yuxuan Liu", "Quandong Wang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Bin Wang", "Bo An"], "canonical_source_url": "https://arxiv.org/abs/2505.12299", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12299/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bc92029e16b5afb4b54837d2ab2304c7b8176548061c9deafdca101754128dc", "submitted_at": "2025-05-18T08:28:05Z", "updated_at": "2026-03-21T09:25:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12299v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2505.12299"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12299"}], "name": "MobileIPL: Enhancing Mobile Agents Thinking Process via Iterative Preference Learning", "publication": null, "published_at": "2025-05-18", "record_sha256": "85a5b6fccdf157cbe91850bd83e0170dea4b2672e927da54ed88c9946236476c", "relations": [], "resource_id": "arxiv:2505.12299", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12299v4", "topics": ["mobile-navigation", "reinforcement-learning", "planning", "data-generation"], "updated_at": "2026-03-21"}, "text": "MobileIPL: Enhancing Mobile Agents Thinking Process via Iterative Preference Learning\nResource type: paper\nReview status: curated\nScope: core\n\nMobileIPL samples trees of action-planning thoughts, propagates rule-based leaf rewards backward, and forms preference pairs for reasoning-level optimization. It also expands screenshot-based instruction data through staged evolution before evaluation on mobile GUI-agent benchmarks.\nPlatforms: unspecified\nTopics: mobile-navigation, reinforcement-learning, planning, data-generation\nPublished: 2025-05-18\nAuthors: Kun Huang, Weikai Xu, Yuxuan Liu, Quandong Wang, Pengzhi Gao, Wei Liu, Jian Luan, Bin Wang, Bo An\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12299v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12299/"}
{"id": "arxiv:2505.12370", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755f885b79e9f3eef9b4cbdb8593fb3cde1ac8a92f86df52d5d4203eb6cad039", "submitted_at": "2025-05-18T11:22:04Z", "updated_at": "2025-05-24T03:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12370v2"}, "authors": ["Xinbin Yuan", "Jian Zhang", "Kaixin Li", "Zhuoxuan Cai", "Lujian Yao", "Jie Chen", "Enguang Wang", "Qibin Hou", "Jinwei Chen", "Peng-Tao Jiang", "Bo Li"], "canonical_source_url": "https://arxiv.org/abs/2505.12370", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12370/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755f885b79e9f3eef9b4cbdb8593fb3cde1ac8a92f86df52d5d4203eb6cad039", "submitted_at": "2025-05-18T11:22:04Z", "updated_at": "2025-05-24T03:01:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12370v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.12370"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12370"}], "name": "Enhancing Visual Grounding for GUI Agents via Self-Evolutionary Reinforcement Learning", "publication": null, "published_at": "2025-05-18", "record_sha256": "cf2a9ce5205ea940ea1399f1a9651a957504032f70ead68926b95f1901dbc903", "relations": [], "resource_id": "arxiv:2505.12370", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12370v2", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24"}, "text": "Enhancing Visual Grounding for GUI Agents via Self-Evolutionary Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nProposes GUI grounding training that combines curated seed examples, continuous location-dependent policy feedback, and iterative refinement guided by attention maps. The study evaluates the resulting reinforcement-learning procedure on three grounding benchmarks, including ScreenSpot-Pro.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning\nPublished: 2025-05-18\nAuthors: Xinbin Yuan, Jian Zhang, Kaixin Li, Zhuoxuan Cai, Lujian Yao, Jie Chen, Enguang Wang, Qibin Hou, Jinwei Chen, Peng-Tao Jiang, Bo Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12370v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12370/"}
{"id": "arxiv:2505.12493", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10d826ea7715ab7054522a9dcba622d5dae3e6c3faee421e4b168e16104d184b", "submitted_at": "2025-05-18T16:34:30Z", "updated_at": "2025-10-10T09:28:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12493v3"}, "authors": ["Longxi Gao", "Li Zhang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2505.12493", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12493/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "10d826ea7715ab7054522a9dcba622d5dae3e6c3faee421e4b168e16104d184b", "submitted_at": "2025-05-18T16:34:30Z", "updated_at": "2025-10-10T09:28:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.12493v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.12493"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12493"}], "name": "GUI-Shift: Enhancing VLM-Based GUI Agents through Self-supervised Reinforcement Learning", "publication": null, "published_at": "2025-05-18", "record_sha256": "4a65de1fee627687f195253e9cd4d1be843bd91d4078bf351796a82a07ac2385", "relations": [], "resource_id": "arxiv:2505.12493", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12493v3", "topics": ["reinforcement-learning", "data-generation", "grounding", "mobile-navigation"], "updated_at": "2025-10-10"}, "text": "GUI-Shift: Enhancing VLM-Based GUI Agents through Self-supervised Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Shift trains an inverse-dynamics task that predicts the initial action connecting two GUI states, using trajectories without natural-language instructions. It combines rule-based reinforcement learning with data filtering and evaluates transfer to Android automation and screenshot grounding.\nPlatforms: android\nTopics: reinforcement-learning, data-generation, grounding, mobile-navigation\nPublished: 2025-05-18\nAuthors: Longxi Gao, Li Zhang, Pengzhi Gao, Wei Liu, Jian Luan, Mengwei Xu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12493v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12493/"}
{"id": "arxiv:2505.12632", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64672e6a3450256a8b04d9db231e885e2dd4e8688235f6c7853d70d6276610ff", "submitted_at": "2025-05-19T02:39:03Z", "updated_at": "2025-05-19T02:39:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.12632v1"}, "authors": ["Yunseok Jang", "Yeda Song", "Sungryull Sohn", "Lajanugen Logeswaran", "Tiange Luo", "Dong-Ki Kim", "Kyunghoon Bae", "Honglak Lee"], "canonical_source_url": "https://arxiv.org/abs/2505.12632", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12632/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64672e6a3450256a8b04d9db231e885e2dd4e8688235f6c7853d70d6276610ff", "submitted_at": "2025-05-19T02:39:03Z", "updated_at": "2025-05-19T02:39:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.12632v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2505.12632"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12632"}], "name": "Scalable Video-to-Dataset Generation for Cross-Platform Mobile Agents", "publication": null, "published_at": "2025-05-19", "record_sha256": "5b7f2dca02ca5b987ba14f234c923a858d6fa4c7529d0c7e4a02b1acb0066fc9", "relations": [], "resource_id": "arxiv:2505.12632", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12632v1", "topics": ["video", "data-generation", "mobile-navigation"], "updated_at": "2025-05-19"}, "text": "Scalable Video-to-Dataset Generation for Cross-Platform Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MONDAY and an automated pipeline that extracts mobile-navigation training data from instructional videos across operating-system platforms.\nPlatforms: cross-platform\nTopics: video, data-generation, mobile-navigation\nPublished: 2025-05-19\nAuthors: Yunseok Jang, Yeda Song, Sungryull Sohn, Lajanugen Logeswaran, Tiange Luo, Dong-Ki Kim, Kyunghoon Bae, Honglak Lee\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12632v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12632/"}
{"id": "arxiv:2505.12842", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce9570eb6a52feb220cdd2eb4c7f37a2ea63082febe48888a2654c974138a82d", "submitted_at": "2025-05-19T08:29:05Z", "updated_at": "2025-11-22T15:07:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12842v4"}, "authors": ["Zheng Wu", "Pengzhou Cheng", "Zongru Wu", "Lingzhong Dong", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2505.12842", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12842/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce9570eb6a52feb220cdd2eb4c7f37a2ea63082febe48888a2654c974138a82d", "submitted_at": "2025-05-19T08:29:05Z", "updated_at": "2025-11-22T15:07:27Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2505.12842v4"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.12842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12842"}], "name": "GEM: Gaussian Embedding Modeling for Out-of-Distribution Detection in GUI Agents", "publication": null, "published_at": "2025-05-19", "record_sha256": "bb9363f1dea5d4f608faf5a46e50a0d28c279eb0266a9f9dd2845f96e85cfc26", "relations": [], "resource_id": "arxiv:2505.12842", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12842v4", "topics": ["robustness", "safety-security", "evaluation", "efficiency"], "updated_at": "2025-11-22"}, "text": "GEM: Gaussian Embedding Modeling for Out-of-Distribution Detection in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGEM detects instructions outside a GUI agent’s capabilities by modeling distances in its input-embedding space with a Gaussian mixture. Experiments across mobile, computer and browser data also examine selective escalation to a cloud model when an input is detected as unfamiliar.\nPlatforms: desktop, web, cross-platform\nTopics: robustness, safety-security, evaluation, efficiency\nPublished: 2025-05-19\nAuthors: Zheng Wu, Pengzhou Cheng, Zongru Wu, Lingzhong Dong, Zhuosheng Zhang\nRecorded arXiv version: v4\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12842v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12842/"}
{"id": "arxiv:2505.12981", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9657a66aebe46151ed56bf76131ffe45f52459c1cf0e000907113776c91259d9", "submitted_at": "2025-05-19T11:17:46Z", "updated_at": "2025-05-20T07:02:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12981v2"}, "authors": ["Liangxuan Wu", "Chao Wang", "Tianming Liu", "Yanjie Zhao", "Haoyu Wang"], "canonical_source_url": "https://arxiv.org/abs/2505.12981", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-12981/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9657a66aebe46151ed56bf76131ffe45f52459c1cf0e000907113776c91259d9", "submitted_at": "2025-05-19T11:17:46Z", "updated_at": "2025-05-20T07:02:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.12981v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "identifiers": {"arxiv": "2505.12981"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.12981"}], "name": "From Assistants to Adversaries: Exploring the Security Risks of Mobile LLM Agents", "publication": null, "published_at": "2025-05-19", "record_sha256": "33b57fc7bc263aa32fe8f57cced683b9bedd0c1c2acadb0e7cb60a1ca828ae95", "relations": [], "resource_id": "arxiv:2505.12981", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.12981v2", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-05-20"}, "text": "From Assistants to Adversaries: Exploring the Security Risks of Mobile LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes security vulnerabilities across OEM system-level, third-party, and research mobile LLM agents, covering GUI and cross-application attack surfaces.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2025-05-19\nAuthors: Liangxuan Wu, Chao Wang, Tianming Liu, Yanjie Zhao, Haoyu Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.12981v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-12981/"}
{"id": "arxiv:2505.13076", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "709ee511ffa39e805fcc58abe066cf2cec6c1707a4a9c64d9cfd107bef7544af", "submitted_at": "2025-05-19T13:10:29Z", "updated_at": "2025-05-19T13:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.13076v1"}, "authors": ["Mykyta Mudryi", "Markiyan Chaklosh", "Grzegorz Wójcik"], "canonical_source_url": "https://arxiv.org/abs/2505.13076", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-13076/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "709ee511ffa39e805fcc58abe066cf2cec6c1707a4a9c64d9cfd107bef7544af", "submitted_at": "2025-05-19T13:10:29Z", "updated_at": "2025-05-19T13:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.13076v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "identifiers": {"arxiv": "2505.13076"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13076"}], "name": "The Hidden Dangers of Browsing AI Agents", "publication": null, "published_at": "2025-05-19", "record_sha256": "29b9969cb22e828f44e1e30ef7e5f56e3d6ef106edfcd2390b936bac385ddbca", "relations": [], "resource_id": "arxiv:2505.13076", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.13076v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-05-19"}, "text": "The Hidden Dangers of Browsing AI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOutlines an end-to-end threat model and empirical security evaluation across architectural layers of autonomous browser agents.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2025-05-19\nAuthors: Mykyta Mudryi, Markiyan Chaklosh, Grzegorz Wójcik\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.13076v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-13076/"}
{"id": "arxiv:2505.13227", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bec4ad578c01ebdfb58bd269562d3ca7a2bd8d2687cde07cb867db0b78cb105", "submitted_at": "2025-05-19T15:09:23Z", "updated_at": "2025-10-24T19:08:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.13227v3"}, "authors": ["Tianbao Xie", "Jiaqi Deng", "Xiaochuan Li", "Junlin Yang", "Haoyuan Wu", "Jixuan Chen", "Wenjing Hu", "Xinyuan Wang", "Yuhui Xu", "Zekun Wang", "Yiheng Xu", "Junli Wang", "Doyen Sahoo", "Tao Yu", "Caiming Xiong"], "canonical_source_url": "https://arxiv.org/abs/2505.13227", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-13227/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bec4ad578c01ebdfb58bd269562d3ca7a2bd8d2687cde07cb867db0b78cb105", "submitted_at": "2025-05-19T15:09:23Z", "updated_at": "2025-10-24T19:08:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.13227v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.13227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13227"}], "name": "Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis", "publication": null, "published_at": "2025-05-19", "record_sha256": "9afabb0ca3d3abf2f42accbfc98f253aa0e63e275061164f464d00cb75d87dad", "relations": [], "resource_id": "arxiv:2505.13227", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.13227v3", "topics": ["grounding", "data-generation", "evaluation"], "updated_at": "2025-10-24"}, "text": "Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OSWorld-G grounding evaluation and the Jedi synthesis pipeline for examples spanning layout reasoning and fine-grained interaction.\nPlatforms: desktop\nTopics: grounding, data-generation, evaluation\nPublished: 2025-05-19\nAuthors: Tianbao Xie, Jiaqi Deng, Xiaochuan Li, Junlin Yang, Haoyuan Wu, Jixuan Chen, Wenjing Hu, Xinyuan Wang, Yuhui Xu, Zekun Wang, Yiheng Xu, Junli Wang, Doyen Sahoo, Tao Yu, Caiming Xiong\nRecorded arXiv version: v3\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.13227v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-13227/"}
{"id": "arxiv:2505.13909", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "026a64046926912a54e6a7c6cf62a17740a6f89a94713615f9750f8bf8302c78", "submitted_at": "2025-05-20T04:20:18Z", "updated_at": "2026-03-03T03:55:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.13909v2"}, "authors": ["Yanheng He", "Jiahe Jin", "Pengfei Liu"], "canonical_source_url": "https://arxiv.org/abs/2505.13909", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-13909/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "026a64046926912a54e6a7c6cf62a17740a6f89a94713615f9750f8bf8302c78", "submitted_at": "2025-05-20T04:20:18Z", "updated_at": "2026-03-03T03:55:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.13909v2"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.13909"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.13909"}], "name": "Efficient Agent Training for Computer Use", "publication": null, "published_at": "2025-05-20", "record_sha256": "49b35095ecc5d2f8dd91ca9582075e3f0e0d993a24a8a8dbaaca960bd15ff18b", "relations": [], "resource_id": "arxiv:2505.13909", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.13909v2", "topics": ["data-generation", "evaluation"], "updated_at": "2026-03-03"}, "text": "Efficient Agent Training for Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nPC Agent-E augments 312 human computer-use trajectories with alternative action decisions synthesized by Claude 3.7 Sonnet. It studies training with this combination and introduces WindowsAgentArena-V2 as an updated evaluation benchmark.\nPlatforms: windows, desktop\nTopics: data-generation, evaluation\nPublished: 2025-05-20\nAuthors: Yanheng He, Jiahe Jin, Pengfei Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.13909v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-13909/"}
{"id": "arxiv:2505.14141", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de6c501a42c3828bdf795ac59c38b982e738c38bd5d0c59cfb6cb10b51e39292", "submitted_at": "2025-05-20T09:45:55Z", "updated_at": "2025-05-20T09:45:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.14141v1"}, "authors": ["Fanglin Mo", "Junzhe Chen", "Haoxuan Zhu", "Xuming Hu"], "canonical_source_url": "https://arxiv.org/abs/2505.14141", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-14141/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de6c501a42c3828bdf795ac59c38b982e738c38bd5d0c59cfb6cb10b51e39292", "submitted_at": "2025-05-20T09:45:55Z", "updated_at": "2025-05-20T09:45:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.14141v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.14141"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14141"}], "name": "Building a Stable Planner: An Extended Finite State Machine Based Planning Module for Mobile GUI Agent", "publication": null, "published_at": "2025-05-20", "record_sha256": "7a309dacadbc0949d49963879091dd70996691caec04a762eaa1bd6cd53e4dc6", "relations": [], "resource_id": "arxiv:2505.14141", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.14141v1", "topics": ["planning", "mobile-navigation"], "updated_at": "2025-05-20"}, "text": "Building a Stable Planner: An Extended Finite State Machine Based Planning Module for Mobile GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nSPlanner models mobile-app functions and configurations with extended finite-state machines, traverses them to construct execution paths, and converts those paths into natural-language plans. A VLM follows these plans to perform GUI actions, with evaluation on AndroidWorld.\nPlatforms: android\nTopics: planning, mobile-navigation\nPublished: 2025-05-20\nAuthors: Fanglin Mo, Junzhe Chen, Haoxuan Zhu, Xuming Hu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.14141v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-14141/"}
{"id": "arxiv:2505.14289", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c256d8b0d774efdeee896860f8c504c0d1593215508fdf0881d3fd3cdc3968c", "submitted_at": "2025-05-20T12:41:05Z", "updated_at": "2026-06-05T10:24:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.14289v2"}, "authors": ["Yijie Lu", "Manman Zhao", "Tianjie Ju", "Zihe Yan", "Xinbei Ma", "Yuan Guo", "Daizong Ding", "Gongshen Liu", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2505.14289", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-14289/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c256d8b0d774efdeee896860f8c504c0d1593215508fdf0881d3fd3cdc3968c", "submitted_at": "2025-05-20T12:41:05Z", "updated_at": "2026-06-05T10:24:33Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.14289v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.14289"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14289"}], "name": "EVA: Evolving Semantic Adversaries for Red-Teaming GUI Agents Against Environmental Injection Attacks", "publication": null, "published_at": "2025-05-20", "record_sha256": "d8080ad2dde2681fcf41aac3651e4b852d140e47e839b2c0fe2bcb11258542a0", "relations": [], "resource_id": "arxiv:2505.14289", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.14289v2", "topics": ["safety-security", "evaluation"], "updated_at": "2026-06-05"}, "text": "EVA: Evolving Semantic Adversaries for Red-Teaming GUI Agents Against Environmental Injection Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces EVA, a red-teaming framework that evolves semantic payloads for GUI environmental-injection tests. It studies the contribution of semantic versus visual deception and distills discovered patterns into reusable evaluation rules.\nPlatforms: unspecified\nTopics: safety-security, evaluation\nPublished: 2025-05-20\nAuthors: Yijie Lu, Manman Zhao, Tianjie Ju, Zihe Yan, Xinbei Ma, Yuan Guo, Daizong Ding, Gongshen Liu, Zhuosheng Zhang\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.14289v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-14289/"}
{"id": "arxiv:2505.14418", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815345ed44d70f40ce9784c212d269ffaaa476e5c2ea5cdbbaaeb98d22eba3cb", "submitted_at": "2025-05-20T14:29:18Z", "updated_at": "2026-06-15T06:42:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.14418v3"}, "authors": ["Pengzhou Cheng", "Haowen Hu", "Zheng Wu", "Zongru Wu", "Tianjie Ju", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2505.14418", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-14418/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815345ed44d70f40ce9784c212d269ffaaa476e5c2ea5cdbbaaeb98d22eba3cb", "submitted_at": "2025-05-20T14:29:18Z", "updated_at": "2026-06-15T06:42:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.14418v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.14418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.14418"}], "name": "Hidden Ghost Hand: Unveiling Backdoor Vulnerabilities in MLLM-Powered Mobile GUI Agents", "publication": null, "published_at": "2025-05-20", "record_sha256": "3244437c25e3ec7deeb1037424a8171e2e6d69934411d749a7f947a37e79d878", "relations": [], "resource_id": "arxiv:2505.14418", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.14418v3", "topics": ["safety-security", "mobile-navigation", "interaction-history"], "updated_at": "2026-06-15"}, "text": "Hidden Ghost Hand: Unveiling Backdoor Vulnerabilities in MLLM-Powered Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies backdoor risks in mobile GUI agents through AgentGhost, which combines goal-level and interaction-history triggers during model training. The paper evaluates multiple mobile agents and a tailored defense.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, interaction-history\nPublished: 2025-05-20\nAuthors: Pengzhou Cheng, Haowen Hu, Zheng Wu, Zongru Wu, Tianjie Ju, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.14418v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-14418/"}
{"id": "arxiv:2505.15259", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb7a88a0cf07427c1878c896dec920852810879c30bad8692358f2e2016539", "submitted_at": "2025-05-21T08:36:18Z", "updated_at": "2025-05-24T15:08:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15259v2"}, "authors": ["Hyunseok Lee", "Jeonghoon Kim", "Beomjun Kim", "Jihoon Tack", "Chansong Jo", "Jaehong Lee", "Cheonbok Park", "Sookyo In", "Jinwoo Shin", "Kang Min Yoo"], "canonical_source_url": "https://arxiv.org/abs/2505.15259", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-15259/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb7a88a0cf07427c1878c896dec920852810879c30bad8692358f2e2016539", "submitted_at": "2025-05-21T08:36:18Z", "updated_at": "2025-05-24T15:08:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15259v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.15259"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15259"}], "name": "ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search", "publication": null, "published_at": "2025-05-21", "record_sha256": "d1d1e99adb6931d0a944dfe688ea7ada77f190817f34b34c263f6913824c421a", "relations": [], "resource_id": "arxiv:2505.15259", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.15259v2", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-24"}, "text": "ReGUIDE: Data Efficient GUI Grounding via Spatial Reasoning and Search\nResource type: paper\nReview status: curated\nScope: core\n\nCombines self-generated localization reasoning, spatial consistency checks, and inference-time search to improve data-efficient web grounding.\nPlatforms: web\nTopics: grounding, reinforcement-learning\nPublished: 2025-05-21\nAuthors: Hyunseok Lee, Jeonghoon Kim, Beomjun Kim, Jihoon Tack, Chansong Jo, Jaehong Lee, Cheonbok Park, Sookyo In, Jinwoo Shin, Kang Min Yoo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.15259v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-15259/"}
{"id": "arxiv:2505.15277", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "440abdc75c0a51478cfedc6ee0a9196ccff641ec5a23849ba56be43448bab15f", "submitted_at": "2025-05-21T08:56:55Z", "updated_at": "2025-11-25T21:34:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15277v2"}, "authors": ["Hyungjoo Chae", "Sunghwan Kim", "Junhee Cho", "Seungone Kim", "Seungjun Moon", "Gyeom Hwangbo", "Dongha Lim", "Minjin Kim", "Yeonjun Hwang", "Minju Gwak", "Dongwook Choi", "Minseok Kang", "Gwanhoon Im", "ByeongUng Cho", "Hyojun Kim", "Jun Hee Han", "Taeyoon Kwon", "Minju Kim", "Beong-woo Kwak", "Dongjin Kang", "Jinyoung Yeo"], "canonical_source_url": "https://arxiv.org/abs/2505.15277", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-15277/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "440abdc75c0a51478cfedc6ee0a9196ccff641ec5a23849ba56be43448bab15f", "submitted_at": "2025-05-21T08:56:55Z", "updated_at": "2025-11-25T21:34:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15277v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.15277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15277"}], "name": "Web-Shepherd: Advancing PRMs for Reinforcing Web Agents", "publication": null, "published_at": "2025-05-21", "record_sha256": "15a9845aa3eba000d93c895bdc1a1f6f8944ee43b6c4278170a2a2ae8810dd37", "relations": [], "resource_id": "arxiv:2505.15277", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.15277v2", "topics": ["reinforcement-learning", "evaluation", "data-generation", "web-navigation"], "updated_at": "2025-11-25"}, "text": "Web-Shepherd: Advancing PRMs for Reinforcing Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWeb-Shepherd is a process reward model for assessing individual steps of web-navigation trajectories. It introduces a collection of 40,000 step-level preference pairs with checklists and a separate benchmark for comparing reward models, then evaluates verification during WebArena-lite navigation.\nPlatforms: web\nTopics: reinforcement-learning, evaluation, data-generation, web-navigation\nPublished: 2025-05-21\nAuthors: Hyungjoo Chae, Sunghwan Kim, Junhee Cho, Seungone Kim, Seungjun Moon, Gyeom Hwangbo, Dongha Lim, Minjin Kim, Yeonjun Hwang, Minju Gwak, Dongwook Choi, Minseok Kang, Gwanhoon Im, ByeongUng Cho, Hyojun Kim, Jun Hee Han, Taeyoon Kwon, Minju Kim, Beong-woo Kwak, Dongjin Kang, Jinyoung Yeo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.15277v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-15277/"}
{"id": "arxiv:2505.15810", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "321f2d627451d60e130c17574994648fae69fc784b0ed03f6560e7f1ef9eb023", "submitted_at": "2025-05-21T17:59:09Z", "updated_at": "2025-05-22T11:15:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15810v2"}, "authors": ["Yuqi Zhou", "Sunhao Dai", "Shuai Wang", "Kaiwen Zhou", "Qinglin Jia", "Jun Xu"], "canonical_source_url": "https://arxiv.org/abs/2505.15810", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-15810/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "321f2d627451d60e130c17574994648fae69fc784b0ed03f6560e7f1ef9eb023", "submitted_at": "2025-05-21T17:59:09Z", "updated_at": "2025-05-22T11:15:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.15810v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.15810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.15810"}], "name": "GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents", "publication": null, "published_at": "2025-05-21", "record_sha256": "f0e0a722bcacb919f1b866b74da0e6ffcd1f57617be3749cbcbcdb80e979aff3", "relations": [], "resource_id": "arxiv:2505.15810", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.15810v2", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-05-22"}, "text": "GUI-G1: Understanding R1-Zero-Like Training for Visual Grounding in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes reasoning templates, reward design, and policy updates in GUI grounding, then proposes targeted adjustments to the training process.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2025-05-21\nAuthors: Yuqi Zhou, Sunhao Dai, Shuai Wang, Kaiwen Zhou, Qinglin Jia, Jun Xu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.15810v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-15810/"}
{"id": "arxiv:2505.16282", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c092569d4d79026ddbdcc3bc9da1cfc617ff44886dab0b36f311cc81083e0fb5", "submitted_at": "2025-05-22T06:24:32Z", "updated_at": "2025-05-22T06:24:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16282v1"}, "authors": ["Fanbin Lu", "Zhisheng Zhong", "Shu Liu", "Chi-Wing Fu", "Jiaya Jia"], "canonical_source_url": "https://arxiv.org/abs/2505.16282", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-16282/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c092569d4d79026ddbdcc3bc9da1cfc617ff44886dab0b36f311cc81083e0fb5", "submitted_at": "2025-05-22T06:24:32Z", "updated_at": "2025-05-22T06:24:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16282v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.16282"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16282"}], "name": "ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay", "publication": null, "published_at": "2025-05-22", "record_sha256": "1054e9d0e49ec587d4030afc8c64043918a3d36c8b6afadc3a967fef7f5f6b22", "relations": [], "resource_id": "arxiv:2505.16282", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.16282v1", "topics": ["reinforcement-learning", "long-horizon"], "updated_at": "2025-05-22"}, "text": "ARPO:End-to-End Policy Optimization for GUI Agents with Experience Replay\nResource type: paper\nReview status: curated\nScope: core\n\nARPO extends group-relative policy optimization with a replay buffer that reuses successful GUI trajectories across training iterations. It filters tasks using baseline performance and compares end-to-end reinforcement learning with offline preference optimization on OSWorld.\nPlatforms: desktop\nTopics: reinforcement-learning, long-horizon\nPublished: 2025-05-22\nAuthors: Fanbin Lu, Zhisheng Zhong, Shu Liu, Chi-Wing Fu, Jiaya Jia\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.16282v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-16282/"}
{"id": "arxiv:2505.16421", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f311fb50b36c234787e652b5689246460153a8f442e83b312048cdf6a4fb1502", "submitted_at": "2025-05-22T09:07:43Z", "updated_at": "2025-10-08T21:16:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.16421v2"}, "authors": ["Zhepei Wei", "Wenlin Yao", "Yao Liu", "Weizhi Zhang", "Qin Lu", "Liang Qiu", "Changlong Yu", "Puyang Xu", "Chao Zhang", "Bing Yin", "Hyokun Yun", "Lihong Li"], "canonical_source_url": "https://arxiv.org/abs/2505.16421", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-16421/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f311fb50b36c234787e652b5689246460153a8f442e83b312048cdf6a4fb1502", "submitted_at": "2025-05-22T09:07:43Z", "updated_at": "2025-10-08T21:16:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.16421v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.16421"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16421"}], "name": "WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning", "publication": null, "published_at": "2025-05-22", "record_sha256": "579ee06847958d6da95b0d4a667af2e96cc4e4d290b8e3880e4220060a652271", "relations": [], "resource_id": "arxiv:2505.16421", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.16421v2", "topics": ["reinforcement-learning", "web-navigation"], "updated_at": "2025-10-08"}, "text": "WebAgent-R1: Training Web Agents via End-to-End Multi-Turn Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nTrains web agents through asynchronous multi-turn environment interaction, using task-success rewards to optimize complete trajectories.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation\nPublished: 2025-05-22\nAuthors: Zhepei Wei, Wenlin Yao, Yao Liu, Weizhi Zhang, Qin Lu, Liang Qiu, Changlong Yu, Puyang Xu, Chao Zhang, Bing Yin, Hyokun Yun, Lihong Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.16421v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-16421/"}
{"id": "arxiv:2505.16422", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93c67222fb0f581e9f9cb3ff59c8d28faec9dc037180060ad5ac962b4469144c", "submitted_at": "2025-05-22T09:08:47Z", "updated_at": "2025-09-05T02:24:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.16422v3"}, "authors": ["Xiaoran Yin", "Xu Luo", "Hao Wu", "Lianli Gao", "Jingkuan Song"], "canonical_source_url": "https://arxiv.org/abs/2505.16422", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-16422/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "93c67222fb0f581e9f9cb3ff59c8d28faec9dc037180060ad5ac962b4469144c", "submitted_at": "2025-05-22T09:08:47Z", "updated_at": "2025-09-05T02:24:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.16422v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2505.16422"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16422"}], "name": "Unlocking Smarter Device Control: Foresighted Planning with a World Model-Driven Code Execution Approach", "publication": null, "published_at": "2025-05-22", "record_sha256": "7bf528ba727d79509967fcf48046b53504d2ea7fae1aca2427db8c8a862e350f", "relations": [], "resource_id": "arxiv:2505.16422", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.16422v3", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2025-09-05"}, "text": "Unlocking Smarter Device Control: Foresighted Planning with a World Model-Driven Code Execution Approach\nResource type: paper\nReview status: curated\nScope: core\n\nFPWC builds and refines a task-oriented world model before planning mobile-device actions. It executes plans as code and evaluates the approach in simulated environments and on real mobile devices.\nPlatforms: unspecified\nTopics: world-models, planning, mobile-navigation\nPublished: 2025-05-22\nAuthors: Xiaoran Yin, Xu Luo, Hao Wu, Lianli Gao, Jingkuan Song\nRecorded arXiv version: v3\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.16422v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-16422/"}
{"id": "arxiv:2505.16827", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "641001c49fccac1e1c0659fe765eba5a30cefd2a7ff5d00c821f79092c45a1df", "submitted_at": "2025-05-22T16:01:06Z", "updated_at": "2025-05-22T16:01:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16827v1"}, "authors": ["Bin Xie", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Jie Liu", "Min Zhang", "Liqiang Nie"], "canonical_source_url": "https://arxiv.org/abs/2505.16827", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-16827/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "641001c49fccac1e1c0659fe765eba5a30cefd2a7ff5d00c821f79092c45a1df", "submitted_at": "2025-05-22T16:01:06Z", "updated_at": "2025-05-22T16:01:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.16827v1"}, "publication": {"authors": ["Bin Xie", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Jie Liu", "Min Zhang", "Liqiang Nie"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.282", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.282/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5650-5667", "pdf_url": "https://aclanthology.org/2025.acl-long.282.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.282/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "40babce7d310777fa65a498c170320124d4d4941f2cb46ad0ff04aae3583e02f", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.acl-long.282/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.16827"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.16827"}, {"rel": "publication", "url": "https://aclanthology.org/2025.acl-long.282/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.acl-long.282.pdf"}], "name": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent", "publication": {"authors": ["Bin Xie", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Jie Liu", "Min Zhang", "Liqiang Nie"], "date_basis": "Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 7], "doi": "10.18653/v1/2025.acl-long.282", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.acl-long.282/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot.", "pages": "5650-5667", "pdf_url": "https://aclanthology.org/2025.acl-long.282.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent", "type": "conference-paper", "url": "https://aclanthology.org/2025.acl-long.282/", "venue": "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", "volume": null}, "published_at": "2025-05-22", "record_sha256": "ae22d9e5e956a8a299bfbe3536d545528ae64ec05e5a04456df1730fd64fff9e", "relations": [], "resource_id": "arxiv:2505.16827", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.16827v1", "topics": ["mobile-navigation", "data-generation", "memory", "planning"], "updated_at": "2025-05-22"}, "text": "GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-explorer generates exploration goals from screenshots and interface structure, then mines observation-action-outcome triples for app-specific transition knowledge. The resulting training-free agent uses this knowledge during mobile task execution, evaluated on SPA-Bench and AndroidWorld.\nPlatforms: android\nTopics: mobile-navigation, data-generation, memory, planning\nPublished: 2025-05-22\nAuthors: Bin Xie, Rui Shao, Gongwei Chen, Kaiwen Zhou, Yinchuan Li, Jie Liu, Min Zhang, Liqiang Nie\nRecorded arXiv version: v1\n\nPublication edition: GUI-explorer: Autonomous Exploration and Mining of Transition-aware Knowledge for GUI Agent\nPublication authors: Bin Xie, Rui Shao, Gongwei Chen, Kaiwen Zhou, Yinchuan Li, Jie Liu, Min Zhang, Liqiang Nie\nPublication edition date (recorded precision): 2025-07\nPublication venue: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)\nPublication publisher: Association for Computational Linguistics\nPublication pages: 5650-5667\nPublication doi: 10.18653/v1/2025.acl-long.282\nPublication date basis: Explicit publisher citation_publication_date: 2025/7. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot.\nPublication source: https://aclanthology.org/2025.acl-long.282/\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.16827v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-16827/"}
{"id": "arxiv:2505.18121", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40cbdc37216c58acd1b7523c48ad735df9bde485405a4e7b439aec85730c5bc0", "submitted_at": "2025-05-23T17:23:11Z", "updated_at": "2025-05-23T17:23:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.18121v1"}, "authors": ["Danyang Zhang", "Situo Zhang", "Ziyue Yang", "Zichen Zhu", "Zihan Zhao", "Ruisheng Cao", "Lu Chen", "Kai Yu"], "canonical_source_url": "https://arxiv.org/abs/2505.18121", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-18121/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40cbdc37216c58acd1b7523c48ad735df9bde485405a4e7b439aec85730c5bc0", "submitted_at": "2025-05-23T17:23:11Z", "updated_at": "2025-05-23T17:23:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.18121v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.18121"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.18121"}], "name": "ProgRM: Build Better GUI Agents with Progress Rewards", "publication": null, "published_at": "2025-05-23", "record_sha256": "7af9f8ffbe33c03af66353f220ca0fe76ee3b42ff45b9c7560e56176361c77b6", "relations": [], "resource_id": "arxiv:2505.18121", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.18121v1", "topics": ["reinforcement-learning", "data-generation"], "updated_at": "2025-05-23"}, "text": "ProgRM: Build Better GUI Agents with Progress Rewards\nResource type: paper\nReview status: curated\nScope: core\n\nProgRM predicts per-step task completion progress to provide intermediate rewards during online GUI-agent training. It derives progress labels by using longest common subsequences to identify key trajectory steps, allowing useful partial progress to contribute even when a trajectory ultimately fails.\nPlatforms: unspecified\nTopics: reinforcement-learning, data-generation\nPublished: 2025-05-23\nAuthors: Danyang Zhang, Situo Zhang, Ziyue Yang, Zichen Zhu, Zihan Zhao, Ruisheng Cao, Lu Chen, Kai Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.18121v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-18121/"}
{"id": "arxiv:2505.18829", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c382ce428e6dd41ec07af7575461c69906ce386fd2bcd8ad9f0d792d2b8f0b3", "submitted_at": "2025-05-24T18:56:00Z", "updated_at": "2025-10-31T21:48:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.18829v2"}, "authors": ["Kai Mei", "Xi Zhu", "Hang Gao", "Shuhang Lin", "Yongfeng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2505.18829", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-18829/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c382ce428e6dd41ec07af7575461c69906ce386fd2bcd8ad9f0d792d2b8f0b3", "submitted_at": "2025-05-24T18:56:00Z", "updated_at": "2025-10-31T21:48:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.18829v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.18829"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.18829"}], "name": "LiteCUA: Computer as MCP Server for Computer-Use Agent on AIOS", "publication": null, "published_at": "2025-05-24", "record_sha256": "d99b7d3459c59f8d10ed46162aed76c2937e9bbd398f862f24c12335d3b439e6", "relations": [], "resource_id": "arxiv:2505.18829", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.18829v2", "topics": ["tool-integration", "automation"], "updated_at": "2025-10-31"}, "text": "LiteCUA: Computer as MCP Server for Computer-Use Agent on AIOS\nResource type: paper\nReview status: curated\nScope: core\n\nAIOS 1.0 exposes computer states and actions through an MCP server to provide structured environmental context for agents. LiteCUA is a lightweight agent built on this interface, evaluated on OSWorld to study the effect of computer-environment contextualization.\nPlatforms: desktop\nTopics: tool-integration, automation\nPublished: 2025-05-24\nAuthors: Kai Mei, Xi Zhu, Hang Gao, Shuhang Lin, Yongfeng Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.18829v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-18829/"}
{"id": "arxiv:2505.19095", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb782466eccc7a271c89a4161c7b142f93f8f0d585cad39a1ccad048cf514426", "submitted_at": "2025-05-25T11:13:03Z", "updated_at": "2025-05-25T11:13:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.19095v1"}, "authors": ["Runliang Niu", "Jinglong Ji", "Yi Chang", "Qi Wang"], "canonical_source_url": "https://arxiv.org/abs/2505.19095", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-19095/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb782466eccc7a271c89a4161c7b142f93f8f0d585cad39a1ccad048cf514426", "submitted_at": "2025-05-25T11:13:03Z", "updated_at": "2025-05-25T11:13:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.19095v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:47Z", "role": "primary", "sha256": "27144160e8981bb9f36b60c391689774913ff3f026eeeffc2923154c404a6529", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2504.03206%2C2504.03601%2C2504.06821%2C2504.09723%2C2504.11301%2C2504.11543%2C2504.12609%2C2504.14233%2C2504.14348%2C2504.14603%2C2504.16073%2C2504.16419%2C2504.17892%2C2504.18580%2C2504.20965%2C2505.00019%2C2505.00693%2C2505.01578%2C2505.02094%2C2505.03570%2C2505.03574%2C2505.05026%2C2505.05849%2C2505.06507%2C2505.07003%2C2505.07473%2C2505.10468%2C2505.10887%2C2505.12053%2C2505.12490%2C2505.12728%2C2505.13291%2C2505.14030%2C2505.14454%2C2505.14646%2C2505.15801%2C2505.15816%2C2505.17020%2C2505.18746%2C2505.19095%2C2505.19713%2C2505.20023%2C2505.20732%2C2505.20795%2C2505.21055%2C2505.22304%2C2505.22429%2C2505.23450%2C2505.23559%2C2506.00641&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:24Z", "role": "primary", "sha256": "645bd2ecbc013dab3cbc601d5a42011d8f7306915ea2ae06baf0767b3962e9ed", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2505.19095v1"}], "identifiers": {"arxiv": "2505.19095"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.19095"}], "name": "ScreenExplorer: Training a Vision-Language Model for Diverse Exploration in Open GUI World", "publication": null, "published_at": "2025-05-25", "record_sha256": "f8a4a5b34d23ded0564ab5f8f6d850fa9f8da9a030c50138915f2efaa307c6eb", "relations": [], "resource_id": "arxiv:2505.19095", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.19095v1", "topics": ["reinforcement-learning", "world-models", "data-generation"], "updated_at": "2025-05-25"}, "text": "ScreenExplorer: Training a Vision-Language Model for Diverse Exploration in Open GUI World\nResource type: paper\nReview status: curated\nScope: core\n\nTrains ScreenExplorer with reinforcement learning in dynamic GUI environments, using a world-model curiosity reward and experience-stream distillation to support sustained exploration.\nPlatforms: unspecified\nTopics: reinforcement-learning, world-models, data-generation\nPublished: 2025-05-25\nAuthors: Runliang Niu, Jinglong Ji, Yi Chang, Qi Wang\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.19095v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-19095/"}
{"id": "arxiv:2505.19897", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4743f6bfd0ca48717184cf55545e33a229dd5ab2d454385d33ea2f1b25c15b0c", "submitted_at": "2025-05-26T12:27:27Z", "updated_at": "2026-04-20T16:29:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.19897v3"}, "authors": ["Qiushi Sun", "Zhoumianze Liu", "Chang Ma", "Zichen Ding", "Fangzhi Xu", "Zhangyue Yin", "Haiteng Zhao", "Zhenyu Wu", "Kanzhi Cheng", "Zhaoyang Liu", "Jianing Wang", "Qintong Li", "Xiangru Tang", "Tianbao Xie", "Xiachong Feng", "Xiang Li", "Ben Kao", "Wenhai Wang", "Biqing Qi", "Lingpeng Kong", "Zhiyong Wu"], "canonical_source_url": "https://arxiv.org/abs/2505.19897", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-19897/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4743f6bfd0ca48717184cf55545e33a229dd5ab2d454385d33ea2f1b25c15b0c", "submitted_at": "2025-05-26T12:27:27Z", "updated_at": "2026-04-20T16:29:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.19897v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2505.19897"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.19897"}], "name": "ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows", "publication": null, "published_at": "2025-05-26", "record_sha256": "489cb27e2a78bbb3dfeaf2d587b1161907c9ec779ae2f5f2bf503636a96a43f1", "relations": [], "resource_id": "arxiv:2505.19897", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.19897v3", "topics": ["evaluation", "science", "workflow"], "updated_at": "2026-04-20"}, "text": "ScienceBoard: Evaluating Multimodal Autonomous Agents in Realistic Scientific Workflows\nResource type: paper\nReview status: curated\nScope: core\n\nScienceBoard combines an interactive environment with professional scientific software and 169 human-curated tasks in areas including biochemistry, astronomy, and geoinformatics. It evaluates computer-use agents on visually rich research workflows and analyzes their execution limitations.\nPlatforms: desktop\nTopics: evaluation, science, workflow\nPublished: 2025-05-26\nAuthors: Qiushi Sun, Zhoumianze Liu, Chang Ma, Zichen Ding, Fangzhi Xu, Zhangyue Yin, Haiteng Zhao, Zhenyu Wu, Kanzhi Cheng, Zhaoyang Liu, Jianing Wang, Qintong Li, Xiangru Tang, Tianbao Xie, Xiachong Feng, Xiang Li, Ben Kao, Wenhai Wang, Biqing Qi, Lingpeng Kong, Zhiyong Wu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.19897v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-19897/"}
{"id": "arxiv:2505.20013", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171071f26a6dc8b9c8851d2f11d60b113983191c905631e431f352711e0d0e26", "submitted_at": "2025-05-26T14:03:37Z", "updated_at": "2025-09-18T11:32:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.20013v2"}, "authors": ["Minda Hu", "Tianqing Fang", "Jianshu Zhang", "Junyu Ma", "Zhisong Zhang", "Jingyan Zhou", "Hongming Zhang", "Haitao Mi", "Dong Yu", "Irwin King"], "canonical_source_url": "https://arxiv.org/abs/2505.20013", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-20013/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "171071f26a6dc8b9c8851d2f11d60b113983191c905631e431f352711e0d0e26", "submitted_at": "2025-05-26T14:03:37Z", "updated_at": "2025-09-18T11:32:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.20013v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.20013"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.20013"}], "name": "WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollback", "publication": null, "published_at": "2025-05-26", "record_sha256": "3357c9c74f03b297adbfa1f51578208538bc611cfc3cd32a7c2eef81ee871b02", "relations": [], "resource_id": "arxiv:2505.20013", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.20013v2", "topics": ["data-generation", "planning", "self-reflection", "web-navigation"], "updated_at": "2025-09-18"}, "text": "WebCoT: Enhancing Web Agent Reasoning by Reconstructing Chain-of-Thought in Reflection, Branching, and Rollback\nResource type: paper\nReview status: curated\nScope: core\n\nWebCoT turns reflection, look-ahead, branching and rollback procedures into chain-of-thought training rationales. Fine-tuning on these reconstructed trajectories is evaluated for web-agent self-improvement and transfer across navigation tasks.\nPlatforms: web\nTopics: data-generation, planning, self-reflection, web-navigation\nPublished: 2025-05-26\nAuthors: Minda Hu, Tianqing Fang, Jianshu Zhang, Junyu Ma, Zhisong Zhang, Jingyan Zhou, Hongming Zhang, Haitao Mi, Dong Yu, Irwin King\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.20013v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-20013/"}
{"id": "arxiv:2505.20660", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2d51388193e19501b580740969a6e08661c6e0a2cf09e2489c46fe4f19209cd", "submitted_at": "2025-05-27T03:09:06Z", "updated_at": "2025-05-27T03:09:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.20660v1"}, "authors": ["Qinzhuo Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2505.20660", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-20660/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2d51388193e19501b580740969a6e08661c6e0a2cf09e2489c46fe4f19209cd", "submitted_at": "2025-05-27T03:09:06Z", "updated_at": "2025-05-27T03:09:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.20660v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.20660"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.20660"}], "name": "BacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism", "publication": null, "published_at": "2025-05-27", "record_sha256": "05ed7fcea32d7233dd5acb838047d775ab3f18e9e564d2b40f743b9fedb0484b", "relations": [], "resource_id": "arxiv:2505.20660", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.20660v1", "topics": ["robustness", "planning", "data-generation", "mobile-navigation"], "updated_at": "2025-05-27"}, "text": "BacktrackAgent: Enhancing GUI Agent with Error Detection and Backtracking Mechanism\nResource type: paper\nReview status: curated\nScope: core\n\nBacktrackAgent adds verification, judgment, and reflection modules to detect erroneous GUI actions and guide recovery. It develops backtracking-oriented training data that includes post-action pages and evaluates task completion and step accuracy on Mobile3M and Auto-UI.\nPlatforms: unspecified\nTopics: robustness, planning, data-generation, mobile-navigation\nPublished: 2025-05-27\nAuthors: Qinzhuo Wu, Pengzhi Gao, Wei Liu, Jian Luan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.20660v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-20660/"}
{"id": "arxiv:2505.21279", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4fe40099905712f9d7ba89bc7616d09b260fe6113beb8122cade54a66d46e3d7", "submitted_at": "2025-05-27T14:49:30Z", "updated_at": "2025-09-26T09:13:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21279v2"}, "authors": ["Shaoqing Zhang", "Kehai Chen", "Zhuosheng Zhang", "Rumei Li", "Rongxiang Weng", "Yang Xiang", "Min Zhang"], "canonical_source_url": "https://arxiv.org/abs/2505.21279", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-21279/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4fe40099905712f9d7ba89bc7616d09b260fe6113beb8122cade54a66d46e3d7", "submitted_at": "2025-05-27T14:49:30Z", "updated_at": "2025-09-26T09:13:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21279v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2505.21279"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21279"}], "name": "XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level", "publication": null, "published_at": "2025-05-27", "record_sha256": "028930646ce0359b958b46945a5a41da6da25560f826e7c39adadebc9ed192be", "relations": [], "resource_id": "arxiv:2505.21279", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.21279v2", "topics": ["evaluation", "grounding"], "updated_at": "2025-09-26"}, "text": "XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level\nResource type: paper\nReview status: curated\nScope: core\n\nXBOUND evaluates GUI agents at the level of a screen state, considering multiple instructions and possible target actions within it. The study separates action matching from instruction unification and compares effects of grounding and trajectory data.\nPlatforms: unspecified\nTopics: evaluation, grounding\nPublished: 2025-05-27\nAuthors: Shaoqing Zhang, Kehai Chen, Zhuosheng Zhang, Rumei Li, Rongxiang Weng, Yang Xiang, Min Zhang\nRecorded arXiv version: v2\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.21279v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-21279/"}
{"id": "arxiv:2505.21496", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f19671a527de76b702e68311656e42feea88249716f68911a41ec80be205e4c", "submitted_at": "2025-05-27T17:58:06Z", "updated_at": "2025-05-27T17:58:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21496v1"}, "authors": ["Han Xiao", "Guozhi Wang", "Yuxiang Chai", "Zimu Lu", "Weifeng Lin", "Hao He", "Lue Fan", "Liuyang Bian", "Rui Hu", "Liang Liu", "Shuai Ren", "Yafei Wen", "Xiaoxin Chen", "Aojun Zhou", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2505.21496", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-21496/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f19671a527de76b702e68311656e42feea88249716f68911a41ec80be205e4c", "submitted_at": "2025-05-27T17:58:06Z", "updated_at": "2025-05-27T17:58:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21496v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.21496"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21496"}], "name": "UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents", "publication": null, "published_at": "2025-05-27", "record_sha256": "ebc5c49a1b80c5aeb007b0365af8670f7f903e177db24e298cae32290b331434", "relations": [], "resource_id": "arxiv:2505.21496", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.21496v1", "topics": ["data-generation", "reinforcement-learning", "mobile-navigation", "interaction-history"], "updated_at": "2025-05-27"}, "text": "UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUI-Genie jointly improves a mobile GUI agent and a reward model that processes screenshots with interaction history and supplies action- and task-level feedback. Its data pipeline uses rule-based verification, corrupted trajectories, and hard negatives to support reward-guided exploration and iterative training.\nPlatforms: unspecified\nTopics: data-generation, reinforcement-learning, mobile-navigation, interaction-history\nPublished: 2025-05-27\nAuthors: Han Xiao, Guozhi Wang, Yuxiang Chai, Zimu Lu, Weifeng Lin, Hao He, Lue Fan, Liuyang Bian, Rui Hu, Liang Liu, Shuai Ren, Yafei Wen, Xiaoxin Chen, Aojun Zhou, Hongsheng Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.21496v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-21496/"}
{"id": "arxiv:2505.21499", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8aade17e67471c87f259d49aef43b3caf764b6ac14f99e1a4114f4a25dda3d6", "submitted_at": "2025-05-27T17:59:05Z", "updated_at": "2025-05-27T17:59:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21499v1"}, "authors": ["Haowei Wang", "Junjie Wang", "Xiaojun Jia", "Rupeng Zhang", "Mingyang Li", "Zhe Liu", "Yang Liu", "Qing Wang"], "canonical_source_url": "https://arxiv.org/abs/2505.21499", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-21499/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8aade17e67471c87f259d49aef43b3caf764b6ac14f99e1a4114f4a25dda3d6", "submitted_at": "2025-05-27T17:59:05Z", "updated_at": "2025-05-27T17:59:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.21499v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.21499"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21499"}], "name": "AdInject: Real-World Black-Box Attacks on Web Agents via Advertising Delivery", "publication": null, "published_at": "2025-05-27", "record_sha256": "c6f90b9bd7c4ff9745cc7e1146f6b5ffa0ac8acdc638bc0387834e1502b470d6", "relations": [], "resource_id": "arxiv:2505.21499", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.21499v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-05-27"}, "text": "AdInject: Real-World Black-Box Attacks on Web Agents via Advertising Delivery\nResource type: paper\nReview status: curated\nScope: core\n\nStudies advertising delivery as an indirect-injection channel for web agents under a black-box threat model. AdInject evaluates how static ad content shaped by website context can redirect browser actions.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2025-05-27\nAuthors: Haowei Wang, Junjie Wang, Xiaojun Jia, Rupeng Zhang, Mingyang Li, Zhe Liu, Yang Liu, Qing Wang\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.21499v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-21499/"}
{"id": "arxiv:2505.21936", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe3bdb1357a375de0b1700a6eda7d5987f358bd47fddf3405c27272a3453f2f", "submitted_at": "2025-05-28T03:42:09Z", "updated_at": "2026-03-01T07:50:09Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.21936v5"}, "authors": ["Zeyi Liao", "Jaylen Jones", "Linxi Jiang", "Yuting Ning", "Eric Fosler-Lussier", "Yu Su", "Zhiqiang Lin", "Huan Sun"], "canonical_source_url": "https://arxiv.org/abs/2505.21936", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-21936/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe3bdb1357a375de0b1700a6eda7d5987f358bd47fddf3405c27272a3453f2f", "submitted_at": "2025-05-28T03:42:09Z", "updated_at": "2026-03-01T07:50:09Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2505.21936v5"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.21936"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21936"}], "name": "RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments", "publication": null, "published_at": "2025-05-28", "record_sha256": "16021cfeb7dbd621979891c9bd105a99df04b538a87bfff470abd68f32f254b4", "relations": [], "resource_id": "arxiv:2505.21936", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.21936v5", "topics": ["safety-security", "evaluation", "sandbox"], "updated_at": "2026-03-01"}, "text": "RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments\nResource type: paper\nReview status: curated\nScope: core\n\nRedTeamCUA combines a virtual-machine OS environment with containerized websites for controlled prompt-injection tests spanning both interfaces. Its RTC-Bench contains 864 examples and can initialize agents at injection points to separate attack susceptibility from navigation failures.\nPlatforms: web, desktop\nTopics: safety-security, evaluation, sandbox\nPublished: 2025-05-28\nAuthors: Zeyi Liao, Jaylen Jones, Linxi Jiang, Yuting Ning, Eric Fosler-Lussier, Yu Su, Zhiqiang Lin, Huan Sun\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.21936v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-21936/"}
{"id": "arxiv:2505.21964", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17c0dac4cbb7c3b7b94496dc225db061a0dc3bda0a34ef40c58ea451524fada7", "submitted_at": "2025-05-28T04:32:05Z", "updated_at": "2025-11-03T08:44:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21964v2"}, "authors": ["Ziyun Zhang", "Xinyi Liu", "Xiaoyi Zhang", "Jun Wang", "Gang Chen", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2505.21964", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-21964/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17c0dac4cbb7c3b7b94496dc225db061a0dc3bda0a34ef40c58ea451524fada7", "submitted_at": "2025-05-28T04:32:05Z", "updated_at": "2025-11-03T08:44:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.21964v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.21964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.21964"}], "name": "UI-Evol: Automatic Knowledge Evolving for Computer Use Agents", "publication": null, "published_at": "2025-05-28", "record_sha256": "7d1142f416fa34a3b962ce73268560d7d9d4765d60bbb1dd76f6b77d7be5743b", "relations": [], "resource_id": "arxiv:2505.21964", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.21964v2", "topics": ["memory", "self-reflection", "robustness"], "updated_at": "2025-11-03"}, "text": "UI-Evol: Automatic Knowledge Evolving for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UI-Evol, which extracts action sequences from agent interactions and critiques external guidance against those traces. Evaluation on OSWorld examines how revised operational knowledge affects execution and behavioral consistency.\nPlatforms: desktop\nTopics: memory, self-reflection, robustness\nPublished: 2025-05-28\nAuthors: Ziyun Zhang, Xinyi Liu, Xiaoyi Zhang, Jun Wang, Gang Chen, Yan Lu\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.21964v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-21964/"}
{"id": "arxiv:2505.22942", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d260f52763eac186e6f0ff614c818f17602658b6be27cfecc3fea9d80579c9a", "submitted_at": "2025-05-28T23:45:28Z", "updated_at": "2025-06-08T07:57:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.22942v2"}, "authors": ["Yuchen Zhuang", "Di Jin", "Jiaao Chen", "Wenqi Shi", "Hanrui Wang", "Chao Zhang"], "canonical_source_url": "https://arxiv.org/abs/2505.22942", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-22942/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d260f52763eac186e6f0ff614c818f17602658b6be27cfecc3fea9d80579c9a", "submitted_at": "2025-05-28T23:45:28Z", "updated_at": "2025-06-08T07:57:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.22942v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2505.22942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.22942"}], "name": "WorkForceAgent-R1: Incentivizing Reasoning Capability in LLM-based Web Agents via Reinforcement Learning", "publication": null, "published_at": "2025-05-28", "record_sha256": "4c6cea6f63fca79dcf57c94b94f7f6560893f5a7028fe8fc4f2d3ac16158ee92", "relations": [], "resource_id": "arxiv:2505.22942", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.22942v2", "topics": ["reinforcement-learning", "workflow"], "updated_at": "2025-06-08"}, "text": "WorkForceAgent-R1: Incentivizing Reasoning Capability in LLM-based Web Agents via Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUses rule-based rewards for output structure and action correctness to train reasoning for enterprise web navigation.\nPlatforms: web\nTopics: reinforcement-learning, workflow\nPublished: 2025-05-28\nAuthors: Yuchen Zhuang, Di Jin, Jiaao Chen, Wenqi Shi, Hanrui Wang, Chao Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.22942v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-22942/"}
{"id": "arxiv:2505.23596", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f0ec81429fb39e56414d3bfe621df6c6adf461e63fc1a1c8886ea2b6850b8ee", "submitted_at": "2025-05-29T16:08:51Z", "updated_at": "2025-11-19T16:48:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.23596v3"}, "authors": ["Linqiang Guo", "Wei Liu", "Yi Wen Heng", "Tse-Hsun", "Chen", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2505.23596", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-23596/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f0ec81429fb39e56414d3bfe621df6c6adf461e63fc1a1c8886ea2b6850b8ee", "submitted_at": "2025-05-29T16:08:51Z", "updated_at": "2025-11-19T16:48:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2505.23596v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2505.23596"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.23596"}], "name": "Agent-SAMA: State-Aware Mobile Assistant", "publication": null, "published_at": "2025-05-29", "record_sha256": "63f80aad44c44b2c4da5c95a72b8da4f00e49bd3364b1608326970399d96c60c", "relations": [], "resource_id": "arxiv:2505.23596", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.23596v3", "topics": ["planning", "memory", "mobile-navigation"], "updated_at": "2025-11-19"}, "text": "Agent-SAMA: State-Aware Mobile Assistant\nResource type: paper\nReview status: curated\nScope: core\n\nModels mobile application execution as a finite-state machine to support planning, verification, and recovery across interface transitions.\nPlatforms: android\nTopics: planning, memory, mobile-navigation\nPublished: 2025-05-29\nAuthors: Linqiang Guo, Wei Liu, Yi Wen Heng, Tse-Hsun, Chen, Yang Wang\nRecorded arXiv version: v3\n\nHugging Face discovery used the label MAPLE for this ID; the record follows the current primary arXiv title, Agent-SAMA. The secondary-title discrepancy is not treated as proof of a historical rename. Experiments not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.23596v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-23596/"}
{"id": "arxiv:2505.23762", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "799e7b387f7d281f0aa9b182f37b215e3baca2041a0a0f6d414ddf1cec0de003", "submitted_at": "2025-05-29T17:59:51Z", "updated_at": "2025-05-29T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.23762v1"}, "authors": ["Chenyu Yang", "Shiqian Su", "Shi Liu", "Xuan Dong", "Yue Yu", "Weijie Su", "Xuehui Wang", "Zhaoyang Liu", "Jinguo Zhu", "Hao Li", "Wenhai Wang", "Yu Qiao", "Xizhou Zhu", "Jifeng Dai"], "canonical_source_url": "https://arxiv.org/abs/2505.23762", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-23762/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "799e7b387f7d281f0aa9b182f37b215e3baca2041a0a0f6d414ddf1cec0de003", "submitted_at": "2025-05-29T17:59:51Z", "updated_at": "2025-05-29T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.23762v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2505.23762"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.23762"}], "name": "ZeroGUI: Automating Online GUI Learning at Zero Human Cost", "publication": null, "published_at": "2025-05-29", "record_sha256": "21b970772e111e9ea74824a39614f651dcb22f0c29a021d63e4c339033e884be", "relations": [], "resource_id": "arxiv:2505.23762", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.23762v1", "topics": ["reinforcement-learning", "data-generation"], "updated_at": "2025-05-29"}, "text": "ZeroGUI: Automating Online GUI Learning at Zero Human Cost\nResource type: paper\nReview status: curated\nScope: core\n\nCombines generated tasks, estimated success rewards, and online reinforcement learning to automate GUI-agent interaction training.\nPlatforms: cross-platform\nTopics: reinforcement-learning, data-generation\nPublished: 2025-05-29\nAuthors: Chenyu Yang, Shiqian Su, Shi Liu, Xuan Dong, Yue Yu, Weijie Su, Xuehui Wang, Zhaoyang Liu, Jinguo Zhu, Hao Li, Wenhai Wang, Yu Qiao, Xizhou Zhu, Jifeng Dai\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.23762v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-23762/"}
{"id": "arxiv:2505.24838", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f913b87f07fdad0e4c45264c58c8d9b9d13ca26695306d9e0644eb856f18ced7", "submitted_at": "2025-05-30T17:39:52Z", "updated_at": "2025-11-08T18:27:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.24838v2"}, "authors": ["Brandon Man", "Ghadi Nehme", "Md Ferdous Alam", "Faez Ahmed"], "canonical_source_url": "https://arxiv.org/abs/2505.24838", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-24838/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f913b87f07fdad0e4c45264c58c8d9b9d13ca26695306d9e0644eb856f18ced7", "submitted_at": "2025-05-30T17:39:52Z", "updated_at": "2025-11-08T18:27:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2505.24838v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}], "identifiers": {"arxiv": "2505.24838"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.24838"}], "name": "VideoCAD: A Dataset and Model for Learning Long-Horizon 3D CAD UI Interactions from Video", "publication": null, "published_at": "2025-05-30", "record_sha256": "2a4a03f0ff7a6b49b4d5785d6f3e85b92312a711ab38206772c4ad20e620a4f2", "relations": [], "resource_id": "arxiv:2505.24838", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.24838v2", "topics": ["video", "data-generation", "science", "long-horizon"], "updated_at": "2025-11-08"}, "text": "VideoCAD: A Dataset and Model for Learning Long-Horizon 3D CAD UI Interactions from Video\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces synthetic CAD interaction videos, a model for learning interface actions from video, and evaluation tasks for spatial and long-horizon reasoning.\nPlatforms: web\nTopics: video, data-generation, science, long-horizon\nPublished: 2025-05-30\nAuthors: Brandon Man, Ghadi Nehme, Md Ferdous Alam, Faez Ahmed\nRecorded arXiv version: v2\n\nThe contribution was reviewed from the retained abstract. Sections 1 and 3 of the exact-version v2 paper identify Onshape as the browser-based interaction environment; the use of pixel-level input does not make the operated surface a desktop application. No experimental results were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.24838v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-24838/"}
{"id": "arxiv:2505.24878", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f2206333ca25b3f21df7ce49adfd8bc56521b9b8ace7a9b686eccf5b67276fa", "submitted_at": "2025-05-30T17:59:55Z", "updated_at": "2025-05-30T17:59:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.24878v1"}, "authors": ["Yaxin Luo", "Zhaoyi Li", "Jiacheng Liu", "Jiacheng Cui", "Xiaohan Zhao", "Zhiqiang Shen"], "canonical_source_url": "https://arxiv.org/abs/2505.24878", "citation_url": "https://awesomecua.github.io/resources/arxiv-2505-24878/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f2206333ca25b3f21df7ce49adfd8bc56521b9b8ace7a9b686eccf5b67276fa", "submitted_at": "2025-05-30T17:59:55Z", "updated_at": "2025-05-30T17:59:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2505.24878v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2505.24878"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2505.24878"}], "name": "Open CaptchaWorld: A Comprehensive Web-based Platform for Testing and Benchmarking Multimodal LLM Agents", "publication": null, "published_at": "2025-05-30", "record_sha256": "c42f79eac3e2cafdecae61f00ee8d8daefe080d723b88660819c61fadfceb441", "relations": [], "resource_id": "arxiv:2505.24878", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2505.24878v1", "topics": ["evaluation", "web-navigation", "screen-understanding"], "updated_at": "2025-05-30"}, "text": "Open CaptchaWorld: A Comprehensive Web-based Platform for Testing and Benchmarking Multimodal LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOpen CaptchaWorld evaluates multimodal agents on interactive browser-based CAPTCHA puzzles. Its benchmark distinguishes task success from the depth of visual reasoning and interaction needed to complete a puzzle.\nPlatforms: web\nTopics: evaluation, web-navigation, screen-understanding\nPublished: 2025-05-30\nAuthors: Yaxin Luo, Zhaoyi Li, Jiacheng Liu, Jiacheng Cui, Xiaohan Zhao, Zhiqiang Shen\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2505.24878v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2505-24878/"}
{"id": "arxiv:2506.00618", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb94fdba775cb0fb54b312fdff9fa71afd3aefa56b812ff31ceccad33eb5f800", "submitted_at": "2025-05-31T16:04:59Z", "updated_at": "2025-06-20T01:24:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.00618v3"}, "authors": ["Jingyi Yang", "Shuai Shao", "Dongrui Liu", "Jing Shao"], "canonical_source_url": "https://arxiv.org/abs/2506.00618", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-00618/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb94fdba775cb0fb54b312fdff9fa71afd3aefa56b812ff31ceccad33eb5f800", "submitted_at": "2025-05-31T16:04:59Z", "updated_at": "2025-06-20T01:24:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.00618v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.00618"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.00618"}], "name": "RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents", "publication": null, "published_at": "2025-05-31", "record_sha256": "313f7855c5ca9b20d6de8caa2c077a4819418380d7005371147801f6a15d535c", "relations": [], "resource_id": "arxiv:2506.00618", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.00618v3", "topics": ["safety-security", "evaluation"], "updated_at": "2025-06-20"}, "text": "RiOSWorld: Benchmarking the Risk of Multimodal Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces RiOSWorld, a computer-use risk benchmark separating user-originated and environmental threats. It evaluates both risky-goal intent and goal completion across web and other computer applications.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-05-31\nAuthors: Jingyi Yang, Shuai Shao, Dongrui Liu, Jing Shao\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.00618v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-00618/"}
{"id": "arxiv:2506.01391", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c0b2a065898ba4f73fcb5e5b2e92d723b0b55ac6caa46d5bf02d44eb7a2cdd9", "submitted_at": "2025-06-02T07:30:29Z", "updated_at": "2025-06-17T02:57:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.01391v2"}, "authors": ["Zhong Zhang", "Yaxi Lu", "Yikun Fu", "Yupeng Huo", "Shenzhi Yang", "Yesai Wu", "Han Si", "Xin Cong", "Haotian Chen", "Yankai Lin", "Jie Xie", "Wei Zhou", "Wang Xu", "Yuanheng Zhang", "Zhou Su", "Zhongwu Zhai", "Xiaoming Liu", "Yudong Mei", "Jianming Xu", "Hongyan Tian", "Chongyi Wang", "Chi Chen", "Yuan Yao", "Zhiyuan Liu", "Maosong Sun"], "canonical_source_url": "https://arxiv.org/abs/2506.01391", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-01391/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c0b2a065898ba4f73fcb5e5b2e92d723b0b55ac6caa46d5bf02d44eb7a2cdd9", "submitted_at": "2025-06-02T07:30:29Z", "updated_at": "2025-06-17T02:57:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.01391v2"}, "publication": {"authors": ["Zhong Zhang", "Yaxi Lu", "Yikun Fu", "Yupeng Huo", "Shenzhi Yang", "Yesai Wu", "Han Si (司函)", "Xin Cong", "Haotian Chen", "Yankai Lin (林衍凯)", "Jie Xie", "Wei Zhou", "Wang Xu", "Yuanheng Zhang", "Zhou Su", "Zhongwu Zhai", "Xiaoming Liu", "Yudong Mei", "Jianming Xu", "Hongyan Tian", "Chongyi Wang", "Chi Chen", "Yuan Yao", "Zhiyuan Liu", "Maosong Sun (孙茂松)"], "date_basis": "Explicit publisher citation_publication_date: 2025/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 11], "doi": "10.18653/v1/2025.emnlp-demos.12", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.emnlp-demos.12/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "155-180", "pdf_url": "https://aclanthology.org/2025.emnlp-demos.12.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2025.emnlp-demos.12/", "venue": "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations", "volume": null}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:09:59Z", "role": "primary", "sha256": "5de5a828e06a63c1627b8ad9f9a46c1643d8b4b7d8ed6156fd9790b6bd083155", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.publisher", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://aclanthology.org/2025.emnlp-demos.12/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2506.01391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.01391"}, {"rel": "publication", "url": "https://aclanthology.org/2025.emnlp-demos.12/"}, {"rel": "publication_pdf", "url": "https://aclanthology.org/2025.emnlp-demos.12.pdf"}], "name": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning", "publication": {"authors": ["Zhong Zhang", "Yaxi Lu", "Yikun Fu", "Yupeng Huo", "Shenzhi Yang", "Yesai Wu", "Han Si (司函)", "Xin Cong", "Haotian Chen", "Yankai Lin (林衍凯)", "Jie Xie", "Wei Zhou", "Wang Xu", "Yuanheng Zhang", "Zhou Su", "Zhongwu Zhai", "Xiaoming Liu", "Yudong Mei", "Jianming Xu", "Hongyan Tian", "Chongyi Wang", "Chi Chen", "Yuan Yao", "Zhiyuan Liu", "Maosong Sun (孙茂松)"], "date_basis": "Explicit publisher citation_publication_date: 2025/11. Preserve supplied precision; do not substitute conference event dates.", "date_parts": [2025, 11], "doi": "10.18653/v1/2025.emnlp-demos.12", "event_dates": null, "evidence_urls": ["https://aclanthology.org/2025.emnlp-demos.12/"], "identity_basis": "Same work: normalized title, ordered author correspondence and the specific contribution in the complete publisher abstract were reviewed against the retained arXiv identity.", "issue": null, "notes": "Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.", "pages": "155-180", "pdf_url": "https://aclanthology.org/2025.emnlp-demos.12.pdf", "publisher": "Association for Computational Linguistics", "series": null, "title": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning", "type": "conference-paper", "url": "https://aclanthology.org/2025.emnlp-demos.12/", "venue": "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations", "volume": null}, "published_at": "2025-06-02", "record_sha256": "87b956b3a31e410e687fd0329171648576009eaef38b92df8380e3d59cf2b1e4", "relations": [], "resource_id": "arxiv:2506.01391", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.01391v2", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2025-06-17"}, "text": "AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning\nResource type: paper\nReview status: curated\nScope: core\n\nCombines grounding pretraining, bilingual interaction trajectories, and reinforcement fine-tuning for mobile GUI actions.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation\nPublished: 2025-06-02\nAuthors: Zhong Zhang, Yaxi Lu, Yikun Fu, Yupeng Huo, Shenzhi Yang, Yesai Wu, Han Si, Xin Cong, Haotian Chen, Yankai Lin, Jie Xie, Wei Zhou, Wang Xu, Yuanheng Zhang, Zhou Su, Zhongwu Zhai, Xiaoming Liu, Yudong Mei, Jianming Xu, Hongyan Tian, Chongyi Wang, Chi Chen, Yuan Yao, Zhiyuan Liu, Maosong Sun\nRecorded arXiv version: v2\n\nPublication edition: AgentCPM-GUI: Building Mobile-Use Agents with Reinforcement Fine-Tuning\nPublication authors: Zhong Zhang, Yaxi Lu, Yikun Fu, Yupeng Huo, Shenzhi Yang, Yesai Wu, Han Si (司函), Xin Cong, Haotian Chen, Yankai Lin (林衍凯), Jie Xie, Wei Zhou, Wang Xu, Yuanheng Zhang, Zhou Su, Zhongwu Zhai, Xiaoming Liu, Yudong Mei, Jianming Xu, Hongyan Tian, Chongyi Wang, Chi Chen, Yuan Yao, Zhiyuan Liu, Maosong Sun (孙茂松)\nPublication edition date (recorded precision): 2025-11\nPublication venue: Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: System Demonstrations\nPublication publisher: Association for Computational Linguistics\nPublication pages: 155-180\nPublication doi: 10.18653/v1/2025.emnlp-demos.12\nPublication date basis: Explicit publisher citation_publication_date: 2025/11. Preserve supplied precision; do not substitute conference event dates.\nPublication notes: Publisher proceedings edition; retained separately from the arXiv snapshot. Publisher spelling, initials, name annotations or author lists differ from the retained preprint.\nPublication source: https://aclanthology.org/2025.emnlp-demos.12/\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.01391v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-01391/"}
{"id": "arxiv:2506.01616", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ec2467703bc5d0f97ef8d207289ed56f6d12b811688d20b7833eac06038757b", "submitted_at": "2025-06-02T12:56:27Z", "updated_at": "2025-06-02T12:56:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.01616v1"}, "authors": ["Xiao Yang", "Jiawei Chen", "Jun Luo", "Zhengwei Fang", "Yinpeng Dong", "Hang Su", "Jun Zhu"], "canonical_source_url": "https://arxiv.org/abs/2506.01616", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-01616/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ec2467703bc5d0f97ef8d207289ed56f6d12b811688d20b7833eac06038757b", "submitted_at": "2025-06-02T12:56:27Z", "updated_at": "2025-06-02T12:56:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.01616v1"}}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "d69da19edaed185ad3071a97db3bb3fcfbdb14c0b514b7d27794b39fbe8e2541", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md"}], "identifiers": {"arxiv": "2506.01616"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.01616"}], "name": "MLA-Trust: Benchmarking Trustworthiness of Multimodal LLM Agents in GUI Environments", "publication": null, "published_at": "2025-06-02", "record_sha256": "08e6c162540f7447bc9e75ff4ef9abd0d9f00f4bc24b63dec3561e1e070aec0a", "relations": [], "resource_id": "arxiv:2506.01616", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.01616v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2025-06-02"}, "text": "MLA-Trust: Benchmarking Trustworthiness of Multimodal LLM Agents in GUI Environments\nResource type: paper\nReview status: curated\nScope: core\n\nBenchmarks truthfulness, controllability, safety, and privacy of multimodal LLM agents across interactive web and mobile GUI tasks.\nPlatforms: web, android, cross-platform\nTopics: safety-security, evaluation, robustness\nPublished: 2025-06-02\nAuthors: Xiao Yang, Jiawei Chen, Jun Luo, Zhengwei Fang, Yinpeng Dong, Hang Su, Jun Zhu\nRecorded arXiv version: v1\n\nThe contribution was reviewed from the retained paper abstract; Android setup is supported separately by the official project README. No benchmark results were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.01616v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-01616/"}
{"id": "arxiv:2506.02456", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cd988217d1e3371bec827b472abf46c205ac407b5b60b7b581d7105f86e8bc5", "submitted_at": "2025-06-03T05:21:50Z", "updated_at": "2026-03-01T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02456v2"}, "authors": ["Tri Cao", "Bennett Lim", "Yue Liu", "Yuan Sui", "Yuexin Li", "Shumin Deng", "Lin Lu", "Nay Oo", "Shuicheng Yan", "Bryan Hooi"], "canonical_source_url": "https://arxiv.org/abs/2506.02456", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-02456/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cd988217d1e3371bec827b472abf46c205ac407b5b60b7b581d7105f86e8bc5", "submitted_at": "2025-06-03T05:21:50Z", "updated_at": "2026-03-01T16:12:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02456v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.02456"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02456"}], "name": "VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agents", "publication": null, "published_at": "2025-06-03", "record_sha256": "d56778f5d2d46aa15721b3790a5367543cf7d8314ad8c21d7dd167c2eca55d43", "relations": [], "resource_id": "arxiv:2506.02456", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.02456v2", "topics": ["safety-security", "evaluation"], "updated_at": "2026-03-01"}, "text": "VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VPI-Bench, interactive web-platform variants containing visually embedded malicious instructions. It evaluates computer-use and browser-use agents under these threats and examines system-prompt defenses.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-06-03\nAuthors: Tri Cao, Bennett Lim, Yue Liu, Yuan Sui, Yuexin Li, Shumin Deng, Lin Lu, Nay Oo, Shuicheng Yan, Bryan Hooi\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.02456v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-02456/"}
{"id": "arxiv:2506.02539", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "401ca18cadbf38045fa20985354988571b69e2280b1a090aff0e128cdf3c70e0", "submitted_at": "2025-06-03T07:25:49Z", "updated_at": "2025-08-07T21:48:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.02539v3"}, "authors": ["Thong Q. Nguyen", "Shubhang Desai", "Raja Hasnain Anwar", "Firoz Shaik", "Vishwas Suryanarayanan", "Vishal Chowdhary"], "canonical_source_url": "https://arxiv.org/abs/2506.02539", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-02539/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "401ca18cadbf38045fa20985354988571b69e2280b1a090aff0e128cdf3c70e0", "submitted_at": "2025-06-03T07:25:49Z", "updated_at": "2025-08-07T21:48:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.02539v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.02539"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02539"}], "name": "VerificAgent: Domain-Specific Memory Verification for Scalable Oversight of Aligned Computer-Use Agents", "publication": null, "published_at": "2025-06-03", "record_sha256": "46279538c725dccf3a4364ed19b8fb1756d03f23989927ea1b6ffeda9b24708f", "relations": [], "resource_id": "arxiv:2506.02539", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.02539v3", "topics": ["memory", "human-in-the-loop", "safety-security"], "updated_at": "2025-08-07"}, "text": "VerificAgent: Domain-Specific Memory Verification for Scalable Oversight of Aligned Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VerificAgent, which grows domain-specific memories from task trajectories and subjects them to human fact-checking before deployment. Its OSWorld evaluation treats persistent memory as an auditable constraint on later agent behavior.\nPlatforms: desktop\nTopics: memory, human-in-the-loop, safety-security\nPublished: 2025-06-03\nAuthors: Thong Q. Nguyen, Shubhang Desai, Raja Hasnain Anwar, Firoz Shaik, Vishwas Suryanarayanan, Vishal Chowdhary\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.02539v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-02539/"}
{"id": "arxiv:2506.02839", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bebdf1c8f34582408ef0050245f6b1e454d1bd2e75903a4e0817a4f3c00fb4db", "submitted_at": "2025-06-03T13:08:17Z", "updated_at": "2025-06-03T13:08:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.02839v1"}, "authors": ["Yougang Lyu", "Xiaoyu Zhang", "Lingyong Yan", "Maarten de Rijke", "Zhaochun Ren", "Xiuying Chen"], "canonical_source_url": "https://arxiv.org/abs/2506.02839", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-02839/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bebdf1c8f34582408ef0050245f6b1e454d1bd2e75903a4e0817a4f3c00fb4db", "submitted_at": "2025-06-03T13:08:17Z", "updated_at": "2025-06-03T13:08:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.02839v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2506.02839"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02839"}], "name": "DeepShop: A Benchmark for Deep Research Shopping Agents", "publication": null, "published_at": "2025-06-03", "record_sha256": "aaa6096bdef795b2e574737924890ee70f197d68dbde6df2f0e948c05ed75619", "relations": [], "resource_id": "arxiv:2506.02839", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.02839v1", "topics": ["evaluation", "web-navigation"], "updated_at": "2025-06-03"}, "text": "DeepShop: A Benchmark for Deep Research Shopping Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDeepShop evaluates shopping agents with evolving query complexity across product attributes, filters and sorting preferences. Its evaluation combines checks of these individual requirements with overall task success.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2025-06-03\nAuthors: Yougang Lyu, Xiaoyu Zhang, Lingyong Yan, Maarten de Rijke, Zhaochun Ren, Xiuying Chen\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.02839v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-02839/"}
{"id": "arxiv:2506.02865", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89c2d658e03204b56f6d273da9572820f9a23ca359a441ea87baa451a93c5b58", "submitted_at": "2025-06-03T13:29:03Z", "updated_at": "2025-06-11T09:33:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02865v2"}, "authors": ["Mathieu Andreux", "Breno Baldas Skuk", "Hamza Benchekroun", "Emilien Biré", "Antoine Bonnet", "Riaz Bordie", "Nathan Bout", "Matthias Brunel", "Pierre-Louis Cedoz", "Antoine Chassang", "Mickaël Chen", "Alexandra D. Constantinou", "Antoine d'Andigné", "Hubert de La Jonquière", "Aurélien Delfosse", "Ludovic Denoyer", "Alexis Deprez", "Augustin Derupti", "Michael Eickenberg", "Mathïs Federico", "Charles Kantor", "Xavier Koegler", "Yann Labbé", "Matthew C. H. Lee", "Erwan Le Jumeau de Kergaradec", "Amir Mahla", "Avshalom Manevich", "Adrien Maret", "Charles Masson", "Rafaël Maurin", "Arturo Mena", "Philippe Modard", "Axel Moyal", "Axel Nguyen Kerbel", "Julien Revelle", "Mats L. Richter", "María Santos", "Laurent Sifre", "Maxime Theillard", "Marc Thibault", "Louis Thiry", "Léo Tronchon", "Nicolas Usunier", "Tony Wu"], "canonical_source_url": "https://arxiv.org/abs/2506.02865", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-02865/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "89c2d658e03204b56f6d273da9572820f9a23ca359a441ea87baa451a93c5b58", "submitted_at": "2025-06-03T13:29:03Z", "updated_at": "2025-06-11T09:33:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.02865v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.02865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.02865"}], "name": "Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights", "publication": null, "published_at": "2025-06-03", "record_sha256": "c53eac2f066e81e9bb9044ed36f8163b824f77aa256acbb2a5aa4e1966b6f562", "relations": [], "resource_id": "arxiv:2506.02865", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.02865v2", "topics": ["web-navigation", "grounding", "data-generation", "evaluation"], "updated_at": "2025-06-11"}, "text": "Surfer-H Meets Holo1: Cost-Efficient Web Agent Powered by Open Weights\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes Surfer-H, a web agent paired with the Holo1 vision-language model family trained on web content, synthetic examples, and agent-generated data. The work introduces WebClick for web-interface localization and evaluates browser task execution on WebVoyager.\nPlatforms: web\nTopics: web-navigation, grounding, data-generation, evaluation\nPublished: 2025-06-03\nAuthors: Mathieu Andreux, Breno Baldas Skuk, Hamza Benchekroun, Emilien Biré, Antoine Bonnet, Riaz Bordie, Nathan Bout, Matthias Brunel, Pierre-Louis Cedoz, Antoine Chassang, Mickaël Chen, Alexandra D. Constantinou, Antoine d'Andigné, Hubert de La Jonquière, Aurélien Delfosse, Ludovic Denoyer, Alexis Deprez, Augustin Derupti, Michael Eickenberg, Mathïs Federico, Charles Kantor, Xavier Koegler, Yann Labbé, Matthew C. H. Lee, Erwan Le Jumeau de Kergaradec, Amir Mahla, Avshalom Manevich, Adrien Maret, Charles Masson, Rafaël Maurin, Arturo Mena, Philippe Modard, Axel Moyal, Axel Nguyen Kerbel, Julien Revelle, Mats L. Richter, María Santos, Laurent Sifre, Maxime Theillard, Marc Thibault, Louis Thiry, Léo Tronchon, Nicolas Usunier, Tony Wu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.02865v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-02865/"}
{"id": "arxiv:2506.03095", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "068eae2cab9899967850f21616a701f3eeb258b9654c83d34f68be266e48854e", "submitted_at": "2025-06-03T17:27:04Z", "updated_at": "2025-06-03T17:27:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03095v1"}, "authors": ["Man Luo", "David Cobbley", "Xin Su", "Shachar Rosenman", "Vasudev Lal", "Shao-Yen Tseng", "Phillip Howard"], "canonical_source_url": "https://arxiv.org/abs/2506.03095", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-03095/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "068eae2cab9899967850f21616a701f3eeb258b9654c83d34f68be266e48854e", "submitted_at": "2025-06-03T17:27:04Z", "updated_at": "2025-06-03T17:27:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03095v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.03095"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03095"}], "name": "DPO Learning with LLMs-Judge Signal for Computer Use Agents", "publication": null, "published_at": "2025-06-03", "record_sha256": "7ee0c358fe3219fb8959df5d6e40a723a932e2d35f919d148b6a7e9b8790baf2", "relations": [], "resource_id": "arxiv:2506.03095", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.03095v1", "topics": ["data-generation", "evaluation", "efficiency"], "updated_at": "2025-06-03"}, "text": "DPO Learning with LLMs-Judge Signal for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses an LLM judge to evaluate and filter synthetic GUI interaction trajectories for training a compact vision-language agent. OSWorld experiments assess the policy designed for local execution.\nPlatforms: desktop\nTopics: data-generation, evaluation, efficiency\nPublished: 2025-06-03\nAuthors: Man Luo, David Cobbley, Xin Su, Shachar Rosenman, Vasudev Lal, Shao-Yen Tseng, Phillip Howard\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.03095v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-03095/"}
{"id": "arxiv:2506.03143", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b27d6caaaa3802629fdbf1e7bbe5557ed567f840cee88e45ba7e1fe12c548d", "submitted_at": "2025-06-03T17:59:08Z", "updated_at": "2025-06-03T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03143v1"}, "authors": ["Qianhui Wu", "Kanzhi Cheng", "Rui Yang", "Chaoyun Zhang", "Jianwei Yang", "Huiqiang Jiang", "Jian Mu", "Baolin Peng", "Bo Qiao", "Reuben Tan", "Si Qin", "Lars Liden", "Qingwei Lin", "Huan Zhang", "Tong Zhang", "Jianbing Zhang", "Dongmei Zhang", "Jianfeng Gao"], "canonical_source_url": "https://arxiv.org/abs/2506.03143", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-03143/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b27d6caaaa3802629fdbf1e7bbe5557ed567f840cee88e45ba7e1fe12c548d", "submitted_at": "2025-06-03T17:59:08Z", "updated_at": "2025-06-03T17:59:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.03143v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.03143"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03143"}], "name": "GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents", "publication": null, "published_at": "2025-06-03", "record_sha256": "57b7634d9027251287097211a7d45dc1cc365d996b759c449b515c2f0270272c", "relations": [], "resource_id": "arxiv:2506.03143", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.03143v1", "topics": ["grounding"], "updated_at": "2025-06-03"}, "text": "GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses an attention-based action head and a verifier to select relevant screen regions without generating coordinates as text.\nPlatforms: cross-platform\nTopics: grounding\nPublished: 2025-06-03\nAuthors: Qianhui Wu, Kanzhi Cheng, Rui Yang, Chaoyun Zhang, Jianwei Yang, Huiqiang Jiang, Jian Mu, Baolin Peng, Bo Qiao, Reuben Tan, Si Qin, Lars Liden, Qingwei Lin, Huan Zhang, Tong Zhang, Jianbing Zhang, Dongmei Zhang, Jianfeng Gao\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.03143v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-03143/"}
{"id": "arxiv:2506.03533", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bccee69e390b5170ff264e689c209d13317e7d3be9b37e6d3fbed0e0d089943", "submitted_at": "2025-06-04T03:27:56Z", "updated_at": "2026-03-02T22:31:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.03533v2"}, "authors": ["Apurva Gandhi", "Graham Neubig"], "canonical_source_url": "https://arxiv.org/abs/2506.03533", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-03533/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bccee69e390b5170ff264e689c209d13317e7d3be9b37e6d3fbed0e0d089943", "submitted_at": "2025-06-04T03:27:56Z", "updated_at": "2026-03-02T22:31:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.03533v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.03533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.03533"}], "name": "Go-Browse: Training Web Agents with Structured Exploration", "publication": null, "published_at": "2025-06-04", "record_sha256": "71c5c6b5dcf30027f709226746ad1361b0805f238e8c09864f9089f0b79f3d0a", "relations": [], "resource_id": "arxiv:2506.03533", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.03533v2", "topics": ["data-generation", "web-navigation"], "updated_at": "2026-03-02"}, "text": "Go-Browse: Training Web Agents with Structured Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nGo-Browse frames web-environment exploration as graph search so that information can be reused across data-collection episodes. It collects 10,000 successful trajectories and 40,000 interaction steps on WebArena and uses them to fine-tune a web agent.\nPlatforms: web\nTopics: data-generation, web-navigation\nPublished: 2025-06-04\nAuthors: Apurva Gandhi, Graham Neubig\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.03533v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-03533/"}
{"id": "arxiv:2506.04135", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4c980694819dc5eb4ff322e6a703cccfe0b1e45efb60f22b120592d03edf36", "submitted_at": "2025-06-04T16:26:56Z", "updated_at": "2025-10-18T12:11:31Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.04135v4"}, "authors": ["Pei Yang", "Hai Ci", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2506.04135", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-04135/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a4c980694819dc5eb4ff322e6a703cccfe0b1e45efb60f22b120592d03edf36", "submitted_at": "2025-06-04T16:26:56Z", "updated_at": "2025-10-18T12:11:31Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.04135v4"}}, "environments": ["macos", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.04135"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.04135"}], "name": "macOSWorld: A Multilingual Interactive Benchmark for GUI Agents", "publication": null, "published_at": "2025-06-04", "record_sha256": "c7073edbb095fedb66e63561aec498ec7c1df3376107e8cad0541320c659058c", "relations": [], "resource_id": "arxiv:2506.04135", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.04135v4", "topics": ["evaluation", "workflow", "safety-security"], "updated_at": "2025-10-18"}, "text": "macOSWorld: A Multilingual Interactive Benchmark for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nmacOSWorld provides 202 interactive tasks across 30 applications, with instructions and operating-system interfaces in five languages. It evaluates GUI agents on macOS workflows and includes a separate subset for deception-attack testing.\nPlatforms: macos, desktop\nTopics: evaluation, workflow, safety-security\nPublished: 2025-06-04\nAuthors: Pei Yang, Hai Ci, Mike Zheng Shou\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.04135v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-04135/"}
{"id": "arxiv:2506.04614", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42be944224a9b7d2c8e7042968ab31d373a0bafa304c0e669728cb0f10cc813d", "submitted_at": "2025-06-05T04:12:36Z", "updated_at": "2025-11-17T07:48:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.04614v2"}, "authors": ["Yuyang Wanyan", "Xi Zhang", "Haiyang Xu", "Haowei Liu", "Junyang Wang", "Jiabo Ye", "Yutong Kou", "Ming Yan", "Fei Huang", "Xiaoshan Yang", "Weiming Dong", "Changsheng Xu"], "canonical_source_url": "https://arxiv.org/abs/2506.04614", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-04614/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42be944224a9b7d2c8e7042968ab31d373a0bafa304c0e669728cb0f10cc813d", "submitted_at": "2025-06-05T04:12:36Z", "updated_at": "2025-11-17T07:48:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.04614v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2506.04614"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.04614"}], "name": "Look Before You Leap: A GUI-Critic-R1 Model for Pre-Operative Error Diagnosis in GUI Automation", "publication": null, "published_at": "2025-06-05", "record_sha256": "c0ae939a02d5eb2286070b21cd18585743546f062189e0c622f99103c466476c", "relations": [], "resource_id": "arxiv:2506.04614", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.04614v2", "topics": ["planning", "robustness"], "updated_at": "2025-11-17"}, "text": "Look Before You Leap: A GUI-Critic-R1 Model for Pre-Operative Error Diagnosis in GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nUses a pre-execution critic and reinforcement training to evaluate proposed GUI actions and provide corrective suggestions before execution.\nPlatforms: cross-platform\nTopics: planning, robustness\nPublished: 2025-06-05\nAuthors: Yuyang Wanyan, Xi Zhang, Haiyang Xu, Haowei Liu, Junyang Wang, Jiabo Ye, Yutong Kou, Ming Yan, Fei Huang, Xiaoshan Yang, Weiming Dong, Changsheng Xu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.04614v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-04614/"}
{"id": "arxiv:2506.05079", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bad447e8bca06f720a9476492de06978c73a356a38b00671203f8af0b81f1952", "submitted_at": "2025-06-05T14:27:40Z", "updated_at": "2026-05-03T08:25:53Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.05079v4"}, "authors": ["Shengcheng Yu", "Yuchen Ling", "Chunrong Fang", "Quan Zhou", "Yi Zhao", "Chunyang Chen", "Shaomin Zhu", "Zhenyu Chen"], "canonical_source_url": "https://arxiv.org/abs/2506.05079", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-05079/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bad447e8bca06f720a9476492de06978c73a356a38b00671203f8af0b81f1952", "submitted_at": "2025-06-05T14:27:40Z", "updated_at": "2026-05-03T08:25:53Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.05079v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2506.05079"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.05079"}], "name": "Scenario-Guided LLM-based Mobile App GUI Testing", "publication": null, "published_at": "2025-06-05", "record_sha256": "8759026049e9b6cdf6124f92bcdb6d8d6d7dc6a6b468b633774da3038c0cbe49", "relations": [], "resource_id": "arxiv:2506.05079", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.05079v4", "topics": ["automation", "mobile-navigation", "planning"], "updated_at": "2026-05-03"}, "text": "Scenario-Guided LLM-based Mobile App GUI Testing\nResource type: paper\nReview status: curated\nScope: core\n\nScenGen organizes mobile GUI testing around business scenarios rather than unrestricted exploration. Five collaborating agents observe the interface, choose and execute actions, check scenario completion, and record operations and runtime bugs.\nPlatforms: unspecified\nTopics: automation, mobile-navigation, planning\nPublished: 2025-06-05\nAuthors: Shengcheng Yu, Yuchen Ling, Chunrong Fang, Quan Zhou, Yi Zhao, Chunyang Chen, Shaomin Zhu, Zhenyu Chen\nRecorded arXiv version: v4\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.05079v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-05079/"}
{"id": "arxiv:2506.07217", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bf4f91de32fdf3bd9b81ba64ace02e4a7d02ff094794f30dfd560389f334cea", "submitted_at": "2025-06-08T16:45:31Z", "updated_at": "2025-06-30T08:31:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07217v2"}, "authors": ["Zihan Deng", "Changyu Du", "Stavros Nousias", "André Borrmann"], "canonical_source_url": "https://arxiv.org/abs/2506.07217", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-07217/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bf4f91de32fdf3bd9b81ba64ace02e4a7d02ff094794f30dfd560389f334cea", "submitted_at": "2025-06-08T16:45:31Z", "updated_at": "2025-06-30T08:31:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07217v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.07217"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07217"}], "name": "BIMgent: Towards Autonomous Building Modeling via Computer-use Agents", "publication": null, "published_at": "2025-06-08", "record_sha256": "4fa8db62fd2683ad8fb37580c1b7da93d2bb96413f879c59c0f1963ebaa9a77c", "relations": [], "resource_id": "arxiv:2506.07217", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.07217v2", "topics": ["industry", "planning", "workflow", "evaluation"], "updated_at": "2025-06-30"}, "text": "BIMgent: Towards Autonomous Building Modeling via Computer-use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nBIMgent plans software-specific workflows and executes GUI actions for building-information modeling. It evaluates multimodal building design and reconstruction tasks in professional authoring software.\nPlatforms: unspecified\nTopics: industry, planning, workflow, evaluation\nPublished: 2025-06-08\nAuthors: Zihan Deng, Changyu Du, Stavros Nousias, André Borrmann\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.07217v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-07217/"}
{"id": "arxiv:2506.07385", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bea3e76b3124ebbcd8f7817b3e32d667d8aaae6e31f1b75ac3fc8c39e7637c6f", "submitted_at": "2025-06-09T03:09:48Z", "updated_at": "2025-06-09T03:09:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07385v1"}, "authors": ["Ruofan Liu", "Xiwen Teoh", "Yun Lin", "Guanjie Chen", "Ruofei Ren", "Denys Poshyvanyk", "Jin Song Dong"], "canonical_source_url": "https://arxiv.org/abs/2506.07385", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-07385/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bea3e76b3124ebbcd8f7817b3e32d667d8aaae6e31f1b75ac3fc8c39e7637c6f", "submitted_at": "2025-06-09T03:09:48Z", "updated_at": "2025-06-09T03:09:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:30Z", "role": "primary", "sha256": "23030fd266519162c51c7f4a87bc38e75bb171ef7fa95e30c2d7afd820a0bc3f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2506.07385v1"}], "identifiers": {"arxiv": "2506.07385"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07385"}], "name": "GUIPilot: A Consistency-based Mobile GUI Testing Approach for Detecting Application-specific Bugs", "publication": null, "published_at": "2025-06-09", "record_sha256": "099c09556e6ed5adf74cdf3c2fd1453e6e7bb1956eec37bdac3773fa6b163600", "relations": [], "resource_id": "arxiv:2506.07385", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.07385v1", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2025-06-09"}, "text": "GUIPilot: A Consistency-based Mobile GUI Testing Approach for Detecting Application-specific Bugs\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUIPilot, which compares mobile design mock-ups with implemented screens through widget alignment and tests expected transitions by executing inferred UI actions. It evaluates both visual inconsistencies and workflow-specific bugs.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, robustness\nPublished: 2025-06-09\nAuthors: Ruofan Liu, Xiwen Teoh, Yun Lin, Guanjie Chen, Ruofei Ren, Denys Poshyvanyk, Jin Song Dong\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.07385v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-07385/"}
{"id": "arxiv:2506.07672", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2334e017fe49b39dc5d460963988f05d11051f2101bebe23be76005809fafd45", "submitted_at": "2025-06-09T11:50:33Z", "updated_at": "2025-06-09T11:50:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07672v1"}, "authors": ["Yunhe Yan", "Shihe Wang", "Jiajun Du", "Yexuan Yang", "Yuxuan Shan", "Qichen Qiu", "Xianqing Jia", "Xinge Wang", "Xin Yuan", "Xu Han", "Mao Qin", "Yinxiao Chen", "Chen Peng", "Shangguang Wang", "Mengwei Xu"], "canonical_source_url": "https://arxiv.org/abs/2506.07672", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-07672/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2334e017fe49b39dc5d460963988f05d11051f2101bebe23be76005809fafd45", "submitted_at": "2025-06-09T11:50:33Z", "updated_at": "2025-06-09T11:50:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.07672v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:12:09Z", "role": "primary", "sha256": "f64464b1516369a87f0fbe28828e32ff61687dae4ce063aedddd8bd45a7d40eb", "supports": ["scope", "summary", "environments"], "url": "https://arxiv.org/html/2506.07672"}], "identifiers": {"arxiv": "2506.07672"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07672"}], "name": "MCPWorld: A Unified Benchmarking Testbed for API, GUI, and Hybrid Computer Use Agents", "publication": null, "published_at": "2025-06-09", "record_sha256": "f0091353d4911e325da66191ee2af7f280671e0c948a81ba4236a23bb2bb0d1d", "relations": [], "resource_id": "arxiv:2506.07672", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.07672v1", "topics": ["evaluation", "tool-integration", "sandbox"], "updated_at": "2025-06-09"}, "text": "MCPWorld: A Unified Benchmarking Testbed for API, GUI, and Hybrid Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMCPWorld evaluates API, GUI, and hybrid computer-use agents with 201 tasks in applications whose source can be instrumented and extended with MCP support. It verifies completion from application behavior and packages the testbed in containers with GPU support.\nPlatforms: desktop\nTopics: evaluation, tool-integration, sandbox\nPublished: 2025-06-09\nAuthors: Yunhe Yan, Shihe Wang, Jiajun Du, Yexuan Yang, Yuxuan Shan, Qichen Qiu, Xianqing Jia, Xinge Wang, Xin Yuan, Xu Han, Mao Qin, Yinxiao Chen, Chen Peng, Shangguang Wang, Mengwei Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. The platform or GUI evaluation setting was additionally checked in the primary full text.\n\nPrimary source: https://arxiv.org/abs/2506.07672v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-07672/"}
{"id": "arxiv:2506.07976", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4b6c8b297e88dae689ba2a48724a441698cd5309a3bbe1198f38343e88aceb1", "submitted_at": "2025-06-09T17:50:02Z", "updated_at": "2025-06-10T12:50:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07976v2"}, "authors": ["Junhong Shen", "Hao Bai", "Lunjun Zhang", "Yifei Zhou", "Amrith Setlur", "Shengbang Tong", "Diego Caples", "Nan Jiang", "Tong Zhang", "Ameet Talwalkar", "Aviral Kumar"], "canonical_source_url": "https://arxiv.org/abs/2506.07976", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-07976/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e4b6c8b297e88dae689ba2a48724a441698cd5309a3bbe1198f38343e88aceb1", "submitted_at": "2025-06-09T17:50:02Z", "updated_at": "2025-06-10T12:50:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.07976v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2506.07976"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.07976"}], "name": "Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction", "publication": null, "published_at": "2025-06-09", "record_sha256": "3198b48c9eeb240d2021aead9c238c49e2ee16f266d553bf901b022f160dfc4d", "relations": [], "resource_id": "arxiv:2506.07976", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.07976v2", "topics": ["reinforcement-learning", "web-navigation", "planning", "long-horizon"], "updated_at": "2025-06-10"}, "text": "Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nTest-Time Interaction scales the interaction horizon so web agents can explore, backtrack and revise plans within a rollout. Its online reinforcement-learning curriculum adjusts rollout lengths and is evaluated on WebVoyager and WebArena.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, planning, long-horizon\nPublished: 2025-06-09\nAuthors: Junhong Shen, Hao Bai, Lunjun Zhang, Yifei Zhou, Amrith Setlur, Shengbang Tong, Diego Caples, Nan Jiang, Tong Zhang, Ameet Talwalkar, Aviral Kumar\nRecorded arXiv version: v2\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.07976v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-07976/"}
{"id": "arxiv:2506.08012", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce3b7d2c25b5a99f110f00be7389d4230d625ac1f7f33d5ab81f3d6f441f3610", "submitted_at": "2025-06-09T17:59:57Z", "updated_at": "2025-06-09T17:59:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08012v1"}, "authors": ["Penghao Wu", "Shengnan Ma", "Bo Wang", "Jiaheng Yu", "Lewei Lu", "Ziwei Liu"], "canonical_source_url": "https://arxiv.org/abs/2506.08012", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-08012/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce3b7d2c25b5a99f110f00be7389d4230d625ac1f7f33d5ab81f3d6f441f3610", "submitted_at": "2025-06-09T17:59:57Z", "updated_at": "2025-06-09T17:59:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08012v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2506.08012"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08012"}], "name": "GUI-Reflection: Empowering Multimodal GUI Models with Self-Reflection Behavior", "publication": null, "published_at": "2025-06-09", "record_sha256": "fa4484f3d74322c98c7895aa9cb486c29d1bd44b7e49c15e8d1a6fffb66269e4", "relations": [], "resource_id": "arxiv:2506.08012", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.08012v1", "topics": ["robustness", "data-generation", "evaluation", "mobile-navigation"], "updated_at": "2025-06-09"}, "text": "GUI-Reflection: Empowering Multimodal GUI Models with Self-Reflection Behavior\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Reflection synthesizes reflection and error-correction examples from successful trajectories and adds dedicated learning stages for these behaviors. It introduces a reflection task suite, a mobile environment for online data collection, and iterative reflection tuning.\nPlatforms: unspecified\nTopics: robustness, data-generation, evaluation, mobile-navigation\nPublished: 2025-06-09\nAuthors: Penghao Wu, Shengnan Ma, Bo Wang, Jiaheng Yu, Lewei Lu, Ziwei Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.08012v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-08012/"}
{"id": "arxiv:2506.08136", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43aca3b0b8a1a0d5729a4e90dc1516bbb68e3611de3220876cdb29afc81d847f", "submitted_at": "2025-06-09T18:39:48Z", "updated_at": "2026-05-11T08:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.08136v3"}, "authors": ["Zefang Liu", "Yinzhu Quan"], "canonical_source_url": "https://arxiv.org/abs/2506.08136", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-08136/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43aca3b0b8a1a0d5729a4e90dc1516bbb68e3611de3220876cdb29afc81d847f", "submitted_at": "2025-06-09T18:39:48Z", "updated_at": "2026-05-11T08:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.08136v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:34Z", "role": "primary", "sha256": "2f1d53911fe997bbba01c9d8a2fa7002f18837492448e15056fd7a7a6c78f97d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2506.08136v3"}], "identifiers": {"arxiv": "2506.08136"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08136"}], "name": "EconWebArena: Benchmarking Autonomous Agents on Economic Tasks in Realistic Web Environments", "publication": null, "published_at": "2025-06-09", "record_sha256": "fd79794a447a6379032cfb03dcc6f25141524f14dd223bf9af96bcab13331076", "relations": [], "resource_id": "arxiv:2506.08136", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.08136v3", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2026-05-11"}, "text": "EconWebArena: Benchmarking Autonomous Agents on Economic Tasks in Realistic Web Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces EconWebArena, a curated benchmark of economic-data tasks requiring live website navigation, multimodal interpretation, and precise extraction. It evaluates visual grounding, planning, and interaction design within multistep workflows.\nPlatforms: web\nTopics: evaluation, web-navigation, planning\nPublished: 2025-06-09\nAuthors: Zefang Liu, Yinzhu Quan\nRecorded arXiv version: v3\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.08136v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-08136/"}
{"id": "arxiv:2506.08972", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8dbc84aa582b94775d490cf84e17eff05be84af7dbcc2de7de61b886fa3c9178", "submitted_at": "2025-06-10T16:45:29Z", "updated_at": "2025-06-10T16:45:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08972v1"}, "authors": ["Yuan Guo", "Tingjia Miao", "Zheng Wu", "Pengzhou Cheng", "Ming Zhou", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2506.08972", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-08972/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8dbc84aa582b94775d490cf84e17eff05be84af7dbcc2de7de61b886fa3c9178", "submitted_at": "2025-06-10T16:45:29Z", "updated_at": "2025-06-10T16:45:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.08972v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}], "identifiers": {"arxiv": "2506.08972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:42Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.08972"}], "name": "Atomic-to-Compositional Generalization for Mobile Agents with A New Benchmark and Scheduling System", "publication": null, "published_at": "2025-06-10", "record_sha256": "1b1970245125087eda8d079ee98c5b86c335c916527265d3b61281fa0cf4f366", "relations": [], "resource_id": "arxiv:2506.08972", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.08972v1", "topics": ["evaluation", "long-horizon", "planning", "mobile-navigation"], "updated_at": "2025-06-10"}, "text": "Atomic-to-Compositional Generalization for Mobile Agents with A New Benchmark and Scheduling System\nResource type: paper\nReview status: curated\nScope: core\n\nUI-NEXUS evaluates compositional mobile tasks in local utility apps and online Chinese and English service apps. AGENT-NEXUS decomposes those long tasks into self-contained atomic subtasks to study generalization beyond isolated screens or short action chains.\nPlatforms: unspecified\nTopics: evaluation, long-horizon, planning, mobile-navigation\nPublished: 2025-06-10\nAuthors: Yuan Guo, Tingjia Miao, Zheng Wu, Pengzhou Cheng, Ming Zhou, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.08972v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-08972/"}
{"id": "arxiv:2506.09373", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b00deba927bda03c079621f70decdeb946ec51a6982531a2086c43033ac9925c", "submitted_at": "2025-06-11T03:43:30Z", "updated_at": "2026-04-21T11:50:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.09373v3"}, "authors": ["Jiaqi Tang", "Yu Xia", "Yi-Feng Wu", "Yuwei Hu", "Yuhui Chen", "Qing-Guo Chen", "Xiaogang Xu", "Xiangyu Wu", "Hao Lu", "Yanqing Ma", "Shiyin Lu", "Qifeng Chen"], "canonical_source_url": "https://arxiv.org/abs/2506.09373", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-09373/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b00deba927bda03c079621f70decdeb946ec51a6982531a2086c43033ac9925c", "submitted_at": "2025-06-11T03:43:30Z", "updated_at": "2026-04-21T11:50:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.09373v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.09373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.09373"}], "name": "LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization", "publication": null, "published_at": "2025-06-11", "record_sha256": "71ac154dd39b347f534ea28ce176cbaf257dd563110b0f66d9048732587824d5", "relations": [], "resource_id": "arxiv:2506.09373", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.09373v3", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-21"}, "text": "LPO: Towards Accurate GUI Agent Interaction via Location Preference Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nLocation Preference Optimization uses an information-entropy signal to select interaction regions and a distance-based reward to assess click locations. It trains GUI localization through preference optimization and evaluates both offline grounding and online interaction.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning\nPublished: 2025-06-11\nAuthors: Jiaqi Tang, Yu Xia, Yi-Feng Wu, Yuwei Hu, Yuhui Chen, Qing-Guo Chen, Xiaogang Xu, Xiangyu Wu, Hao Lu, Yanqing Ma, Shiyin Lu, Qifeng Chen\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.09373v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-09373/"}
{"id": "arxiv:2506.10387", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b353606ce29a5bc60d263e86827a825af46883e25da5bd290b48affa9dd3838", "submitted_at": "2025-06-12T06:21:19Z", "updated_at": "2025-06-12T06:21:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.10387v1"}, "authors": ["Yuquan Xie", "Zaijing Li", "Rui Shao", "Gongwei Chen", "Kaiwen Zhou", "Yinchuan Li", "Dongmei Jiang", "Liqiang Nie"], "canonical_source_url": "https://arxiv.org/abs/2506.10387", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-10387/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b353606ce29a5bc60d263e86827a825af46883e25da5bd290b48affa9dd3838", "submitted_at": "2025-06-12T06:21:19Z", "updated_at": "2025-06-12T06:21:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.10387v1"}}, "environments": ["android", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.10387"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.10387"}], "name": "Mirage-1: Augmenting and Updating GUI Agent with Hierarchical Multimodal Skills", "publication": null, "published_at": "2025-06-12", "record_sha256": "2c33fbd0eb15469545ee51d0bf0dfcd9b1544de563a97933fb8a0b3aaffdcfd6", "relations": [], "resource_id": "arxiv:2506.10387", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.10387v1", "topics": ["planning", "memory", "long-horizon", "mobile-navigation", "web-navigation"], "updated_at": "2025-06-12"}, "text": "Mirage-1: Augmenting and Updating GUI Agent with Hierarchical Multimodal Skills\nResource type: paper\nReview status: curated\nScope: core\n\nMirage-1 organizes interaction trajectories into execution, core, and meta-skills, then uses them to guide Monte Carlo tree search during online GUI tasks. It introduces AndroidLH for long-horizon evaluation and tests the approach on Android and web-navigation benchmarks.\nPlatforms: android, web, cross-platform\nTopics: planning, memory, long-horizon, mobile-navigation, web-navigation\nPublished: 2025-06-12\nAuthors: Yuquan Xie, Zaijing Li, Rui Shao, Gongwei Chen, Kaiwen Zhou, Yinchuan Li, Dongmei Jiang, Liqiang Nie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.10387v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-10387/"}
{"id": "arxiv:2506.11127", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cce7c2e1623ec841e8afed03c2e8e0c6f969f3dd05d53ddc4ef11e74310c21c", "submitted_at": "2025-06-10T12:16:27Z", "updated_at": "2025-11-26T07:48:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.11127v3"}, "authors": ["Wenkang Han", "Zhixiong Zeng", "Jing Huang", "Shu Jiang", "Liming Zheng", "Longrong Yang", "Haibo Qiu", "Chang Yao", "Jingyuan Chen", "Lin Ma"], "canonical_source_url": "https://arxiv.org/abs/2506.11127", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-11127/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3cce7c2e1623ec841e8afed03c2e8e0c6f969f3dd05d53ddc4ef11e74310c21c", "submitted_at": "2025-06-10T12:16:27Z", "updated_at": "2025-11-26T07:48:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.11127v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.11127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.11127"}], "name": "UITron-Speech: Towards Automated GUI Agents Based on Speech Instructions", "publication": null, "published_at": "2025-06-10", "record_sha256": "94b764d8df598944a510d5ad9647544f4da2a3f95867c8af87ebdce85929e1d7", "relations": [], "resource_id": "arxiv:2506.11127", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.11127v3", "topics": ["grounding", "action-prediction", "data-generation"], "updated_at": "2025-11-26"}, "text": "UITron-Speech: Towards Automated GUI Agents Based on Speech Instructions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UITron-Speech, which predicts GUI actions from spoken instructions and screenshots. It combines synthetic speech data and mixed-modality training with a training-free grounding refinement step.\nPlatforms: unspecified\nTopics: grounding, action-prediction, data-generation\nPublished: 2025-06-10\nAuthors: Wenkang Han, Zhixiong Zeng, Jing Huang, Shu Jiang, Liming Zheng, Longrong Yang, Haibo Qiu, Chang Yao, Jingyuan Chen, Lin Ma\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.11127v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-11127/"}
{"id": "arxiv:2506.13205", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390304ce515886fa15fb8cd74d8b2df161a75b9adc0f4d2fad1f65976c6a6ce7", "submitted_at": "2025-06-16T08:09:32Z", "updated_at": "2025-09-05T14:19:03Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2506.13205v6"}, "authors": ["Xuan Wang", "Siyuan Liang", "Zhe Liu", "Yi Yu", "Aishan Liu", "Yuliang Lu", "Xitong Gao", "Ee-Chien Chang"], "canonical_source_url": "https://arxiv.org/abs/2506.13205", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-13205/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "390304ce515886fa15fb8cd74d8b2df161a75b9adc0f4d2fad1f65976c6a6ce7", "submitted_at": "2025-06-16T08:09:32Z", "updated_at": "2025-09-05T14:19:03Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2506.13205v6"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:08Z", "role": "primary", "sha256": "00d275cd5407cec847fa9d43fe73cfc4384c5570b68a5e058c97921799be0bec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2302.01560%2C2303.16563%2C2305.15486%2C2305.16291%2C2305.17144%2C2306.00003%2C2306.00349%2C2306.17882%2C2310.01557%2C2310.08235%2C2310.08367%2C2311.05997%2C2311.17600%2C2312.02519%2C2312.11865%2C2402.01118%2C2402.10196%2C2402.11208%2C2403.10249%2C2403.12037%2C2404.06647%2C2404.12241%2C2404.13246%2C2407.15325%2C2409.00138%2C2410.14573%2C2410.18927%2C2410.19012%2C2411.09523%2C2411.10349%2C2411.13543%2C2412.01547%2C2501.05366%2C2501.09004%2C2502.07463%2C2502.14538%2C2503.02732%2C2503.08175%2C2503.11387%2C2504.10147%2C2504.11703%2C2504.14064%2C2504.15810%2C2504.16208%2C2504.20635%2C2504.21776%2C2505.12981%2C2505.13076%2C2506.01616%2C2506.13205&max_results=50"}], "identifiers": {"arxiv": "2506.13205"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:08Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.13205"}], "name": "Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents", "publication": null, "published_at": "2025-06-16", "record_sha256": "06be03c46b80290cae102f2e8a6d5e7e99e9e191f48a8983e7c6231bdcf4c9cd", "relations": [], "resource_id": "arxiv:2506.13205", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.13205v6", "topics": ["safety-security", "mobile-navigation", "robustness"], "updated_at": "2025-09-05"}, "text": "Poison Once, Control Anywhere: Clean-Text Visual Backdoors in VLM-based Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDemonstrates clean-text visual backdoor attacks (VIBMA) that compromise fine-tuned VLM mobile agents through visual triggers while keeping text instructions intact.\nPlatforms: android\nTopics: safety-security, mobile-navigation, robustness\nPublished: 2025-06-16\nAuthors: Xuan Wang, Siyuan Liang, Zhe Liu, Yi Yu, Aishan Liu, Yuliang Lu, Xitong Gao, Ee-Chien Chang\nRecorded arXiv version: v6\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.13205v6\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-13205/"}
{"id": "arxiv:2506.14205", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae36b271f9b4799093831cabee772845edfaeca97107414e424e170ccf89904d", "submitted_at": "2025-06-17T05:46:52Z", "updated_at": "2026-03-02T00:51:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14205v2"}, "authors": ["Jingxu Xie", "Dylan Xu", "Xuandong Zhao", "Dawn Song"], "canonical_source_url": "https://arxiv.org/abs/2506.14205", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-14205/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae36b271f9b4799093831cabee772845edfaeca97107414e424e170ccf89904d", "submitted_at": "2025-06-17T05:46:52Z", "updated_at": "2026-03-02T00:51:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14205v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:13+00:00", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["scope", "environments"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}], "identifiers": {"arxiv": "2506.14205"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14205"}], "name": "AgentSynth: Scalable Task Generation for Generalist Computer-Use Agents", "publication": null, "published_at": "2025-06-17", "record_sha256": "cfb74bbe80378fcf5418272160ade3ec0ca3936292b28eda63294a62faef476d", "relations": [], "resource_id": "arxiv:2506.14205", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.14205v2", "topics": ["data-generation", "evaluation", "long-horizon"], "updated_at": "2026-03-02"}, "text": "AgentSynth: Scalable Task Generation for Generalist Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAgentSynth composes easily generated subtasks into longer computer-use tasks and trajectories with adjustable difficulty. Its evaluation uses the number of subtasks to probe how agent success changes as task complexity grows.\nPlatforms: desktop, web\nTopics: data-generation, evaluation, long-horizon\nPublished: 2025-06-17\nAuthors: Jingxu Xie, Dylan Xu, Xuandong Zhao, Dawn Song\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.14205v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-14205/"}
{"id": "arxiv:2506.14477", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9e64118f03c2a9f3f55e35f79c38835f1eb823265c83c933fdceca2610b9904", "submitted_at": "2025-06-17T12:50:35Z", "updated_at": "2025-06-17T12:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.14477v1"}, "authors": ["Jingqi Yang", "Zhilong Song", "Jiawei Chen", "Mingli Song", "Sheng Zhou", "linjun sun", "Xiaogang Ouyang", "Chun Chen", "Can Wang"], "canonical_source_url": "https://arxiv.org/abs/2506.14477", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-14477/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9e64118f03c2a9f3f55e35f79c38835f1eb823265c83c933fdceca2610b9904", "submitted_at": "2025-06-17T12:50:35Z", "updated_at": "2025-06-17T12:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.14477v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.14477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14477"}], "name": "GUI-Robust: A Comprehensive Dataset for Testing GUI Agent Robustness in Real-World Anomalies", "publication": null, "published_at": "2025-06-17", "record_sha256": "c1a709329580857c120cd0b153ca4ce50a6139d906a7743408938d66ebe47b98", "relations": [], "resource_id": "arxiv:2506.14477", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.14477v1", "topics": ["robustness", "evaluation", "data-generation"], "updated_at": "2025-06-17"}, "text": "GUI-Robust: A Comprehensive Dataset for Testing GUI Agent Robustness in Real-World Anomalies\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds GUI task data with real-world interaction anomalies and evaluates how agent behavior changes under those conditions.\nPlatforms: cross-platform\nTopics: robustness, evaluation, data-generation\nPublished: 2025-06-17\nAuthors: Jingqi Yang, Zhilong Song, Jiawei Chen, Mingli Song, Sheng Zhou, linjun sun, Xiaogang Ouyang, Chun Chen, Can Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.14477v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-14477/"}
{"id": "arxiv:2506.14866", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f551b861012fc4eddbf010a619bc65e77b524b05958761b738a88bcae377e9a9", "submitted_at": "2025-06-17T17:59:31Z", "updated_at": "2025-10-29T10:34:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14866v2"}, "authors": ["Thomas Kuntz", "Agatha Duzan", "Hao Zhao", "Francesco Croce", "Zico Kolter", "Nicolas Flammarion", "Maksym Andriushchenko"], "canonical_source_url": "https://arxiv.org/abs/2506.14866", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-14866/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f551b861012fc4eddbf010a619bc65e77b524b05958761b738a88bcae377e9a9", "submitted_at": "2025-06-17T17:59:31Z", "updated_at": "2025-10-29T10:34:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.14866v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.14866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.14866"}], "name": "OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents", "publication": null, "published_at": "2025-06-17", "record_sha256": "eb0c9c51f53899966b7e4dad2822122969ddee7ed63aedc01a8bba5678979d1a", "relations": [], "resource_id": "arxiv:2506.14866", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.14866v2", "topics": ["safety-security", "evaluation"], "updated_at": "2025-10-29"}, "text": "OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nExtends OSWorld with tasks covering user misuse, prompt injection, and model misbehavior, with assessment of both task success and safety.\nPlatforms: desktop\nTopics: safety-security, evaluation\nPublished: 2025-06-17\nAuthors: Thomas Kuntz, Agatha Duzan, Hao Zhao, Francesco Croce, Zico Kolter, Nicolas Flammarion, Maksym Andriushchenko\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.14866v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-14866/"}
{"id": "arxiv:2506.15425", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3884b1807cbee971364947aa2f3756937c4c774d9942bb4da3d0fabb69bda07", "submitted_at": "2025-06-18T12:55:35Z", "updated_at": "2025-06-18T12:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.15425v1"}, "authors": ["Xingjian Tao", "Yiwei Wang", "Yujun Cai", "Zhicheng Yang", "Jing Tang"], "canonical_source_url": "https://arxiv.org/abs/2506.15425", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-15425/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3884b1807cbee971364947aa2f3756937c4c774d9942bb4da3d0fabb69bda07", "submitted_at": "2025-06-18T12:55:35Z", "updated_at": "2025-06-18T12:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.15425v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.15425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.15425"}], "name": "Understanding GUI Agent Localization Biases through Logit Sharpness", "publication": null, "published_at": "2025-06-18", "record_sha256": "ff7d0961f0a511e0122c49fd83ddcf055da62956d6213b1df03803577115a1e9", "relations": [], "resource_id": "arxiv:2506.15425", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.15425v1", "topics": ["grounding", "evaluation", "robustness", "training-free"], "updated_at": "2025-06-18"}, "text": "Understanding GUI Agent Localization Biases through Logit Sharpness\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a finer-grained taxonomy of GUI localization errors and a logit-based Peak Sharpness Score for analyzing coordinate-prediction uncertainty. A context-aware cropping method refines the visual input without retraining the grounding model.\nPlatforms: unspecified\nTopics: grounding, evaluation, robustness, training-free\nPublished: 2025-06-18\nAuthors: Xingjian Tao, Yiwei Wang, Yujun Cai, Zhicheng Yang, Jing Tang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.15425v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-15425/"}
{"id": "arxiv:2506.15677", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e8f714b234a1aad60d0306df23c90410aebfd257fb5906644d0382479e8f3d0", "submitted_at": "2025-06-18T17:58:17Z", "updated_at": "2025-07-29T22:40:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.15677v3"}, "authors": ["Yining Hong", "Rui Sun", "Bingxuan Li", "Xingcheng Yao", "Maxine Wu", "Alexander Chien", "Da Yin", "Ying Nian Wu", "Zhecan James Wang", "Kai-Wei Chang"], "canonical_source_url": "https://arxiv.org/abs/2506.15677", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-15677/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e8f714b234a1aad60d0306df23c90410aebfd257fb5906644d0382479e8f3d0", "submitted_at": "2025-06-18T17:58:17Z", "updated_at": "2025-07-29T22:40:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.15677v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.15677"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.15677"}], "name": "Embodied Web Agents: Bridging Physical-Digital Realms for Integrated Agent Intelligence", "publication": null, "published_at": "2025-06-18", "record_sha256": "d8ca294202ba2fe5c671bc2b5cb222f5339f6199744585562d7554839a0bbbd5", "relations": [], "resource_id": "arxiv:2506.15677", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.15677v3", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2025-07-29"}, "text": "Embodied Web Agents: Bridging Physical-Digital Realms for Integrated Agent Intelligence\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces environments that couple 3D physical-world simulations with functional web interfaces. The accompanying benchmark evaluates tasks requiring agents to coordinate online information and physical actions across cooking, navigation and shopping scenarios.\nPlatforms: web\nTopics: evaluation, web-navigation, planning\nPublished: 2025-06-18\nAuthors: Yining Hong, Rui Sun, Bingxuan Li, Xingcheng Yao, Maxine Wu, Alexander Chien, Da Yin, Ying Nian Wu, Zhecan James Wang, Kai-Wei Chang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The physical component is a simulation; the supported catalog connection is its functional web interaction.\n\nPrimary source: https://arxiv.org/abs/2506.15677v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-15677/"}
{"id": "arxiv:2506.16042", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df3f654ef1364aeb6cdaeaea1c6334da949a798b2f8256c9ddc4e6a88a588e19", "submitted_at": "2025-06-19T05:26:40Z", "updated_at": "2026-05-18T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.16042v2"}, "authors": ["Reyna Abhyankar", "Qi Qi", "Yiying Zhang"], "canonical_source_url": "https://arxiv.org/abs/2506.16042", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-16042/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df3f654ef1364aeb6cdaeaea1c6334da949a798b2f8256c9ddc4e6a88a588e19", "submitted_at": "2025-06-19T05:26:40Z", "updated_at": "2026-05-18T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.16042v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.16042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.16042"}], "name": "OSWorld-Human: Benchmarking the Efficiency of Computer-Use Agents", "publication": null, "published_at": "2025-06-19", "record_sha256": "69f65313433ab315c30d59b3ef7ef9806bf976e2a543a8900d2ae0b1ae449cbc", "relations": [], "resource_id": "arxiv:2506.16042", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.16042v2", "topics": ["evaluation", "efficiency", "observability"], "updated_at": "2026-05-18"}, "text": "OSWorld-Human: Benchmarking the Efficiency of Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOSWorld-Human adds human-determined trajectories to OSWorld to evaluate the efficiency of desktop computer-use agents. The study compares 16 agents and analyzes model-call latency, changes in step duration during task execution, and excess action steps.\nPlatforms: desktop\nTopics: evaluation, efficiency, observability\nPublished: 2025-06-19\nAuthors: Reyna Abhyankar, Qi Qi, Yiying Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.16042v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-16042/"}
{"id": "arxiv:2506.17318", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d346389013d465f578417a7fca80ee527f5e7b582e93eb73dae148f11479971", "submitted_at": "2025-06-18T14:29:02Z", "updated_at": "2025-06-18T14:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.17318v1"}, "authors": ["Atharv Singh Patlan", "Ashwin Hebbar", "Pramod Viswanath", "Prateek Mittal"], "canonical_source_url": "https://arxiv.org/abs/2506.17318", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-17318/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d346389013d465f578417a7fca80ee527f5e7b582e93eb73dae148f11479971", "submitted_at": "2025-06-18T14:29:02Z", "updated_at": "2025-06-18T14:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.17318v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.17318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.17318"}], "name": "Context manipulation attacks : Web agents are susceptible to corrupted memory", "publication": null, "published_at": "2025-06-18", "record_sha256": "939d9723f402d56b9fb13153f53aa03e45e54c1697f02712d0b3e52a354576c8", "relations": [], "resource_id": "arxiv:2506.17318", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.17318v1", "topics": ["safety-security", "memory", "evaluation"], "updated_at": "2025-06-18"}, "text": "Context manipulation attacks : Web agents are susceptible to corrupted memory\nResource type: paper\nReview status: curated\nScope: core\n\nStudies context manipulation in web agents by altering stored task representations rather than webpage prompts alone. Evaluation on Browser-use and Agent-E examines memory-directed attacks and the implications for secure context handling.\nPlatforms: web\nTopics: safety-security, memory, evaluation\nPublished: 2025-06-18\nAuthors: Atharv Singh Patlan, Ashwin Hebbar, Pramod Viswanath, Prateek Mittal\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.17318v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-17318/"}
{"id": "arxiv:2506.17913", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f12d2ae58243b5f89a6a510e8f3ef4073a605c4bafdf63c37fd38db57ea84602", "submitted_at": "2025-06-22T06:30:52Z", "updated_at": "2026-07-21T05:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.17913v2"}, "authors": ["Jinjie Wei", "Jiyao Liu", "Lihao Liu", "Ming Hu", "Junzhi Ning", "Mingcheng Li", "Weijie Yin", "Junjun He", "Xiao Liang", "Chao Feng", "Dingkang Yang"], "canonical_source_url": "https://arxiv.org/abs/2506.17913", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-17913/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f12d2ae58243b5f89a6a510e8f3ef4073a605c4bafdf63c37fd38db57ea84602", "submitted_at": "2025-06-22T06:30:52Z", "updated_at": "2026-07-21T05:30:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2506.17913v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.17913"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.17913"}], "name": "Learning, Reasoning, Refinement: A Framework for Kahneman's Dual-System Intelligence in GUI Agents", "publication": null, "published_at": "2025-06-22", "record_sha256": "3336a42b7223f7eb5cd31d483c6155fe4dfae5d3aec3880bc2c4835b605942f0", "relations": [], "resource_id": "arxiv:2506.17913", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.17913v2", "topics": ["grounding", "planning", "reinforcement-learning", "evaluation"], "updated_at": "2026-07-21"}, "text": "Learning, Reasoning, Refinement: A Framework for Kahneman's Dual-System Intelligence in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCogniGUI combines hierarchical visual parsing of interface elements with a GRPO-based grounding agent that compares alternative interaction paths. Its ScreenSeek benchmark evaluates navigation across applications, changing interface states, and consistency between interfaces.\nPlatforms: unspecified\nTopics: grounding, planning, reinforcement-learning, evaluation\nPublished: 2025-06-22\nAuthors: Jinjie Wei, Jiyao Liu, Lihao Liu, Ming Hu, Junzhi Ning, Mingcheng Li, Weijie Yin, Junjun He, Xiao Liang, Chao Feng, Dingkang Yang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.17913v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-17913/"}
{"id": "arxiv:2506.18158", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaa62e24a7a269fea49471f17028a3d5d8068d934e450f53bc002ae80fb7cb03", "submitted_at": "2025-06-22T20:17:46Z", "updated_at": "2025-06-22T20:17:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.18158v1"}, "authors": ["Xinzge Gao", "Chuanrui Hu", "Bin Chen", "Teng Li"], "canonical_source_url": "https://arxiv.org/abs/2506.18158", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-18158/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaa62e24a7a269fea49471f17028a3d5d8068d934e450f53bc002ae80fb7cb03", "submitted_at": "2025-06-22T20:17:46Z", "updated_at": "2025-06-22T20:17:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2506.18158v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.18158"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.18158"}], "name": "Chain-of-Memory: Enhancing GUI Agents for Cross-Application Navigation", "publication": null, "published_at": "2025-06-22", "record_sha256": "b5a29f5580909366e15b52b5141f05ba2c9f588d0ddf8cd308460b9c850af2e1", "relations": [], "resource_id": "arxiv:2506.18158", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.18158v1", "topics": ["memory", "state-tracking", "cross-app", "data-generation"], "updated_at": "2025-06-22"}, "text": "Chain-of-Memory: Enhancing GUI Agents for Cross-Application Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nChain-of-Memory explicitly records action descriptions and task-relevant screen information in short- and long-term memory for GUI navigation. GUI Odyssey-CoM provides 111,000 screen-action pairs with memory annotations for learning and evaluating information retention across applications.\nPlatforms: unspecified\nTopics: memory, state-tracking, cross-app, data-generation\nPublished: 2025-06-22\nAuthors: Xinzge Gao, Chuanrui Hu, Bin Chen, Teng Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.18158v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-18158/"}
{"id": "arxiv:2506.20332", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9620eb7d8fa5f5c2a318ff516d45b8a4cfd35cf0c235b96bab3bba4e5a6aab9b", "submitted_at": "2025-06-25T11:34:43Z", "updated_at": "2026-04-25T13:54:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.20332v4"}, "authors": ["Jihao Gu", "Qihang Ai", "Yingyao Wang", "Pi Bu", "Jingxuan Xing", "Zekun Zhu", "Wei Jiang", "Ziming Wang", "Yingxiu Zhao", "Ming-Liang Zhang", "Jun Song", "Yuning Jiang", "Bo Zheng"], "canonical_source_url": "https://arxiv.org/abs/2506.20332", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-20332/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9620eb7d8fa5f5c2a318ff516d45b8a4cfd35cf0c235b96bab3bba4e5a6aab9b", "submitted_at": "2025-06-25T11:34:43Z", "updated_at": "2026-04-25T13:54:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2506.20332v4"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2506.20332"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.20332"}], "name": "Mobile-R1: Towards Interactive Capability for VLM-Based Mobile Agent via Systematic Training", "publication": null, "published_at": "2025-06-25", "record_sha256": "73a4daef331d0999eb0be3b8ffd9e4cc46bbcb13473907fad6aeb6708ac94169", "relations": [], "resource_id": "arxiv:2506.20332", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.20332v4", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": "2026-04-25"}, "text": "Mobile-R1: Towards Interactive Capability for VLM-Based Mobile Agent via Systematic Training\nResource type: paper\nReview status: curated\nScope: core\n\nUses a staged curriculum from response-format alignment through action feedback to interactive task-level training for mobile agents.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation\nPublished: 2025-06-25\nAuthors: Jihao Gu, Qihang Ai, Yingyao Wang, Pi Bu, Jingxuan Xing, Zekun Zhu, Wei Jiang, Ziming Wang, Yingxiu Zhao, Ming-Liang Zhang, Jun Song, Yuning Jiang, Bo Zheng\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.20332v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-20332/"}
{"id": "arxiv:2506.23491", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb3678b0d0e645a87ce9392b606c1d67dcbc40656880c57d1cbb8ddd8ede7ea1", "submitted_at": "2025-06-30T03:33:02Z", "updated_at": "2025-07-18T06:03:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.23491v3"}, "authors": ["ZongHan Hsieh", "Tzer-Jen Wei", "ShengJing Yang"], "canonical_source_url": "https://arxiv.org/abs/2506.23491", "citation_url": "https://awesomecua.github.io/resources/arxiv-2506-23491/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb3678b0d0e645a87ce9392b606c1d67dcbc40656880c57d1cbb8ddd8ede7ea1", "submitted_at": "2025-06-30T03:33:02Z", "updated_at": "2025-07-18T06:03:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2506.23491v3"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2506.23491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2506.23491"}], "name": "ZonUI-3B: A Lightweight Vision-Language Model for Cross-Resolution GUI Grounding", "publication": null, "published_at": "2025-06-30", "record_sha256": "97906505000f76269896ca84036f70ec58fba6270f5f1078a7781c1c410c71e3", "relations": [], "resource_id": "arxiv:2506.23491", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2506.23491v3", "topics": ["grounding", "data-generation", "efficiency"], "updated_at": "2025-07-18"}, "text": "ZonUI-3B: A Lightweight Vision-Language Model for Cross-Resolution GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nZonUI-3B trains GUI grounding with a curated 24,000-example collection of mobile, desktop, and web screenshots. Its two-stage procedure first learns across platforms and then specializes on high-resolution images, with ablations on data diversity and redundant examples.\nPlatforms: web, desktop, cross-platform\nTopics: grounding, data-generation, efficiency\nPublished: 2025-06-30\nAuthors: ZongHan Hsieh, Tzer-Jen Wei, ShengJing Yang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2506.23491v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2506-23491/"}
{"id": "arxiv:2507.00008", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e423834b07e74155214597c0fbb66c26b949e2b726568a65617a670a9ab157fa", "submitted_at": "2025-06-12T03:13:21Z", "updated_at": "2025-09-05T17:21:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.00008v2"}, "authors": ["Hang Wu", "Hongkai Chen", "Yujun Cai", "Chang Liu", "Qingwen Ye", "Ming-Hsuan Yang", "Yiwei Wang"], "canonical_source_url": "https://arxiv.org/abs/2507.00008", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-00008/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e423834b07e74155214597c0fbb66c26b949e2b726568a65617a670a9ab157fa", "submitted_at": "2025-06-12T03:13:21Z", "updated_at": "2025-09-05T17:21:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.00008v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.00008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00008"}], "name": "DiMo-GUI: Advancing Test-time Scaling in GUI Grounding via Modality-Aware Visual Reasoning", "publication": null, "published_at": "2025-06-12", "record_sha256": "8a188ed4cf1663da279124d1e9d687e3ccb97d3124bd7b39d13afabea06d7ef6", "relations": [], "resource_id": "arxiv:2507.00008", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.00008v2", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-09-05"}, "text": "DiMo-GUI: Advancing Test-time Scaling in GUI Grounding via Modality-Aware Visual Reasoning\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates textual and iconic UI elements and iteratively focuses visual regions to refine grounding without additional model training.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\nPublished: 2025-06-12\nAuthors: Hang Wu, Hongkai Chen, Yujun Cai, Chang Liu, Qingwen Ye, Ming-Hsuan Yang, Yiwei Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.00008v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-00008/"}
{"id": "arxiv:2507.00210", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8972688fb95908c024a85e3ffc258be61280ee8d22e6ab97dfb0c4997b83445", "submitted_at": "2025-06-30T19:24:45Z", "updated_at": "2025-06-30T19:24:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00210v1"}, "authors": ["Imene Kerboua", "Sahar Omidi Shayegan", "Megh Thakkar", "Xing Han Lù", "Massimo Caccia", "Véronique Eglin", "Alexandre Aussem", "Jérémy Espinas", "Alexandre Lacoste"], "canonical_source_url": "https://arxiv.org/abs/2507.00210", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-00210/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b8972688fb95908c024a85e3ffc258be61280ee8d22e6ab97dfb0c4997b83445", "submitted_at": "2025-06-30T19:24:45Z", "updated_at": "2025-06-30T19:24:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00210v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.00210"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00210"}], "name": "LineRetriever: Planning-Aware Observation Reduction for Web Agents", "publication": null, "published_at": "2025-06-30", "record_sha256": "a9c6a46a55d1c7a0cda7b810530dd9239c015ac6e1b30995cb675e0e182444e5", "relations": [], "resource_id": "arxiv:2507.00210", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.00210v1", "topics": ["web-navigation", "efficiency", "planning"], "updated_at": "2025-06-30"}, "text": "LineRetriever: Planning-Aware Observation Reduction for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nLineRetriever uses a language model to select observation lines that matter for upcoming web-navigation actions. It reduces DOM or accessibility-tree context by considering the planning horizon instead of relying solely on semantic similarity.\nPlatforms: web\nTopics: web-navigation, efficiency, planning\nPublished: 2025-06-30\nAuthors: Imene Kerboua, Sahar Omidi Shayegan, Megh Thakkar, Xing Han Lù, Massimo Caccia, Véronique Eglin, Alexandre Aussem, Jérémy Espinas, Alexandre Lacoste\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.00210v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-00210/"}
{"id": "arxiv:2507.00841", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78dd523ad6123fe1ee5164685a3853719c9a4442adac8661050f147ff6a9f619", "submitted_at": "2025-07-01T15:10:00Z", "updated_at": "2025-07-01T15:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00841v1"}, "authors": ["Siyuan Liang", "Tianmeng Fang", "Zhe Liu", "Aishan Liu", "Yan Xiao", "Jinyuan He", "Ee-Chien Chang", "Xiaochun Cao"], "canonical_source_url": "https://arxiv.org/abs/2507.00841", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-00841/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78dd523ad6123fe1ee5164685a3853719c9a4442adac8661050f147ff6a9f619", "submitted_at": "2025-07-01T15:10:00Z", "updated_at": "2025-07-01T15:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.00841v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:56.912207+00:00", "role": "primary", "sha256": "a25da6700a238167d5e9915cf6fb3a4bcd910cefc300f4be6eed56271be26f00", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2507.00841v1"}], "identifiers": {"arxiv": "2507.00841"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:16.030926+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00841"}], "name": "SafeMobile: Chain-level Jailbreak Detection and Automated Evaluation for Multimodal Mobile Agents", "publication": null, "published_at": "2025-07-01", "record_sha256": "484528d7f70223a1481fb684ef334e42c83755de460752e95c56385aa5dfb903", "relations": [], "resource_id": "arxiv:2507.00841", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.00841v1", "topics": ["safety-security", "evaluation", "mobile-navigation"], "updated_at": "2025-07-01"}, "text": "SafeMobile: Chain-level Jailbreak Detection and Automated Evaluation for Multimodal Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies chain-level risk detection for mobile multimodal agents and a language-model-assisted evaluator, assessing safety interventions together with functional preservation.\nPlatforms: android\nTopics: safety-security, evaluation, mobile-navigation\nPublished: 2025-07-01\nAuthors: Siyuan Liang, Tianmeng Fang, Zhe Liu, Aishan Liu, Yan Xiao, Jinyuan He, Ee-Chien Chang, Xiaochun Cao\nRecorded arXiv version: v1\n\nThe reviewed evaluation uses Android emulation and a specific MobileSafetyBench-derived setup. Its judge and score threshold are evaluator choices, not universal safety measurements, and its task population is not the whole later MobileSafetyBench corpus. No results were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (II.B Defense objective; III Chain-level guard and GPTJudge; IV.B Evaluation setup; V.D Judge validity); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2507.00841v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-00841/"}
{"id": "arxiv:2507.00938", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1745ec9cf10c465c0f389f885df6db502f1884cca271bccea1b45acd2c787af9", "submitted_at": "2025-07-01T16:43:57Z", "updated_at": "2026-09-24T12:58:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.00938v4"}, "authors": ["Zihao Sun", "Zijing Shi", "Ling Chen"], "canonical_source_url": "https://arxiv.org/abs/2507.00938", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-00938/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1745ec9cf10c465c0f389f885df6db502f1884cca271bccea1b45acd2c787af9", "submitted_at": "2025-07-01T16:43:57Z", "updated_at": "2026-09-24T12:58:22Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.00938v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.00938"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.00938"}], "name": "WebArxiv: A Reproducible Benchmark for Evaluating Multimodal Web Agents on arXiv Tasks", "publication": null, "published_at": "2025-07-01", "record_sha256": "c7d420cd521766395283cf24366fbdb289196fc30d212b1a31e18c236fa5631d", "relations": [], "resource_id": "arxiv:2507.00938", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.00938v4", "topics": ["evaluation", "web-navigation", "memory"], "updated_at": "2026-09-24"}, "text": "WebArxiv: A Reproducible Benchmark for Evaluating Multimodal Web Agents on arXiv Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nWebArxiv provides 510 static-snapshot tasks with deterministic answers for scholarly browsing, including constrained paper retrieval, content extraction, and cross-paper comparison. The study also tests dynamic retrieval of interaction history to address repetitive or incomplete agent reasoning.\nPlatforms: web\nTopics: evaluation, web-navigation, memory\nPublished: 2025-07-01\nAuthors: Zihao Sun, Zijing Shi, Ling Chen\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.00938v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-00938/"}
{"id": "arxiv:2507.03730", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "774e2515b563f75d4561daaa6a4b672423d917a3f50e98308ab448e2ae0731e5", "submitted_at": "2025-07-04T17:37:15Z", "updated_at": "2025-07-04T17:37:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.03730v1"}, "authors": ["Gongwei Chen", "Xurui Zhou", "Rui Shao", "Yibo Lyu", "Kaiwen Zhou", "Shuai Wang", "Wentao Li", "Yinchuan Li", "Zhongang Qi", "Liqiang Nie"], "canonical_source_url": "https://arxiv.org/abs/2507.03730", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-03730/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "774e2515b563f75d4561daaa6a4b672423d917a3f50e98308ab448e2ae0731e5", "submitted_at": "2025-07-04T17:37:15Z", "updated_at": "2025-07-04T17:37:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.03730v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.03730"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.03730"}], "name": "Less is More: Empowering GUI Agent with Context-Aware Simplification", "publication": null, "published_at": "2025-07-04", "record_sha256": "d6daec76c3db65bdbf33652acbe4465117af035a223e5c6fcf4db110a311a9bb", "relations": [], "resource_id": "arxiv:2507.03730", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.03730v1", "topics": ["efficiency", "interaction-history", "web-navigation", "mobile-navigation"], "updated_at": "2025-07-04"}, "text": "Less is More: Empowering GUI Agent with Context-Aware Simplification\nResource type: paper\nReview status: curated\nScope: core\n\nSimpAgent prunes distracting interface elements with visual masks and compresses interaction history using consistency-based guidance. The study evaluates these two forms of context reduction during GUI navigation in web and mobile environments.\nPlatforms: web, cross-platform\nTopics: efficiency, interaction-history, web-navigation, mobile-navigation\nPublished: 2025-07-04\nAuthors: Gongwei Chen, Xurui Zhou, Rui Shao, Yibo Lyu, Kaiwen Zhou, Shuai Wang, Wentao Li, Yinchuan Li, Zhongang Qi, Liqiang Nie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.03730v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-03730/"}
{"id": "arxiv:2507.04103", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15fa527a869ab66f4f39900ab07ea8b8db930703a8e950ed1e195e1e66164099", "submitted_at": "2025-07-05T17:12:33Z", "updated_at": "2026-02-13T17:24:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.04103v4"}, "authors": ["Dheeraj Vattikonda", "Santhoshi Ravichandran", "Emiliano Penaloza", "Hadi Nekoei", "Megh Thakkar", "Thibault Le Sellier de Chezelles", "Nicolas Gontier", "Miguel Muñoz-Mármol", "Sahar Omidi Shayegan", "Stefania Raimondo", "Xue Liu", "Alexandre Drouin", "Laurent Charlin", "Alexandre Piché", "Alexandre Lacoste", "Massimo Caccia"], "canonical_source_url": "https://arxiv.org/abs/2507.04103", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-04103/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15fa527a869ab66f4f39900ab07ea8b8db930703a8e950ed1e195e1e66164099", "submitted_at": "2025-07-05T17:12:33Z", "updated_at": "2026-02-13T17:24:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.04103v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.04103"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.04103"}], "name": "How to Train Your LLM Web Agent: A Statistical Diagnosis", "publication": null, "published_at": "2025-07-05", "record_sha256": "7c596e89c72d2483192a44803cc560ba77c794021902a146429d853ececb386d", "relations": [], "resource_id": "arxiv:2507.04103", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.04103v4", "topics": ["imitation-learning", "reinforcement-learning", "efficiency", "evaluation"], "updated_at": "2026-02-13"}, "text": "How to Train Your LLM Web Agent: A Statistical Diagnosis\nResource type: paper\nReview status: curated\nScope: core\n\nStudies compute allocation for web-agent post-training with teacher-generated supervised fine-tuning followed by on-policy reinforcement learning. It samples 1,370 training configurations and uses bootstrap analysis to examine hyperparameter sensitivity on WorkArena and MiniWoB++.\nPlatforms: web\nTopics: imitation-learning, reinforcement-learning, efficiency, evaluation\nPublished: 2025-07-05\nAuthors: Dheeraj Vattikonda, Santhoshi Ravichandran, Emiliano Penaloza, Hadi Nekoei, Megh Thakkar, Thibault Le Sellier de Chezelles, Nicolas Gontier, Miguel Muñoz-Mármol, Sahar Omidi Shayegan, Stefania Raimondo, Xue Liu, Alexandre Drouin, Laurent Charlin, Alexandre Piché, Alexandre Lacoste, Massimo Caccia\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.04103v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-04103/"}
{"id": "arxiv:2507.04227", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85ebf8b3b5916e66e4b4221c73bb29519156c41e8d1695bcc46d275dd50cf58d", "submitted_at": "2025-07-06T03:31:36Z", "updated_at": "2026-04-14T14:47:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.04227v2"}, "authors": ["Guohong Liu", "Jialei Ye", "Jiacheng Liu", "Yuanchun Li", "Wei Liu", "Pengzhi Gao", "Jian Luan", "Yunxin Liu"], "canonical_source_url": "https://arxiv.org/abs/2507.04227", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-04227/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "85ebf8b3b5916e66e4b4221c73bb29519156c41e8d1695bcc46d275dd50cf58d", "submitted_at": "2025-07-06T03:31:36Z", "updated_at": "2026-04-14T14:47:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.04227v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.04227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.04227"}], "name": "Mobile GUI Agents under Real-world Threats: Are We There Yet?", "publication": null, "published_at": "2025-07-06", "record_sha256": "89ef78a12c92a56ec377b544b9e600afce5a41d3cfdf911ab3adcb15ef22f5ba", "relations": [], "resource_id": "arxiv:2507.04227", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.04227v2", "topics": ["mobile-navigation", "safety-security", "evaluation"], "updated_at": "2026-04-14"}, "text": "Mobile GUI Agents under Real-world Threats: Are We There Yet?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces an app-content instrumentation framework and complementary interactive and static test suites for mobile GUI agents. It studies how untrusted third-party content changes task execution and state understanding.\nPlatforms: unspecified\nTopics: mobile-navigation, safety-security, evaluation\nPublished: 2025-07-06\nAuthors: Guohong Liu, Jialei Ye, Jiacheng Liu, Yuanchun Li, Wei Liu, Pengzhi Gao, Jian Luan, Yunxin Liu\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.04227v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-04227/"}
{"id": "arxiv:2507.05445", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ad2f299303076e3474cf247cf25d20ba63b1f1aa8601ab74243c00896bb56e4", "submitted_at": "2025-07-07T19:50:21Z", "updated_at": "2025-07-07T19:50:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05445v1"}, "authors": ["Daniel Jones", "Giorgio Severi", "Martin Pouliot", "Gary Lopez", "Joris de Gruyter", "Santiago Zanella-Beguelin", "Justin Song", "Blake Bullwinkel", "Pamela Cortez", "Amanda Minnich"], "canonical_source_url": "https://arxiv.org/abs/2507.05445", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-05445/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3ad2f299303076e3474cf247cf25d20ba63b1f1aa8601ab74243c00896bb56e4", "submitted_at": "2025-07-07T19:50:21Z", "updated_at": "2025-07-07T19:50:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05445v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.05445"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05445"}], "name": "A Systematization of Security Vulnerabilities in Computer Use Agents", "publication": null, "published_at": "2025-07-07", "record_sha256": "6c5c01488edb4dd8e26e2fec6ac534cc36d888f81522c03e42bcbe3f2cef9dd2", "relations": [], "resource_id": "arxiv:2507.05445", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.05445v1", "topics": ["safety-security", "evaluation"], "updated_at": "2025-07-07"}, "text": "A Systematization of Security Vulnerabilities in Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes security boundaries in computer-use agents through a risk taxonomy and concrete adversarial case studies. It links observed failures to input provenance, interface-action binding, memory, and delegation, and proposes CUA-specific evaluation principles.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2025-07-07\nAuthors: Daniel Jones, Giorgio Severi, Martin Pouliot, Gary Lopez, Joris de Gruyter, Santiago Zanella-Beguelin, Justin Song, Blake Bullwinkel, Pamela Cortez, Amanda Minnich\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. This is a threat analysis with case studies; the annotation does not imply a standardized public benchmark or independently validated defenses.\n\nPrimary source: https://arxiv.org/abs/2507.05445v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-05445/"}
{"id": "arxiv:2507.05673", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aff9b208c446bd57f24258734557e0cd5caaf9146af68e594ec6a7e575919a6f", "submitted_at": "2025-07-08T04:56:57Z", "updated_at": "2025-07-08T04:56:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05673v1"}, "authors": ["Joonhyung Park", "Peng Tang", "Sagnik Das", "Srikar Appalaraju", "Kunwar Yashraj Singh", "R. Manmatha", "Shabnam Ghadar"], "canonical_source_url": "https://arxiv.org/abs/2507.05673", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-05673/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aff9b208c446bd57f24258734557e0cd5caaf9146af68e594ec6a7e575919a6f", "submitted_at": "2025-07-08T04:56:57Z", "updated_at": "2025-07-08T04:56:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05673v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.05673"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05673"}], "name": "R-VLM: Region-Aware Vision Language Model for Precise GUI Grounding", "publication": null, "published_at": "2025-07-08", "record_sha256": "1e395e03fb3519b8787106a3aad47695c46dcf1b80d90b6abcdfa46721607277", "relations": [], "resource_id": "arxiv:2507.05673", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.05673v1", "topics": ["grounding", "screen-understanding", "web-navigation"], "updated_at": "2025-07-08"}, "text": "R-VLM: Region-Aware Vision Language Model for Precise GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nR-VLM localizes GUI elements through zoomed-in region proposals and an intersection-over-union-aware training objective. It evaluates the grounding model on ScreenSpot and AgentStudio and studies its effect on navigation in AITW and Mind2Web.\nPlatforms: web, cross-platform\nTopics: grounding, screen-understanding, web-navigation\nPublished: 2025-07-08\nAuthors: Joonhyung Park, Peng Tang, Sagnik Das, Srikar Appalaraju, Kunwar Yashraj Singh, R. Manmatha, Shabnam Ghadar\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.05673v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-05673/"}
{"id": "arxiv:2507.05720", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45d36dd32b30bf4faee6d9a72e3cdfe91edff465bbd886bdfc3c27c3da475b82", "submitted_at": "2025-07-08T07:07:53Z", "updated_at": "2025-07-08T07:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05720v1"}, "authors": ["Yucheng Shi", "Wenhao Yu", "Zaitang Li", "Yonglin Wang", "Hongming Zhang", "Ninghao Liu", "Haitao Mi", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2507.05720", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-05720/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "45d36dd32b30bf4faee6d9a72e3cdfe91edff465bbd886bdfc3c27c3da475b82", "submitted_at": "2025-07-08T07:07:53Z", "updated_at": "2025-07-08T07:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.05720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.05720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05720"}], "name": "MobileGUI-RL: Advancing Mobile GUI Agent through Reinforcement Learning in Online Environment", "publication": null, "published_at": "2025-07-08", "record_sha256": "eef8efaefbf47236cf0781ee3b77f94b0b139cf054b1eccea9d0ac3b92aed1df", "relations": [], "resource_id": "arxiv:2507.05720", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.05720v1", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation", "efficiency"], "updated_at": "2025-07-08"}, "text": "MobileGUI-RL: Advancing Mobile GUI Agent through Reinforcement Learning in Online Environment\nResource type: paper\nReview status: curated\nScope: core\n\nMobileGUI-RL constructs an online training curriculum through agent exploration and task filtering. It adapts GRPO with trajectory-aware advantages and rewards for task completion and execution efficiency, evaluated on three online mobile-agent benchmarks.\nPlatforms: unspecified\nTopics: mobile-navigation, reinforcement-learning, data-generation, efficiency\nPublished: 2025-07-08\nAuthors: Yucheng Shi, Wenhao Yu, Zaitang Li, Yonglin Wang, Hongming Zhang, Ninghao Liu, Haitao Mi, Dong Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.05720v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-05720/"}
{"id": "arxiv:2507.05791", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c1f659a30e1a0bd04219b9383927ce72e5f5b0d20783a89b850036d087155a4", "submitted_at": "2025-07-08T08:52:18Z", "updated_at": "2025-10-03T23:50:19Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2507.05791v5"}, "authors": ["Yan Yang", "Dongxu Li", "Yutong Dai", "Yuhao Yang", "Ziyang Luo", "Zirui Zhao", "Zhiyuan Hu", "Junzhe Huang", "Amrita Saha", "Zeyuan Chen", "Ran Xu", "Liyuan Pan", "Silvio Savarese", "Caiming Xiong", "Junnan Li"], "canonical_source_url": "https://arxiv.org/abs/2507.05791", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-05791/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c1f659a30e1a0bd04219b9383927ce72e5f5b0d20783a89b850036d087155a4", "submitted_at": "2025-07-08T08:52:18Z", "updated_at": "2025-10-03T23:50:19Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2507.05791v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2507.05791"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.05791"}], "name": "GTA1: GUI Test-time Scaling Agent", "publication": null, "published_at": "2025-07-08", "record_sha256": "96ddfa9591fd329724e448a6c516eb9a8fd893daa538cc03fd38825bdffe56eb", "relations": [], "resource_id": "arxiv:2507.05791", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.05791v5", "topics": ["planning", "grounding", "reinforcement-learning"], "updated_at": "2025-10-03"}, "text": "GTA1: GUI Test-time Scaling Agent\nResource type: paper\nReview status: curated\nScope: core\n\nGTA1 samples multiple action proposals at each GUI step and uses a judge to select among them. A separate grounding model is trained with reinforcement learning to map selected proposals to visual targets, linking planning-time search with click localization.\nPlatforms: unspecified\nTopics: planning, grounding, reinforcement-learning\nPublished: 2025-07-08\nAuthors: Yan Yang, Dongxu Li, Yutong Dai, Yuhao Yang, Ziyang Luo, Zirui Zhao, Zhiyuan Hu, Junzhe Huang, Amrita Saha, Zeyuan Chen, Ran Xu, Liyuan Pan, Silvio Savarese, Caiming Xiong, Junnan Li\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.05791v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-05791/"}
{"id": "arxiv:2507.06134", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db56d120a7f04f9181a107bcd75c61a3eb3ac36939e2a0517d622332c70f3eac", "submitted_at": "2025-07-08T16:18:54Z", "updated_at": "2026-02-16T22:34:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06134v2"}, "authors": ["Sanidhya Vijayvargiya", "Aditya Bharat Soni", "Xuhui Zhou", "Zora Zhiruo Wang", "Nouha Dziri", "Graham Neubig", "Maarten Sap"], "canonical_source_url": "https://arxiv.org/abs/2507.06134", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-06134/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db56d120a7f04f9181a107bcd75c61a3eb3ac36939e2a0517d622332c70f3eac", "submitted_at": "2025-07-08T16:18:54Z", "updated_at": "2026-02-16T22:34:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06134v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2507.06134"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.06134"}], "name": "OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety", "publication": null, "published_at": "2025-07-08", "record_sha256": "ada9212477b4bfb6656aee266ec16e650a1edd4ddee73d01a0001ef082b13c77", "relations": [], "resource_id": "arxiv:2507.06134", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.06134v2", "topics": ["safety-security", "evaluation"], "updated_at": "2026-02-16"}, "text": "OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates agent safety across real browser, code, filesystem, shell, and messaging tools using rule-based and model-based assessments.\nPlatforms: web, desktop\nTopics: safety-security, evaluation\nPublished: 2025-07-08\nAuthors: Sanidhya Vijayvargiya, Aditya Bharat Soni, Xuhui Zhou, Zora Zhiruo Wang, Nouha Dziri, Graham Neubig, Maarten Sap\nRecorded arXiv version: v2\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.06134v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-06134/"}
{"id": "arxiv:2507.06899", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010a3a54ba37417c97e52434211375447418dd2e7ce1da6c339532534f0a920e", "submitted_at": "2025-07-09T14:36:00Z", "updated_at": "2025-09-24T14:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06899v2"}, "authors": ["Ziang Ye", "Yang Zhang", "Wentao Shi", "Xiaoyu You", "Fuli Feng", "Tat-Seng Chua"], "canonical_source_url": "https://arxiv.org/abs/2507.06899", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-06899/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010a3a54ba37417c97e52434211375447418dd2e7ce1da6c339532534f0a920e", "submitted_at": "2025-07-09T14:36:00Z", "updated_at": "2025-09-24T14:33:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.06899v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.06899"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.06899"}], "name": "VisualTrap: A Stealthy Backdoor Attack on GUI Agents via Visual Grounding Manipulation", "publication": null, "published_at": "2025-07-09", "record_sha256": "b94dddd7f7e9aa8db53c99e1fbb097ef11c96f47ffcc98278110f0e2439a1f66", "relations": [], "resource_id": "arxiv:2507.06899", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.06899v2", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2025-09-24"}, "text": "VisualTrap: A Stealthy Backdoor Attack on GUI Agents via Visual Grounding Manipulation\nResource type: paper\nReview status: curated\nScope: core\n\nStudies backdoors that redirect GUI visual grounding even when the task plan is correct. VisualTrap evaluates whether poisoned grounding training transfers to downstream tasks and across web, mobile, and desktop interface settings.\nPlatforms: web, desktop\nTopics: safety-security, grounding, evaluation\nPublished: 2025-07-09\nAuthors: Ziang Ye, Yang Zhang, Wentao Shi, Xiaoyu You, Fuli Feng, Tat-Seng Chua\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.06899v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-06899/"}
{"id": "arxiv:2507.10610", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5db0008a053357b6dba38b55f29642ddbcbf38761c64684fd1725183ff8e56d3", "submitted_at": "2025-07-13T08:36:09Z", "updated_at": "2026-04-07T11:46:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.10610v3"}, "authors": ["Zihe Yan", "Jiaping Gui", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2507.10610", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-10610/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5db0008a053357b6dba38b55f29642ddbcbf38761c64684fd1725183ff8e56d3", "submitted_at": "2025-07-13T08:36:09Z", "updated_at": "2026-04-07T11:46:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.10610v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.10610"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.10610"}], "name": "LaSM: Layer-wise Scaling Mechanism for Defending Pop-up Attack on GUI Agents", "publication": null, "published_at": "2025-07-13", "record_sha256": "102f20c763d06959d45698c543867210375fa3f963dd496a1e392a7e653edd72", "relations": [], "resource_id": "arxiv:2507.10610", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.10610v3", "topics": ["safety-security", "grounding", "training-free"], "updated_at": "2026-04-07"}, "text": "LaSM: Layer-wise Scaling Mechanism for Defending Pop-up Attack on GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces LaSM, a training-free defense that selectively scales attention and MLP components to improve alignment with task-relevant screen regions. It evaluates the approach against pop-up-based GUI injection attacks.\nPlatforms: unspecified\nTopics: safety-security, grounding, training-free\nPublished: 2025-07-13\nAuthors: Zihe Yan, Jiaping Gui, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.10610v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-10610/"}
{"id": "arxiv:2507.12844", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fe2d9d877a0666d46a065bd8b673c282c879ef0a2620e3b267793a91858b519", "submitted_at": "2025-07-17T07:10:27Z", "updated_at": "2025-11-12T13:34:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.12844v2"}, "authors": ["Joel Nitu", "Heidrun Mühle", "Andreas Stöckl"], "canonical_source_url": "https://arxiv.org/abs/2507.12844", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-12844/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9fe2d9d877a0666d46a065bd8b673c282c879ef0a2620e3b267793a91858b519", "submitted_at": "2025-07-17T07:10:27Z", "updated_at": "2025-11-12T13:34:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.12844v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.12844"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.12844"}], "name": "Machine-Readable Ads: Accessibility and Trust Patterns for AI Web Agents interacting with Online Advertisements", "publication": null, "published_at": "2025-07-17", "record_sha256": "d1783dffa23036332c03db2f3806e6c1beae580dd7da591fc8e4bd41685fe23e", "relations": [], "resource_id": "arxiv:2507.12844", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.12844v2", "topics": ["evaluation", "ui-design", "safety-security"], "updated_at": "2025-11-12"}, "text": "Machine-Readable Ads: Accessibility and Trust Patterns for AI Web Agents interacting with Online Advertisements\nResource type: paper\nReview status: curated\nScope: core\n\nCompares DOM-based and pixel-based web agents on a cloned news site containing advertisements, consent dialogs and paywalls. The study examines how interface presentation affects ad interaction and consequential subscription choices.\nPlatforms: web\nTopics: evaluation, ui-design, safety-security\nPublished: 2025-07-17\nAuthors: Joel Nitu, Heidrun Mühle, Andreas Stöckl\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.12844v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-12844/"}
{"id": "arxiv:2507.14293", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8f082fd40174e01f40e97d64f7aa3d90e5f76e831df8f93afdc21aa2e35bfd5", "submitted_at": "2025-07-18T18:06:27Z", "updated_at": "2025-07-18T18:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14293v1"}, "authors": ["Boyuan Zheng", "Zeyi Liao", "Scott Salisbury", "Zeyuan Liu", "Michael Lin", "Qinyuan Zheng", "Zifan Wang", "Xiang Deng", "Dawn Song", "Huan Sun", "Yu Su"], "canonical_source_url": "https://arxiv.org/abs/2507.14293", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-14293/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8f082fd40174e01f40e97d64f7aa3d90e5f76e831df8f93afdc21aa2e35bfd5", "submitted_at": "2025-07-18T18:06:27Z", "updated_at": "2025-07-18T18:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14293v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.14293"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.14293"}], "name": "WebGuard: Building a Generalizable Guardrail for Web Agents", "publication": null, "published_at": "2025-07-18", "record_sha256": "795dbcbab60402290870641d259b2483dc3c85e1853492f451224d8ac4e9d86b", "relations": [], "resource_id": "arxiv:2507.14293", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.14293v1", "topics": ["safety-security", "reward-models", "evaluation"], "updated_at": "2025-07-18"}, "text": "WebGuard: Building a Generalizable Guardrail for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebGuard, a human-annotated dataset for predicting outcomes and risk levels of state-changing website actions. It provides generalization splits and evaluates specialized guardrail models.\nPlatforms: web\nTopics: safety-security, reward-models, evaluation\nPublished: 2025-07-18\nAuthors: Boyuan Zheng, Zeyi Liao, Scott Salisbury, Zeyuan Liu, Michael Lin, Qinyuan Zheng, Zifan Wang, Xiang Deng, Dawn Song, Huan Sun, Yu Su\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.14293v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-14293/"}
{"id": "arxiv:2507.14799", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "79963ac401384f0c23ecc65d58e616daacc93103456ceb39241dcd6d71c6591c", "submitted_at": "2025-07-20T03:10:13Z", "updated_at": "2025-07-20T03:10:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14799v1"}, "authors": ["Sam Johnson", "Viet Pham", "Thai Le"], "canonical_source_url": "https://arxiv.org/abs/2507.14799", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-14799/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "79963ac401384f0c23ecc65d58e616daacc93103456ceb39241dcd6d71c6591c", "submitted_at": "2025-07-20T03:10:13Z", "updated_at": "2025-07-20T03:10:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.14799v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.14799"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.14799"}], "name": "Manipulating LLM Web Agents with Indirect Prompt Injection Attack via HTML Accessibility Tree", "publication": null, "published_at": "2025-07-20", "record_sha256": "27e31ae3012ea91218f333574aa78c5c4fc12c6adf4052c960926f256847faaf", "relations": [], "resource_id": "arxiv:2507.14799", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.14799v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2025-07-20"}, "text": "Manipulating LLM Web Agents with Indirect Prompt Injection Attack via HTML Accessibility Tree\nResource type: paper\nReview status: curated\nScope: core\n\nStudies indirect prompt injection through HTML content consumed as accessibility trees. The evaluation tests how manipulated page observations redirect browser-agent actions on real websites.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2025-07-20\nAuthors: Sam Johnson, Viet Pham, Thai Le\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.14799v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-14799/"}
{"id": "arxiv:2507.15846", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a2a49373259cb5903fad8c2fb48a7380732642bedd81a644e735af4f028c63", "submitted_at": "2025-07-21T17:53:42Z", "updated_at": "2025-07-28T16:54:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.15846v3"}, "authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Xuyang Liu", "Shuheng Shen", "Changhua Meng", "Wen Wang", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Jun Xiao", "Yueting Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2507.15846", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-15846/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2a2a49373259cb5903fad8c2fb48a7380732642bedd81a644e735af4f028c63", "submitted_at": "2025-07-21T17:53:42Z", "updated_at": "2025-07-28T16:54:13Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2507.15846v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.15846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.15846"}], "name": "GUI-G$^2$: Gaussian Reward Modeling for GUI Grounding", "publication": null, "published_at": "2025-07-21", "record_sha256": "4a19568e984ee31904b28e28ef006df27304d93c741e0625f439d4badfbf99bd", "relations": [], "resource_id": "arxiv:2507.15846", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.15846v3", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": "2025-07-28"}, "text": "GUI-G$^2$: Gaussian Reward Modeling for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-G2 replaces binary grounding rewards with Gaussian point rewards and spatial-coverage rewards. An adaptive variance tied to element dimensions supplies continuous training feedback for targets of different sizes, evaluated on the ScreenSpot benchmark family.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, reward-models\nPublished: 2025-07-21\nAuthors: Fei Tang, Zhangxuan Gu, Zhengxi Lu, Xuyang Liu, Shuheng Shen, Changhua Meng, Wen Wang, Wenqi Zhang, Yongliang Shen, Weiming Lu, Jun Xiao, Yueting Zhuang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.15846v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-15846/"}
{"id": "arxiv:2507.16853", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d58a188742fef5862b299c1dcf17bb6be139e81aafd3d2e1bdbe26f88001384", "submitted_at": "2025-07-21T09:37:05Z", "updated_at": "2025-07-21T09:37:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.16853v1"}, "authors": ["Ning Li", "Xiangmou Qu", "Jiamu Zhou", "Jun Wang", "Muning Wen", "Kounianhua Du", "Xingyu Lou", "Qiuying Peng", "Jun Wang", "Weinan Zhang"], "canonical_source_url": "https://arxiv.org/abs/2507.16853", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-16853/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d58a188742fef5862b299c1dcf17bb6be139e81aafd3d2e1bdbe26f88001384", "submitted_at": "2025-07-21T09:37:05Z", "updated_at": "2025-07-21T09:37:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.16853v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.16853"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.16853"}], "name": "MobileUse: A GUI Agent with Hierarchical Reflection for Autonomous Mobile Operation", "publication": null, "published_at": "2025-07-21", "record_sha256": "d52c14ad82df1014f748e3e9c9b5ebb621d83b17398835a0bc309e8de3800fdf", "relations": [], "resource_id": "arxiv:2507.16853", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.16853v1", "topics": ["mobile-navigation", "self-reflection", "robustness", "long-horizon"], "updated_at": "2025-07-21"}, "text": "MobileUse: A GUI Agent with Hierarchical Reflection for Autonomous Mobile Operation\nResource type: paper\nReview status: curated\nScope: core\n\nMobileUse reflects on errors at action and task timescales, invoking reflection when needed and using proactive exploration to learn unfamiliar apps. It evaluates this mobile execution framework on AndroidWorld and AndroidLab and describes a toolkit for physical-device operation.\nPlatforms: android\nTopics: mobile-navigation, self-reflection, robustness, long-horizon\nPublished: 2025-07-21\nAuthors: Ning Li, Xiangmou Qu, Jiamu Zhou, Jun Wang, Muning Wen, Kounianhua Du, Xingyu Lou, Qiuying Peng, Jun Wang, Weinan Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.16853v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-16853/"}
{"id": "arxiv:2507.18165", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3c271f330c5bf27e2831efd96fb1d6f843bd79942791686a866df1b4812de2f", "submitted_at": "2025-07-24T08:02:35Z", "updated_at": "2025-07-24T08:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.18165v1"}, "authors": ["Yuheng Zhao", "Xueli Shu", "Liwen Fan", "Lin Gao", "Yu Zhang", "Siming Chen"], "canonical_source_url": "https://arxiv.org/abs/2507.18165", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-18165/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3c271f330c5bf27e2831efd96fb1d6f843bd79942791686a866df1b4812de2f", "submitted_at": "2025-07-24T08:02:35Z", "updated_at": "2025-07-24T08:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.18165v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:16.030926+00:00", "role": "primary", "sha256": "4dcd1ba66ba21c0f2cb2aeb00c8bcab7c0cee4da37ff2fa6c16b45066aaa9db0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:30.529800+00:00", "role": "primary", "sha256": "7ca9569113479791a0482f4c196002a8459f1826eb6a3c9fab8a701b33360593", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2507.18165v1"}], "identifiers": {"arxiv": "2507.18165"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:16.030926+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.18165"}], "name": "ProactiveVA: Proactive Visual Analytics with LLM-Based UI Agent", "publication": null, "published_at": "2025-07-24", "record_sha256": "ed1ebfca61b51dfe96dd6aef9f7c4addea8864196a5a5aaff412130f518d42d2", "relations": [], "resource_id": "arxiv:2507.18165", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2507.18165v1", "topics": ["human-in-the-loop", "planning"], "updated_at": "2025-07-24"}, "text": "ProactiveVA: Proactive Visual Analytics with LLM-Based UI Agent\nResource type: paper\nReview status: curated\nScope: adjacent\n\nStudies proactive assistance in visual-analytics interfaces, using interaction logs to detect support needs and provide context-aware suggestions and guidance.\nPlatforms: unspecified\nTopics: human-in-the-loop, planning\nPublished: 2025-07-24\nAuthors: Yuheng Zhao, Xueli Shu, Liwen Fan, Lin Gao, Yu Zhang, Siming Chen\nRecorded arXiv version: v1\n\nAdjacent work on embedded assistance in two instrumented visual-analytics systems. Its in-lab user-study evidence is context-specific, and the paper discusses clarification and intervention-control trade-offs. General GUI-agent performance is not inferred. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (5.3 User study; 5.3.2 Observed issues; 6.1 Design trade-off; 6.2 Limitations); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2507.18165v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-18165/"}
{"id": "arxiv:2507.19132", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcd5a34bf6187f48da8e08842678e1e21cd558b2bcc7aab2611ceee7bc0d0179", "submitted_at": "2025-07-25T10:14:53Z", "updated_at": "2025-07-25T10:14:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19132v1"}, "authors": ["Xuetian Chen", "Yinghao Chen", "Xinfeng Yuan", "Zhuo Peng", "Lu Chen", "Yuekeng Li", "Zhoujia Zhang", "Yingqian Huang", "Leyan Huang", "Jiaqing Liang", "Tianbao Xie", "Zhiyong Wu", "Qiushi Sun", "Biqing Qi", "Bowen Zhou"], "canonical_source_url": "https://arxiv.org/abs/2507.19132", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-19132/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcd5a34bf6187f48da8e08842678e1e21cd558b2bcc7aab2611ceee7bc0d0179", "submitted_at": "2025-07-25T10:14:53Z", "updated_at": "2025-07-25T10:14:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19132v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.19132"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.19132"}], "name": "OS-MAP: How Far Can Computer-Using Agents Go in Breadth and Depth?", "publication": null, "published_at": "2025-07-25", "record_sha256": "6e9588db30f95b266de663ce24004fb696fddc28c97ffa13b6c1ffda47c17706", "relations": [], "resource_id": "arxiv:2507.19132", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.19132v1", "topics": ["evaluation", "automation", "planning"], "updated_at": "2025-07-25"}, "text": "OS-MAP: How Far Can Computer-Using Agents Go in Breadth and Depth?\nResource type: paper\nReview status: curated\nScope: core\n\nOS-MAP organizes 416 computer-use tasks across 15 applications by a five-level automation taxonomy and a hierarchy of user demands. The benchmark separates required autonomy from generalization scope to diagnose where agents struggle as task demands increase.\nPlatforms: unspecified\nTopics: evaluation, automation, planning\nPublished: 2025-07-25\nAuthors: Xuetian Chen, Yinghao Chen, Xinfeng Yuan, Zhuo Peng, Lu Chen, Yuekeng Li, Zhoujia Zhang, Yingqian Huang, Leyan Huang, Jiaqing Liang, Tianbao Xie, Zhiyong Wu, Qiushi Sun, Biqing Qi, Bowen Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.19132v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-19132/"}
{"id": "arxiv:2507.19478", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b91f7bacbac917e374aa6f656dcfbcb2e220ab3227fa2cff295c63a1c005f9f", "submitted_at": "2025-07-25T17:59:26Z", "updated_at": "2025-07-25T17:59:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19478v1"}, "authors": ["Xuehui Wang", "Zhenyu Wu", "JingJing Xie", "Zichen Ding", "Bowen Yang", "Zehao Li", "Zhaoyang Liu", "Qingyun Li", "Xuan Dong", "Zhe Chen", "Weiyun Wang", "Xiangyu Zhao", "Jixuan Chen", "Haodong Duan", "Tianbao Xie", "Chenyu Yang", "Shiqian Su", "Yue Yu", "Yuan Huang", "Yiqian Liu", "Xiao Zhang", "Yanting Zhang", "Xiangyu Yue", "Weijie Su", "Xizhou Zhu", "Wei Shen", "Jifeng Dai", "Wenhai Wang"], "canonical_source_url": "https://arxiv.org/abs/2507.19478", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-19478/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b91f7bacbac917e374aa6f656dcfbcb2e220ab3227fa2cff295c63a1c005f9f", "submitted_at": "2025-07-25T17:59:26Z", "updated_at": "2025-07-25T17:59:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.19478v1"}}, "environments": ["web", "desktop", "windows", "macos", "linux", "ios", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.19478"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.19478"}], "name": "MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents", "publication": null, "published_at": "2025-07-25", "record_sha256": "bfd6e0880753889fe23a4312ad37c1fb91e9df98c5f68bea1ae51ba48e077fca", "relations": [], "resource_id": "arxiv:2507.19478", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.19478v1", "topics": ["evaluation", "grounding", "efficiency", "planning"], "updated_at": "2025-07-25"}, "text": "MMBench-GUI: Hierarchical Multi-Platform Evaluation Framework for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMMBench-GUI evaluates interface understanding, element grounding, task automation, and collaboration across six named desktop, mobile, and web platforms. It introduces the Efficiency-Quality Area metric for examining execution efficiency alongside task quality in online GUI automation.\nPlatforms: web, desktop, windows, macos, linux, ios, android, cross-platform\nTopics: evaluation, grounding, efficiency, planning\nPublished: 2025-07-25\nAuthors: Xuehui Wang, Zhenyu Wu, JingJing Xie, Zichen Ding, Bowen Yang, Zehao Li, Zhaoyang Liu, Qingyun Li, Xuan Dong, Zhe Chen, Weiyun Wang, Xiangyu Zhao, Jixuan Chen, Haodong Duan, Tianbao Xie, Chenyu Yang, Shiqian Su, Yue Yu, Yuan Huang, Yiqian Liu, Xiao Zhang, Yanting Zhang, Xiangyu Yue, Weijie Su, Xizhou Zhu, Wei Shen, Jifeng Dai, Wenhai Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.19478v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-19478/"}
{"id": "arxiv:2507.21071", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "947c85efc189635a118b823b8ca0787ff6b1c922d1af8b21a8686b7df9a27cf8", "submitted_at": "2025-06-09T06:38:41Z", "updated_at": "2026-03-16T09:14:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.21071v2"}, "authors": ["Qinglong Yang", "Haoming Li", "Haotian Zhao", "Xiaokai Yan", "Jingtao Ding", "Fengli Xu", "Yong Li"], "canonical_source_url": "https://arxiv.org/abs/2507.21071", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-21071/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "947c85efc189635a118b823b8ca0787ff6b1c922d1af8b21a8686b7df9a27cf8", "submitted_at": "2025-06-09T06:38:41Z", "updated_at": "2026-03-16T09:14:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2507.21071v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "90a03448678a584f29ef11ae2afe1ef7040b4f3658c492f6e48eafe8fb6d2fa2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2507.21071"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.21071"}], "name": "FingerTip 20K: A Benchmark for Proactive and Personalized Mobile LLM Agents", "publication": null, "published_at": "2025-06-09", "record_sha256": "3a27cf139b678702be9244633cb4248ecead5ea8256cc33d4f2e6ce00b297590", "relations": [], "resource_id": "arxiv:2507.21071", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.21071v2", "topics": ["evaluation", "data-generation", "mobile-navigation", "interaction-history"], "updated_at": "2026-03-16"}, "text": "FingerTip 20K: A Benchmark for Proactive and Personalized Mobile LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nFingerTip 20K gathers 20K human Android demonstrations from users' long-term everyday app use, with user context, to evaluate proactive task suggestion and preference-aware task execution.\nPlatforms: android\nTopics: evaluation, data-generation, mobile-navigation, interaction-history\nPublished: 2025-06-09\nAuthors: Qinglong Yang, Haoming Li, Haotian Zhao, Xiaokai Yan, Jingtao Ding, Fengli Xu, Yong Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.21071v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-21071/"}
{"id": "arxiv:2507.21953", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197dbb0e2734f30349260444fca80a8ac830a5d4000a677d6417f46509737475", "submitted_at": "2025-07-29T16:05:32Z", "updated_at": "2025-07-29T16:05:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.21953v1"}, "authors": ["Yi Kong", "Dianxi Shi", "Guoli Yang", "Zhang ke-di", "Chenlin Huang", "Xiaopeng Li", "Songchang Jin"], "canonical_source_url": "https://arxiv.org/abs/2507.21953", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-21953/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197dbb0e2734f30349260444fca80a8ac830a5d4000a677d6417f46509737475", "submitted_at": "2025-07-29T16:05:32Z", "updated_at": "2025-07-29T16:05:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.21953v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2507.21953"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.21953"}], "name": "MapAgent: Trajectory-Constructed Memory-Augmented Planning for Mobile Task Automation", "publication": null, "published_at": "2025-07-29", "record_sha256": "e2e3975f8a0a4ad19533711082a3e145bde1f8a58daf3d4ef144dbe43b97ceaf", "relations": [], "resource_id": "arxiv:2507.21953", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.21953v1", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2025-07-29"}, "text": "MapAgent: Trajectory-Constructed Memory-Augmented Planning for Mobile Task Automation\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds reusable page memories from past mobile trajectories and retrieves them to support task planning and execution.\nPlatforms: android\nTopics: memory, planning, mobile-navigation\nPublished: 2025-07-29\nAuthors: Yi Kong, Dianxi Shi, Guoli Yang, Zhang ke-di, Chenlin Huang, Xiaopeng Li, Songchang Jin\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.21953v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-21953/"}
{"id": "arxiv:2507.22025", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "741cce21b25e9b29b99e9804a399695bf8a0ae89df6307d9f9bff68e3e856e85", "submitted_at": "2025-07-29T17:22:07Z", "updated_at": "2026-04-08T16:03:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.22025v4"}, "authors": ["Shuquan Lian", "Yuhang Wu", "Jia Ma", "Yifan Ding", "Zihan Song", "Bingqi Chen", "Xiawu Zheng", "Hui Li", "Rongrong Ji"], "canonical_source_url": "https://arxiv.org/abs/2507.22025", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-22025/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "741cce21b25e9b29b99e9804a399695bf8a0ae89df6307d9f9bff68e3e856e85", "submitted_at": "2025-07-29T17:22:07Z", "updated_at": "2026-04-08T16:03:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2507.22025v4"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.22025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.22025"}], "name": "UI-AGILE: Advancing GUI Agents with Effective Reinforcement Learning and Precise Inference-Time Grounding", "publication": null, "published_at": "2025-07-29", "record_sha256": "304e1ed85bc6b5559f33505cb4078ca9e43902bb8445a62b68c59994537b4e88", "relations": [], "resource_id": "arxiv:2507.22025", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.22025v4", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-04-08"}, "text": "UI-AGILE: Advancing GUI Agents with Effective Reinforcement Learning and Precise Inference-Time Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nCombines grounding-oriented reward and sampling improvements with inference-time image decomposition and target selection.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2025-07-29\nAuthors: Shuquan Lian, Yuhang Wu, Jia Ma, Yifan Ding, Zihan Song, Bingqi Chen, Xiawu Zheng, Hui Li, Rongrong Ji\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.22025v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-22025/"}
{"id": "arxiv:2507.22358", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9416ede89036d640412496b5f0def218b69271c5f2449689d4e4afacb43c556", "submitted_at": "2025-07-30T03:49:14Z", "updated_at": "2025-07-30T03:49:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.22358v1"}, "authors": ["Hussein Mozannar", "Gagan Bansal", "Cheng Tan", "Adam Fourney", "Victor Dibia", "Jingya Chen", "Jack Gerrits", "Tyler Payne", "Matheus Kunzler Maldaner", "Madeleine Grunde-McLaughlin", "Eric Zhu", "Griffin Bassman", "Jacob Alber", "Peter Chang", "Ricky Loynd", "Friederike Niedtner", "Ece Kamar", "Maya Murad", "Rafah Hosn", "Saleema Amershi"], "canonical_source_url": "https://arxiv.org/abs/2507.22358", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-22358/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9416ede89036d640412496b5f0def218b69271c5f2449689d4e4afacb43c556", "submitted_at": "2025-07-30T03:49:14Z", "updated_at": "2025-07-30T03:49:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.22358v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2507.22358"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.22358"}], "name": "Magentic-UI: Towards Human-in-the-loop Agentic Systems", "publication": null, "published_at": "2025-07-30", "record_sha256": "c4ab9c5cc2996e4ad3d0573332334f58f9adb9bfb297fba1dbdfe5c2e9aa8896", "relations": [], "resource_id": "arxiv:2507.22358", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.22358v1", "topics": ["human-in-the-loop", "tool-integration", "web-navigation", "safety-security"], "updated_at": "2025-07-30"}, "text": "Magentic-UI: Towards Human-in-the-loop Agentic Systems\nResource type: paper\nReview status: curated\nScope: core\n\nMagentic-UI combines browser interaction, code execution, file operations, and MCP tools in an interface for human-agent collaboration. It studies collaborative planning and task execution, action guards, and persistent memory through benchmark tasks, simulated users, user studies, and safety assessments.\nPlatforms: web\nTopics: human-in-the-loop, tool-integration, web-navigation, safety-security\nPublished: 2025-07-30\nAuthors: Hussein Mozannar, Gagan Bansal, Cheng Tan, Adam Fourney, Victor Dibia, Jingya Chen, Jack Gerrits, Tyler Payne, Matheus Kunzler Maldaner, Madeleine Grunde-McLaughlin, Eric Zhu, Griffin Bassman, Jacob Alber, Peter Chang, Ricky Loynd, Friederike Niedtner, Ece Kamar, Maya Murad, Rafah Hosn, Saleema Amershi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.22358v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-22358/"}
{"id": "arxiv:2507.23779", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a7a32f30d2ada7f739e460429b6453d0e0e56364dda8653b0511ef2e9e6df2a", "submitted_at": "2025-07-31T17:59:09Z", "updated_at": "2025-07-31T17:59:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.23779v1"}, "authors": ["Miaosen Zhang", "Ziqiang Xu", "Jialiang Zhu", "Qi Dai", "Kai Qiu", "Yifan Yang", "Chong Luo", "Tianyi Chen", "Justin Wagle", "Tim Franklin", "Baining Guo"], "canonical_source_url": "https://arxiv.org/abs/2507.23779", "citation_url": "https://awesomecua.github.io/resources/arxiv-2507-23779/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2a7a32f30d2ada7f739e460429b6453d0e0e56364dda8653b0511ef2e9e6df2a", "submitted_at": "2025-07-31T17:59:09Z", "updated_at": "2025-07-31T17:59:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2507.23779v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2507.23779"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2507.23779"}], "name": "Phi-Ground Tech Report: Advancing Perception in GUI Grounding", "publication": null, "published_at": "2025-07-31", "record_sha256": "095f0c2be772111414437d627e5c0e59ea2c0ee7ff20f4f6820b06e51a99ba1f", "relations": [], "resource_id": "arxiv:2507.23779", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2507.23779v1", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-07-31"}, "text": "Phi-Ground Tech Report: Advancing Perception in GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nReports a GUI-grounding model family and empirical design choices covering source data, perception training, and coordinate prediction.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\nPublished: 2025-07-31\nAuthors: Miaosen Zhang, Ziqiang Xu, Jialiang Zhu, Qi Dai, Kai Qiu, Yifan Yang, Chong Luo, Tianyi Chen, Justin Wagle, Tim Franklin, Baining Guo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2507.23779v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2507-23779/"}
{"id": "arxiv:2508.01330", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d89909534af90cbc10b97a992238bde3ae976ac6cc97778b4345635cef552c", "submitted_at": "2025-08-02T11:53:41Z", "updated_at": "2026-04-20T02:44:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.01330v4"}, "authors": ["Zihan Zheng", "Tianle Cui", "Taoran Wang", "Fengtao Wang", "Jiahui Pan", "Lewei He", "Qianglong Chen"], "canonical_source_url": "https://arxiv.org/abs/2508.01330", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-01330/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d89909534af90cbc10b97a992238bde3ae976ac6cc97778b4345635cef552c", "submitted_at": "2025-08-02T11:53:41Z", "updated_at": "2026-04-20T02:44:32Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.01330v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2508.01330"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.01330"}], "name": "NaturalGAIA: A Verifiable Benchmark and Hierarchical Framework for Long-Horizon GUI Tasks", "publication": null, "published_at": "2025-08-02", "record_sha256": "1faf0d0ad855c05805d643bdfd8b05df64ed1cb55556abcc46016eebd5efc223", "relations": [], "resource_id": "arxiv:2508.01330", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.01330v4", "topics": ["evaluation", "long-horizon", "planning"], "updated_at": "2026-04-20"}, "text": "NaturalGAIA: A Verifiable Benchmark and Hierarchical Framework for Long-Horizon GUI Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nNaturalGAIA builds verifiable long-horizon GUI tasks from human interaction intents while separating causal task structure from its verbal description. Its LightManus-Jarvis framework divides dynamic planning and context management from execution based on visual and structural interface information.\nPlatforms: unspecified\nTopics: evaluation, long-horizon, planning\nPublished: 2025-08-02\nAuthors: Zihan Zheng, Tianle Cui, Taoran Wang, Fengtao Wang, Jiahui Pan, Lewei He, Qianglong Chen\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.01330v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-01330/"}
{"id": "arxiv:2508.01858", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7406b48ed9826d51215d955c31c3c67cd2ce9e08bf9a93a55ac4e85a546315f1", "submitted_at": "2025-08-03T17:17:52Z", "updated_at": "2026-07-05T06:17:25Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.01858v3"}, "authors": ["Yuhan Guo", "Cong Guo", "Aiwen Sun", "Hongliang He", "Xinyu Yang", "Yue Lu", "Yingji Zhang", "Xuntao Guo", "Dong Zhang", "Jianzhuang Liu", "Jiang Duan", "Yijia Xiao", "Liangjian Wen", "Hai-Ming Xu", "Yong Dai"], "canonical_source_url": "https://arxiv.org/abs/2508.01858", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-01858/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7406b48ed9826d51215d955c31c3c67cd2ce9e08bf9a93a55ac4e85a546315f1", "submitted_at": "2025-08-03T17:17:52Z", "updated_at": "2026-07-05T06:17:25Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.01858v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:13+00:00", "role": "primary", "sha256": "1c8ad21928ad5e15bba474a060d225b7da000b37878a8e98dfe5ce94008e3086", "supports": ["scope", "environments"], "url": "https://raw.githubusercontent.com/Gnonymous/Web-CogReasoner/HEAD/README.md"}], "identifiers": {"arxiv": "2508.01858"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.01858"}], "name": "Web-CogReasoner: Towards Multimodal Knowledge-Induced Cognitive Reasoning for Web Agents", "publication": null, "published_at": "2025-08-03", "record_sha256": "b79df891e16a746a38c6f54231ef2e1768427db2f7ac5c19c57df6d29a7094ee", "relations": [], "resource_id": "arxiv:2508.01858", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.01858v3", "topics": ["data-generation", "planning", "evaluation", "web-navigation"], "updated_at": "2026-07-05"}, "text": "Web-CogReasoner: Towards Multimodal Knowledge-Induced Cognitive Reasoning for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWeb-CogReasoner separates factual and conceptual web knowledge from procedural reasoning, using a structured curriculum to train a multimodal agent. Web-CogDataset and Web-CogBench support study of these capabilities alongside web-navigation evaluation.\nPlatforms: web\nTopics: data-generation, planning, evaluation, web-navigation\nPublished: 2025-08-03\nAuthors: Yuhan Guo, Cong Guo, Aiwen Sun, Hongliang He, Xinyu Yang, Yue Lu, Yingji Zhang, Xuntao Guo, Dong Zhang, Jianzhuang Liu, Jiang Duan, Yijia Xiao, Liangjian Wen, Hai-Ming Xu, Yong Dai\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.01858v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-01858/"}
{"id": "arxiv:2508.02046", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e892ab93511f4164f37ef79ac0473e3c1eed7d6ec11fb6c8ec82092f26c5ca68", "submitted_at": "2025-08-04T04:28:18Z", "updated_at": "2026-05-02T12:32:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.02046v4"}, "authors": ["Zhihao Luo", "Wentao Yan", "Jingyu Gong", "Min Wang", "Zhizhong Zhang", "Xuhong Wang", "Yuan Xie", "Xin Tan"], "canonical_source_url": "https://arxiv.org/abs/2508.02046", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-02046/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e892ab93511f4164f37ef79ac0473e3c1eed7d6ec11fb6c8ec82092f26c5ca68", "submitted_at": "2025-08-04T04:28:18Z", "updated_at": "2026-05-02T12:32:44Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.02046v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2508.02046"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.02046"}], "name": "NaviMaster: Learning a Unified Policy for GUI and Embodied Navigation Tasks", "publication": null, "published_at": "2025-08-04", "record_sha256": "ac143a44374723bf188d194ec8c865800720b203ef36e1b9dd847d36fabdb846", "relations": [], "resource_id": "arxiv:2508.02046", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.02046v4", "topics": ["reinforcement-learning", "data-generation", "robustness"], "updated_at": "2026-05-02"}, "text": "NaviMaster: Learning a Unified Policy for GUI and Embodied Navigation Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nNaviMaster trains a shared policy for GUI and embodied navigation using a common visual-trajectory formulation. It combines trajectory collection, reinforcement learning and distance-aware rewards, with evaluation across both task families.\nPlatforms: unspecified\nTopics: reinforcement-learning, data-generation, robustness\nPublished: 2025-08-04\nAuthors: Zhihao Luo, Wentao Yan, Jingyu Gong, Min Wang, Zhizhong Zhang, Xuhong Wang, Yuan Xie, Xin Tan\nRecorded arXiv version: v4\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.02046v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-02046/"}
{"id": "arxiv:2508.03700", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5175e5ed0061660fa950512300daa878eb90e5c98bdac8f80d478414b0a2931", "submitted_at": "2025-07-19T12:33:43Z", "updated_at": "2025-09-11T14:28:11Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2508.03700v5"}, "authors": ["Liujian Tang", "Shaokang Dong", "Yijia Huang", "Minqi Xiang", "Hongtao Ruan", "Bin Wang", "Shuo Li", "Zhiheng Xi", "Zhihui Cao", "Hailiang Pang", "Heng Kong", "He Yang", "Mingxu Chai", "Zhilin Gao", "Xingyu Liu", "Yingnan Fu", "Jiaming Liu", "Xuanjing Huang", "Yu-Gang Jiang", "Tao Gui", "Qi Zhang", "Kang Wang", "Yunke Zhang", "Yuran Wang"], "canonical_source_url": "https://arxiv.org/abs/2508.03700", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-03700/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5175e5ed0061660fa950512300daa878eb90e5c98bdac8f80d478414b0a2931", "submitted_at": "2025-07-19T12:33:43Z", "updated_at": "2025-09-11T14:28:11Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2508.03700v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.03700"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.03700"}], "name": "MagicGUI: A Foundational Mobile GUI Agent with Scalable Data Pipeline and Reinforcement Fine-tuning", "publication": null, "published_at": "2025-07-19", "record_sha256": "4090ed182ea51b6d3a4319363c24e794728bcbf0713970ce85f6ffa34bcfcda2", "relations": [], "resource_id": "arxiv:2508.03700", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.03700v5", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning"], "updated_at": "2025-09-11"}, "text": "MagicGUI: A Foundational Mobile GUI Agent with Scalable Data Pipeline and Reinforcement Fine-tuning\nResource type: paper\nReview status: curated\nScope: core\n\nCombines a mobile GUI data pipeline, perception and planning capabilities, and reinforcement fine-tuning for interactive task execution.\nPlatforms: unspecified\nTopics: mobile-navigation, data-generation, reinforcement-learning\nPublished: 2025-07-19\nAuthors: Liujian Tang, Shaokang Dong, Yijia Huang, Minqi Xiang, Hongtao Ruan, Bin Wang, Shuo Li, Zhiheng Xi, Zhihui Cao, Hailiang Pang, Heng Kong, He Yang, Mingxu Chai, Zhilin Gao, Xingyu Liu, Yingnan Fu, Jiaming Liu, Xuanjing Huang, Yu-Gang Jiang, Tao Gui, Qi Zhang, Kang Wang, Yunke Zhang, Yuran Wang\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.03700v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-03700/"}
{"id": "arxiv:2508.03923", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e4272a69c1a93f22bd35032547ffb6021a79c97d179a1075e14aee69dea83a", "submitted_at": "2025-08-05T21:33:36Z", "updated_at": "2026-02-20T10:14:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.03923v3"}, "authors": ["Linxin Song", "Yutong Dai", "Viraj Prabhu", "Jieyu Zhang", "Taiwei Shi", "Li Li", "Junnan Li", "Silvio Savarese", "Zeyuan Chen", "Jieyu Zhao", "Ran Xu", "Caiming Xiong"], "canonical_source_url": "https://arxiv.org/abs/2508.03923", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-03923/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b6e4272a69c1a93f22bd35032547ffb6021a79c97d179a1075e14aee69dea83a", "submitted_at": "2025-08-05T21:33:36Z", "updated_at": "2026-02-20T10:14:08Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.03923v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.03923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.03923"}], "name": "CoAct-1: Computer-using Multi-Agent System with Coding Actions", "publication": null, "published_at": "2025-08-05", "record_sha256": "8bf64a16e960663c8e57073625eadb34ee00f36ef574eab08699507a2e298008", "relations": [], "resource_id": "arxiv:2508.03923", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.03923v3", "topics": ["tool-integration", "automation", "efficiency", "planning"], "updated_at": "2026-02-20"}, "text": "CoAct-1: Computer-using Multi-Agent System with Coding Actions\nResource type: paper\nReview status: curated\nScope: core\n\nCoAct-1 delegates subtasks between a GUI operator and a programmer that executes Python or Bash. Its OSWorld evaluation studies a hybrid action space in which scripts handle suitable file and data operations while visual interaction handles other steps.\nPlatforms: desktop\nTopics: tool-integration, automation, efficiency, planning\nPublished: 2025-08-05\nAuthors: Linxin Song, Yutong Dai, Viraj Prabhu, Jieyu Zhang, Taiwei Shi, Li Li, Junnan Li, Silvio Savarese, Zeyuan Chen, Jieyu Zhao, Ran Xu, Caiming Xiong\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.03923v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-03923/"}
{"id": "arxiv:2508.04025", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "499a462f9f14b8afdd367bb7cf7a7ffb1b2e166c212fb5c36686d3145f08ab5f", "submitted_at": "2025-08-06T02:38:02Z", "updated_at": "2025-08-06T02:38:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04025v1"}, "authors": ["Chao Hao", "Shuai Wang", "Kaiwen Zhou"], "canonical_source_url": "https://arxiv.org/abs/2508.04025", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04025/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "499a462f9f14b8afdd367bb7cf7a7ffb1b2e166c212fb5c36686d3145f08ab5f", "submitted_at": "2025-08-06T02:38:02Z", "updated_at": "2025-08-06T02:38:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04025v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.04025"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04025"}], "name": "Uncertainty-Aware GUI Agent: Adaptive Perception through Component Recommendation and Human-in-the-Loop Refinement", "publication": null, "published_at": "2025-08-06", "record_sha256": "37944b3db67203a5d17f7548003d53fda622a1d109a9fe7cdb32017d2250ba91", "relations": [], "resource_id": "arxiv:2508.04025", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04025v1", "topics": ["mobile-navigation", "human-in-the-loop", "screen-understanding", "evaluation"], "updated_at": "2025-08-06"}, "text": "Uncertainty-Aware GUI Agent: Adaptive Perception through Component Recommendation and Human-in-the-Loop Refinement\nResource type: paper\nReview status: curated\nScope: core\n\nRecAgent narrows attention to recommended UI components and requests user feedback when task intent or decisions remain uncertain. It introduces ComplexAction to evaluate specified single-step GUI actions in difficult interface scenarios.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop, screen-understanding, evaluation\nPublished: 2025-08-06\nAuthors: Chao Hao, Shuai Wang, Kaiwen Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.04025v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04025/"}
{"id": "arxiv:2508.04037", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d54ad74695a60cb040d88fcb85549912d7d4ecf6ca2ed78e636d05f8f17d1795", "submitted_at": "2025-08-06T02:57:22Z", "updated_at": "2026-01-22T05:22:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04037v2"}, "authors": ["Yuhao Cheng", "Liang Tang", "Shuxian Li", "Yukang Huo", "Tiaonan Duan", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan"], "canonical_source_url": "https://arxiv.org/abs/2508.04037", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04037/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d54ad74695a60cb040d88fcb85549912d7d4ecf6ca2ed78e636d05f8f17d1795", "submitted_at": "2025-08-06T02:57:22Z", "updated_at": "2026-01-22T05:22:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04037v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.04037"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04037"}], "name": "Evolving in Tasks: Empowering the Multi-modality Large Language Model as the Computer Use Agent", "publication": null, "published_at": "2025-08-06", "record_sha256": "90a4e794da5923c59bb8d593a8b3bc17301a86d90ba6451393f06eb9bf5250d3", "relations": [], "resource_id": "arxiv:2508.04037", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04037v2", "topics": ["data-generation", "reinforcement-learning", "grounding", "planning"], "updated_at": "2026-01-22"}, "text": "Evolving in Tasks: Empowering the Multi-modality Large Language Model as the Computer Use Agent\nResource type: paper\nReview status: curated\nScope: core\n\nThe Self-Evolution Agent combines automatic generation of verifiable computer-use trajectories with step-wise reinforcement learning for long-horizon training. It also describes a model-enhancement procedure that integrates grounding and planning capabilities without an additional training stage.\nPlatforms: unspecified\nTopics: data-generation, reinforcement-learning, grounding, planning\nPublished: 2025-08-06\nAuthors: Yuhao Cheng, Liang Tang, Shuxian Li, Yukang Huo, Tiaonan Duan, Kaer Huang, Yanzhe Jing, Yiqiang Yan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.04037v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04037/"}
{"id": "arxiv:2508.04389", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b5336c1850c15228880e0625252331b15b79d1cf1e3929abbc01ead70d5f25", "submitted_at": "2025-08-06T12:35:24Z", "updated_at": "2025-08-06T12:35:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04389v1"}, "authors": ["Weitai Kang", "Bin Lei", "Gaowen Liu", "Caiwen Ding", "Yan Yan"], "canonical_source_url": "https://arxiv.org/abs/2508.04389", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04389/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87b5336c1850c15228880e0625252331b15b79d1cf1e3929abbc01ead70d5f25", "submitted_at": "2025-08-06T12:35:24Z", "updated_at": "2025-08-06T12:35:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04389v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:51Z", "role": "primary", "sha256": "b91987360225499418fcd3645d7d5effac91e8202bf7c7e17c9f9e31d25b5be8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.01423%2C2506.03231%2C2506.04301%2C2506.05336%2C2506.07385%2C2506.07398%2C2506.08136%2C2506.08837%2C2506.10953%2C2506.11425%2C2506.11763%2C2506.13538%2C2506.14683%2C2506.15461%2C2506.18959%2C2506.19502%2C2506.20668%2C2506.21252%2C2506.21495%2C2506.21614%2C2507.01599%2C2507.02279%2C2507.02825%2C2507.04036%2C2507.07400%2C2507.09481%2C2507.09884%2C2507.13348%2C2507.14447%2C2507.16704%2C2507.19902%2C2507.21678%2C2507.22925%2C2508.01548%2C2508.02866%2C2508.04002%2C2508.04389%2C2508.05193%2C2508.05294%2C2508.05748%2C2508.07871%2C2508.07999%2C2508.08228%2C2508.09124%2C2508.09981%2C2508.15804%2C2508.16153%2C2508.17398%2C2508.18489%2C2508.19611&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:37Z", "role": "primary", "sha256": "afdfecfc117506609a5bff392254299d88cc84c971469efe0601cbcb02660ecf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2508.04389v1"}], "identifiers": {"arxiv": "2508.04389"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:37Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04389"}], "name": "GuirlVG: Incentivize GUI Visual Grounding via Empirical Exploration on Reinforcement Learning", "publication": null, "published_at": "2025-08-06", "record_sha256": "87a6febaef63f71fa0a1104e09b5a78e23490982ad728e89be155b933e16872d", "relations": [], "resource_id": "arxiv:2508.04389", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04389v1", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-06"}, "text": "GuirlVG: Incentivize GUI Visual Grounding via Empirical Exploration on Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GuirlVG, an empirical study of reinforcement fine-tuning for GUI visual grounding. It analyzes training components and adds an adaptive KL-based mechanism to limit reward over-optimization.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning\nPublished: 2025-08-06\nAuthors: Weitai Kang, Bin Lei, Gaowen Liu, Caiwen Ding, Yan Yan\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.04389v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04389/"}
{"id": "arxiv:2508.04412", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "789cfb56a83974c38e6f5e359dac5b6b3075cb70ac8afc178e36a901ce742041", "submitted_at": "2025-08-06T12:56:54Z", "updated_at": "2026-08-31T14:39:02Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.04412v4"}, "authors": ["Thassilo M. Schiepanski", "Nicholas Piël"], "canonical_source_url": "https://arxiv.org/abs/2508.04412", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04412/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "789cfb56a83974c38e6f5e359dac5b6b3075cb70ac8afc178e36a901ce742041", "submitted_at": "2025-08-06T12:56:54Z", "updated_at": "2026-08-31T14:39:02Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2508.04412v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.04412"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04412"}], "name": "Beyond Pixels: Exploring DOM Downsampling for LLM-Based Web Agents", "publication": null, "published_at": "2025-08-06", "record_sha256": "fa3a1ca859f3083aa04251900afec7d7d4ef989b9fa558df902b6b3bfdf3fdb0", "relations": [], "resource_id": "arxiv:2508.04412", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04412v4", "topics": ["web-navigation", "efficiency", "screen-understanding"], "updated_at": "2026-08-31"}, "text": "Beyond Pixels: Exploring DOM Downsampling for LLM-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nD2Snap reduces serialized DOM observations while preserving actionable elements and features that distinguish their uses. The study compares reduced DOM inputs with screenshot-based observations for a web agent on tasks sampled from Online-Mind2Web.\nPlatforms: web\nTopics: web-navigation, efficiency, screen-understanding\nPublished: 2025-08-06\nAuthors: Thassilo M. Schiepanski, Nicholas Piël\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.04412v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04412/"}
{"id": "arxiv:2508.04482", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14fdd895e52277a3695eeb83e5095472051d139aec63f22f9c63d344c051dad", "submitted_at": "2025-08-06T14:33:45Z", "updated_at": "2025-08-06T14:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04482v1"}, "authors": ["Xueyu Hu", "Tao Xiong", "Biao Yi", "Zishu Wei", "Ruixuan Xiao", "Yurun Chen", "Jiasheng Ye", "Meiling Tao", "Xiangxin Zhou", "Ziyu Zhao", "Yuhuai Li", "Shengze Xu", "Shenzhi Wang", "Xinchen Xu", "Shuofei Qiao", "Zhaokai Wang", "Kun Kuang", "Tieyong Zeng", "Liang Wang", "Jiwei Li", "Yuchen Eleanor Jiang", "Wangchunshu Zhou", "Guoyin Wang", "Keting Yin", "Zhou Zhao", "Hongxia Yang", "Fan Wu", "Shengyu Zhang", "Fei Wu"], "canonical_source_url": "https://arxiv.org/abs/2508.04482", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04482/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14fdd895e52277a3695eeb83e5095472051d139aec63f22f9c63d344c051dad", "submitted_at": "2025-08-06T14:33:45Z", "updated_at": "2025-08-06T14:33:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.04482v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2508.04482"}, "kind": "survey", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04482"}], "name": "OS Agents: A Survey on MLLM-based Agents for General Computing Devices Use", "publication": null, "published_at": "2025-08-06", "record_sha256": "b907b7afc9db647792eae44efda64e25515bca9a7856bf6ff785144ab5a92164", "relations": [], "resource_id": "arxiv:2508.04482", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04482v1", "topics": ["survey"], "updated_at": "2025-08-06"}, "text": "OS Agents: A Survey on MLLM-based Agents for General Computing Devices Use\nResource type: survey\nReview status: curated\nScope: core\n\nSurveys agents that operate computers and phones through operating-system interfaces, covering components, capabilities, foundation models, frameworks, and benchmarks.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2025-08-06\nAuthors: Xueyu Hu, Tao Xiong, Biao Yi, Zishu Wei, Ruixuan Xiao, Yurun Chen, Jiasheng Ye, Meiling Tao, Xiangxin Zhou, Ziyu Zhao, Yuhuai Li, Shengze Xu, Shenzhi Wang, Xinchen Xu, Shuofei Qiao, Zhaokai Wang, Kun Kuang, Tieyong Zeng, Liang Wang, Jiwei Li, Yuchen Eleanor Jiang, Wangchunshu Zhou, Guoyin Wang, Keting Yin, Zhou Zhao, Hongxia Yang, Fan Wu, Shengyu Zhang, Fei Wu\nRecorded arXiv version: v1\n\nPrimary source: https://arxiv.org/abs/2508.04482v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04482/"}
{"id": "arxiv:2508.04700", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268c38e6f11f11771a1251fc522112ff052557b5753dff6cadf2ba49c0ec04b2", "submitted_at": "2025-08-06T17:58:46Z", "updated_at": "2025-08-12T15:11:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04700v2"}, "authors": ["Zeyi Sun", "Ziyu Liu", "Yuhang Zang", "Yuhang Cao", "Xiaoyi Dong", "Tong Wu", "Dahua Lin", "Jiaqi Wang"], "canonical_source_url": "https://arxiv.org/abs/2508.04700", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-04700/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "268c38e6f11f11771a1251fc522112ff052557b5753dff6cadf2ba49c0ec04b2", "submitted_at": "2025-08-06T17:58:46Z", "updated_at": "2025-08-12T15:11:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.04700v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.04700"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.04700"}], "name": "SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience", "publication": null, "published_at": "2025-08-06", "record_sha256": "d68e97663906a6dfd8a5e216f96a9d8d58b227be41581d7ab4d08bf4a9aeb03b", "relations": [], "resource_id": "arxiv:2508.04700", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.04700v2", "topics": ["reinforcement-learning", "data-generation", "world-models", "imitation-learning"], "updated_at": "2025-08-12"}, "text": "SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience\nResource type: paper\nReview status: curated\nScope: core\n\nSEAgent explores unfamiliar software using a world-state assessor and an automatically generated task curriculum. It learns from failed and successful interactions through adversarial imitation and GRPO, then combines specialist experience into a generalist agent evaluated on five OSWorld software environments.\nPlatforms: desktop\nTopics: reinforcement-learning, data-generation, world-models, imitation-learning\nPublished: 2025-08-06\nAuthors: Zeyi Sun, Ziyu Liu, Yuhang Zang, Yuhang Cao, Xiaoyi Dong, Tong Wu, Dahua Lin, Jiaqi Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.04700v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-04700/"}
{"id": "arxiv:2508.05081", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d574d4157b775639c08ed12813892584b011dd3c0271ca2586f9d6a61d00c14", "submitted_at": "2025-08-07T07:05:22Z", "updated_at": "2025-08-07T07:05:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.05081v1"}, "authors": ["Jiarun Liu", "Chunhong Zhang", "Zheng Hu"], "canonical_source_url": "https://arxiv.org/abs/2508.05081", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-05081/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d574d4157b775639c08ed12813892584b011dd3c0271ca2586f9d6a61d00c14", "submitted_at": "2025-08-07T07:05:22Z", "updated_at": "2025-08-07T07:05:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.05081v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.05081"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05081"}], "name": "Cognitive Duality for Adaptive Web Agents", "publication": null, "published_at": "2025-08-07", "record_sha256": "020ecc2e4b9332ac783904bc344815be50b74c8d1ebc48a2458f6ce304f915ca", "relations": [], "resource_id": "arxiv:2508.05081", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.05081v1", "topics": ["planning", "efficiency", "web-navigation"], "updated_at": "2025-08-07"}, "text": "Cognitive Duality for Adaptive Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CogniWeb, which switches between fast reactive processing and slower deliberative reasoning according to task complexity. WebArena experiments examine this combination of learned behaviors and online planning.\nPlatforms: web\nTopics: planning, efficiency, web-navigation\nPublished: 2025-08-07\nAuthors: Jiarun Liu, Chunhong Zhang, Zheng Hu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.05081v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-05081/"}
{"id": "arxiv:2508.05615", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7deba3ec19a878cfd402c5dac61e4dca284d478429bc9204264448062299e99", "submitted_at": "2025-08-07T17:54:27Z", "updated_at": "2025-11-13T12:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05615v2"}, "authors": ["Yong Du", "Yuchen Yan", "Fei Tang", "Zhengxi Lu", "Chang Zong", "Weiming Lu", "Shengpei Jiang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2508.05615", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-05615/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c7deba3ec19a878cfd402c5dac61e4dca284d478429bc9204264448062299e99", "submitted_at": "2025-08-07T17:54:27Z", "updated_at": "2025-11-13T12:54:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05615v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.05615"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05615"}], "name": "Test-Time Reinforcement Learning for GUI Grounding via Region Consistency", "publication": null, "published_at": "2025-08-07", "record_sha256": "d30865b1a7325e558ffd848fc8bb233e61b0f0f3188a702cdd04e43828590c14", "relations": [], "resource_id": "arxiv:2508.05615", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.05615v2", "topics": ["grounding", "reinforcement-learning", "training-free"], "updated_at": "2025-11-13"}, "text": "Test-Time Reinforcement Learning for GUI Grounding via Region Consistency\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-RC aggregates multiple grounding predictions into spatial voting grids to choose a consensus region. GUI-RCPO then turns agreement with that region into a reward for test-time policy updates on unlabeled GUI screenshots.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, training-free\nPublished: 2025-08-07\nAuthors: Yong Du, Yuchen Yan, Fei Tang, Zhengxi Lu, Chang Zong, Weiming Lu, Shengpei Jiang, Yongliang Shen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.05615v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-05615/"}
{"id": "arxiv:2508.05731", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a69d2bdb4d01e6e5c9f08da33281c8247f59af52f3bc08dfed6b7e24da88e194", "submitted_at": "2025-08-07T17:49:56Z", "updated_at": "2025-12-08T10:49:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05731v2"}, "authors": ["Yuhang Liu", "Zeyu Liu", "Shuanghe Zhu", "Pengxiang Li", "Congkai Xie", "Jiasheng Wang", "Xavier Hu", "Xiaotian Han", "Jianbo Yuan", "Xinyao Wang", "Shengyu Zhang", "Hongxia Yang", "Fei Wu"], "canonical_source_url": "https://arxiv.org/abs/2508.05731", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-05731/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a69d2bdb4d01e6e5c9f08da33281c8247f59af52f3bc08dfed6b7e24da88e194", "submitted_at": "2025-08-07T17:49:56Z", "updated_at": "2025-12-08T10:49:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.05731v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.05731"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.05731"}], "name": "InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization", "publication": null, "published_at": "2025-08-07", "record_sha256": "1024c881c5bcd42aed82f7164a83d0ff8aa7118d616e490f1733394aca5952d2", "relations": [], "resource_id": "arxiv:2508.05731", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.05731v2", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-12-08"}, "text": "InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nUses multiple candidate answers and adaptive exploration rewards to improve spatial and semantic alignment in GUI grounding.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2025-08-07\nAuthors: Yuhang Liu, Zeyu Liu, Shuanghe Zhu, Pengxiang Li, Congkai Xie, Jiasheng Wang, Xavier Hu, Xiaotian Han, Jianbo Yuan, Xinyao Wang, Shengyu Zhang, Hongxia Yang, Fei Wu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.05731v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-05731/"}
{"id": "arxiv:2508.09057", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4efa329c3b106219c9b25f675be0c561743085ac999d2964458a5f22ff6e62f", "submitted_at": "2025-08-12T16:18:30Z", "updated_at": "2025-08-14T16:36:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.09057v2"}, "authors": ["Zeyu Huang", "Juyuan Wang", "Longfeng Chen", "Boyi Xiao", "Leng Cai", "Yawen Zeng", "Jin Xu"], "canonical_source_url": "https://arxiv.org/abs/2508.09057", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-09057/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4efa329c3b106219c9b25f675be0c561743085ac999d2964458a5f22ff6e62f", "submitted_at": "2025-08-12T16:18:30Z", "updated_at": "2025-08-14T16:36:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.09057v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2508.09057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09057"}], "name": "MVISU-Bench: Benchmarking Mobile Agents for Real-World Tasks by Multi-App, Vague, Interactive, Single-App and Unethical Instructions", "publication": null, "published_at": "2025-08-12", "record_sha256": "405e24eb5f164cc9bcaa351eed594e48d5390d851e8dddb0bdd8c2107cbd07c3", "relations": [], "resource_id": "arxiv:2508.09057", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.09057v2", "topics": ["evaluation", "mobile-navigation", "human-in-the-loop", "safety-security"], "updated_at": "2025-08-14"}, "text": "MVISU-Bench: Benchmarking Mobile Agents for Real-World Tasks by Multi-App, Vague, Interactive, Single-App and Unethical Instructions\nResource type: paper\nReview status: curated\nScope: core\n\nMVISU-Bench evaluates bilingual mobile tasks involving cross-app work, vague requests, user interaction and unethical instructions alongside ordinary single-app tasks. Its Aider prompting module is designed to clarify intent and mitigate risky execution within agent workflows.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, human-in-the-loop, safety-security\nPublished: 2025-08-12\nAuthors: Zeyu Huang, Juyuan Wang, Longfeng Chen, Boyi Xiao, Leng Cai, Yawen Zeng, Jin Xu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.09057v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-09057/"}
{"id": "arxiv:2508.09123", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e46f27128d066358fcee6279b26bb2b708dacaf1f7b5794b755520fe892243", "submitted_at": "2025-08-12T17:52:32Z", "updated_at": "2025-10-04T17:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.09123v3"}, "authors": ["Xinyuan Wang", "Bowen Wang", "Dunjie Lu", "Junlin Yang", "Tianbao Xie", "Junli Wang", "Jiaqi Deng", "Xiaole Guo", "Yiheng Xu", "Chen Henry Wu", "Zhennan Shen", "Zhuokai Li", "Ryan Li", "Xiaochuan Li", "Junda Chen", "Boyuan Zheng", "Peihang Li", "Fangyu Lei", "Ruisheng Cao", "Yeqiao Fu", "Dongchan Shin", "Martin Shin", "Jiarui Hu", "Yuyan Wang", "Jixuan Chen", "Yuxiao Ye", "Danyang Zhang", "Dikang Du", "Hao Hu", "Huarong Chen", "Zaida Zhou", "Haotian Yao", "Ziwei Chen", "Qizheng Gu", "Yipu Wang", "Heng Wang", "Diyi Yang", "Victor Zhong", "Flood Sung", "Y. Charles", "Zhilin Yang", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2508.09123", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-09123/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40e46f27128d066358fcee6279b26bb2b708dacaf1f7b5794b755520fe892243", "submitted_at": "2025-08-12T17:52:32Z", "updated_at": "2025-10-04T17:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.09123v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.09123"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09123"}], "name": "OpenCUA: Open Foundations for Computer-Use Agents", "publication": null, "published_at": "2025-08-12", "record_sha256": "0049b7910b34e349f5dab42c0151007a13f8fba2ac60ac54ee0354749dd5e484", "relations": [], "resource_id": "arxiv:2508.09123", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.09123v3", "topics": ["data-generation", "planning"], "updated_at": "2025-10-04"}, "text": "OpenCUA: Open Foundations for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents demonstration collection infrastructure, the AgentNet dataset, and a training pipeline for open computer-use models.\nPlatforms: desktop\nTopics: data-generation, planning\nPublished: 2025-08-12\nAuthors: Xinyuan Wang, Bowen Wang, Dunjie Lu, Junlin Yang, Tianbao Xie, Junli Wang, Jiaqi Deng, Xiaole Guo, Yiheng Xu, Chen Henry Wu, Zhennan Shen, Zhuokai Li, Ryan Li, Xiaochuan Li, Junda Chen, Boyuan Zheng, Peihang Li, Fangyu Lei, Ruisheng Cao, Yeqiao Fu, Dongchan Shin, Martin Shin, Jiarui Hu, Yuyan Wang, Jixuan Chen, Yuxiao Ye, Danyang Zhang, Dikang Du, Hao Hu, Huarong Chen, Zaida Zhou, Haotian Yao, Ziwei Chen, Qizheng Gu, Yipu Wang, Heng Wang, Diyi Yang, Victor Zhong, Flood Sung, Y. Charles, Zhilin Yang, Tao Yu\nRecorded arXiv version: v3\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.09123v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-09123/"}
{"id": "arxiv:2508.09241", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8c91a1f85602760c706e56426c0d2d870bbc101ca500aa3c7c391938c3f1676", "submitted_at": "2025-08-12T15:12:42Z", "updated_at": "2025-08-12T15:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.09241v1"}, "authors": ["Fengxian Ji", "Jingpu Yang", "Zirui Song", "Yuanxi Wang", "Zhexuan Cui", "Yuke Li", "Qian Jiang", "Miao Fang", "Xiuying Chen"], "canonical_source_url": "https://arxiv.org/abs/2508.09241", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-09241/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8c91a1f85602760c706e56426c0d2d870bbc101ca500aa3c7c391938c3f1676", "submitted_at": "2025-08-12T15:12:42Z", "updated_at": "2025-08-12T15:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.09241v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.09241"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.09241"}], "name": "FineState-Bench: A Comprehensive Benchmark for Fine-Grained State Control in GUI Agents", "publication": null, "published_at": "2025-08-12", "record_sha256": "299dc1c98645b4727f02abab278bcb26392831814e7c899c831d270a0aaa1701", "relations": [], "resource_id": "arxiv:2508.09241", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.09241v1", "topics": ["evaluation", "grounding", "observability", "screen-understanding"], "updated_at": "2025-08-12"}, "text": "FineState-Bench: A Comprehensive Benchmark for Fine-Grained State Control in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nFineState-Bench evaluates fine-grained GUI perception and control across desktop, web, and mobile interfaces. Its Visual Diagnostic Assistant separates perception and localization contributions in controlled interaction experiments.\nPlatforms: desktop, web, cross-platform\nTopics: evaluation, grounding, observability, screen-understanding\nPublished: 2025-08-12\nAuthors: Fengxian Ji, Jingpu Yang, Zirui Song, Yuanxi Wang, Zhexuan Cui, Yuke Li, Qian Jiang, Miao Fang, Xiuying Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.09241v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-09241/"}
{"id": "arxiv:2508.10833", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db88c4ba15c48210f0b618a8b5d2b9c02184f2e4375d45adf22b9a0a395abc45", "submitted_at": "2025-08-14T16:58:07Z", "updated_at": "2025-08-15T14:49:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.10833v2"}, "authors": ["Zhangxuan Gu", "Zhengwen Zeng", "Zhenyu Xu", "Xingran Zhou", "Shuheng Shen", "Yunfei Liu", "Beitong Zhou", "Changhua Meng", "Tianyu Xia", "Weizhi Chen", "Yue Wen", "Jingya Dou", "Fei Tang", "Jinzhen Lin", "Yulin Liu", "Zhenlin Guo", "Yichen Gong", "Heng Jia", "Changlong Gao", "Yuan Guo", "Yong Deng", "Zhenyu Guo", "Liang Chen", "Weiqiang Wang"], "canonical_source_url": "https://arxiv.org/abs/2508.10833", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-10833/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db88c4ba15c48210f0b618a8b5d2b9c02184f2e4375d45adf22b9a0a395abc45", "submitted_at": "2025-08-14T16:58:07Z", "updated_at": "2025-08-15T14:49:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.10833v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2508.10833"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.10833"}], "name": "UI-Venus Technical Report: Building High-performance UI Agents with RFT", "publication": null, "published_at": "2025-08-14", "record_sha256": "4405fb0d971fe352c34657b22f1279fd80cd052dc7cf78f7cc1047b6c947087c", "relations": [], "resource_id": "arxiv:2508.10833", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.10833v2", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-08-15"}, "text": "UI-Venus Technical Report: Building High-performance UI Agents with RFT\nResource type: paper\nReview status: curated\nScope: core\n\nStudies screenshot-based GUI grounding and navigation using curated interaction data, reward design, and reinforcement fine-tuning.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2025-08-14\nAuthors: Zhangxuan Gu, Zhengwen Zeng, Zhenyu Xu, Xingran Zhou, Shuheng Shen, Yunfei Liu, Beitong Zhou, Changhua Meng, Tianyu Xia, Weizhi Chen, Yue Wen, Jingya Dou, Fei Tang, Jinzhen Lin, Yulin Liu, Zhenlin Guo, Yichen Gong, Heng Jia, Changlong Gao, Yuan Guo, Yong Deng, Zhenyu Guo, Liang Chen, Weiqiang Wang\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.10833v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-10833/"}
{"id": "arxiv:2508.11360", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea849d8e7640d7a40d582431233418cf1cbd29585ca6c78638919b8832265445", "submitted_at": "2025-08-15T09:55:02Z", "updated_at": "2026-03-13T11:10:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.11360v2"}, "authors": ["Songqin Nong", "Xiaoxuan Tang", "Jingxuan Xu", "Sheng Zhou", "Jianfeng Chen", "Tao Jiang", "Wenhao Xu"], "canonical_source_url": "https://arxiv.org/abs/2508.11360", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-11360/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea849d8e7640d7a40d582431233418cf1cbd29585ca6c78638919b8832265445", "submitted_at": "2025-08-15T09:55:02Z", "updated_at": "2026-03-13T11:10:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.11360v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2508.11360"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.11360"}], "name": "CRAFT-GUI: Curriculum-Reinforced Agent For GUI Tasks", "publication": null, "published_at": "2025-08-15", "record_sha256": "87242c1b590fdbe5f76695db9cddf32b9f281eecb264d5c6e9c3a09928ee5f6b", "relations": [], "resource_id": "arxiv:2508.11360", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.11360v2", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": "2026-03-13"}, "text": "CRAFT-GUI: Curriculum-Reinforced Agent For GUI Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nCRAFT-GUI organizes GUI trajectories by difficulty for curriculum-based GRPO training. Its reward combines rule-based signals with model judgments, with evaluation on Android Control and an internal online benchmark.\nPlatforms: android\nTopics: reinforcement-learning, reward-models, mobile-navigation\nPublished: 2025-08-15\nAuthors: Songqin Nong, Xiaoxuan Tang, Jingxuan Xu, Sheng Zhou, Jianfeng Chen, Tao Jiang, Wenhao Xu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.11360v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-11360/"}
{"id": "arxiv:2508.13024", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0307d242a90f0650b5d5fdfbaf9653842d3ce1461d2653b0b1ba6b21c31ee2dc", "submitted_at": "2025-08-18T15:41:22Z", "updated_at": "2026-04-30T09:17:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13024v3"}, "authors": ["Ralph Peeters", "Aaron Steiner", "Luca Schwarz", "Julian Yuya Caspary", "Christian Bizer"], "canonical_source_url": "https://arxiv.org/abs/2508.13024", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-13024/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0307d242a90f0650b5d5fdfbaf9653842d3ce1461d2653b0b1ba6b21c31ee2dc", "submitted_at": "2025-08-18T15:41:22Z", "updated_at": "2026-04-30T09:17:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13024v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.13024"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.13024"}], "name": "WebMall -- A Multi-Shop Benchmark for Evaluating Web Agents", "publication": null, "published_at": "2025-08-18", "record_sha256": "9e582f52239a63ffce28e8a75790d7975a7bc166f8aa8e928475ee0dcd37a7ea", "relations": [], "resource_id": "arxiv:2508.13024", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.13024v3", "topics": ["evaluation", "web-navigation", "sandbox"], "updated_at": "2026-04-30"}, "text": "WebMall -- A Multi-Shop Benchmark for Evaluating Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebMall provides four simulated online shops with heterogeneous product data for comparison shopping, related-product searches, and checkout tasks. It evaluates web agents with different observation formats and memory configurations in a fixed offline environment.\nPlatforms: web\nTopics: evaluation, web-navigation, sandbox\nPublished: 2025-08-18\nAuthors: Ralph Peeters, Aaron Steiner, Luca Schwarz, Julian Yuya Caspary, Christian Bizer\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.13024v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-13024/"}
{"id": "arxiv:2508.13634", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfb50e10184cded6c18d4c49401145bc7cfa00b0ba21bce964abe8f459af5626", "submitted_at": "2025-08-19T08:47:44Z", "updated_at": "2026-01-20T06:39:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13634v3"}, "authors": ["Jikai Chen", "Long Chen", "Dong Wang", "Qinglin Su", "Zhixuan Chu", "Bingguang Hao", "Leilei Gan", "Chenyi Zhuang", "Jinjie Gu"], "canonical_source_url": "https://arxiv.org/abs/2508.13634", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-13634/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfb50e10184cded6c18d4c49401145bc7cfa00b0ba21bce964abe8f459af5626", "submitted_at": "2025-08-19T08:47:44Z", "updated_at": "2026-01-20T06:39:19Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2508.13634v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.13634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.13634"}], "name": "V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking", "publication": null, "published_at": "2025-08-19", "record_sha256": "a0f46a36dd64693bca1d553d7209115e378e2384631ec3119671aea094fed9c6", "relations": [], "resource_id": "arxiv:2508.13634", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.13634v3", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-01-20"}, "text": "V2P: Visual Attention Calibration for GUI Grounding via Background Suppression and Center Peaking\nResource type: paper\nReview status: curated\nScope: core\n\nV2P trains GUI grounding by suppressing attention to background regions and emphasizing the center of the target element. It represents target importance with size-dependent Gaussian heatmaps and evaluates the contribution of each component on ScreenSpot-v2 and ScreenSpot-Pro.\nPlatforms: unspecified\nTopics: grounding, screen-understanding\nPublished: 2025-08-19\nAuthors: Jikai Chen, Long Chen, Dong Wang, Qinglin Su, Zhixuan Chu, Bingguang Hao, Leilei Gan, Chenyi Zhuang, Jinjie Gu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.13634v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-13634/"}
{"id": "arxiv:2508.14040", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "383d8e80619d58e72ebb9c49229cd9522a0b3586b4bfa8f00b29b1834adca5b9", "submitted_at": "2025-08-19T17:59:45Z", "updated_at": "2025-10-21T05:18:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.14040v2"}, "authors": ["Hanyu Lai", "Xiao Liu", "Yanxiao Zhao", "Han Xu", "Hanchen Zhang", "Bohao Jing", "Yanyu Ren", "Shuntian Yao", "Yuxiao Dong", "Jie Tang"], "canonical_source_url": "https://arxiv.org/abs/2508.14040", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-14040/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "383d8e80619d58e72ebb9c49229cd9522a0b3586b4bfa8f00b29b1834adca5b9", "submitted_at": "2025-08-19T17:59:45Z", "updated_at": "2025-10-21T05:18:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.14040v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.14040"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.14040"}], "name": "ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents", "publication": null, "published_at": "2025-08-19", "record_sha256": "50eea590972465317d7e0a5413b0f1ac3568b498a90a32596d7a1e0ea71bb52d", "relations": [], "resource_id": "arxiv:2508.14040", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.14040v2", "topics": ["reinforcement-learning", "tool-integration"], "updated_at": "2025-10-21"}, "text": "ComputerRL: Scaling End-to-End Online Reinforcement Learning for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines API and GUI actions with distributed desktop reinforcement learning and alternating supervised training to address training instability.\nPlatforms: desktop\nTopics: reinforcement-learning, tool-integration\nPublished: 2025-08-19\nAuthors: Hanyu Lai, Xiao Liu, Yanxiao Zhao, Han Xu, Hanchen Zhang, Bohao Jing, Yanyu Ren, Shuntian Yao, Yuxiao Dong, Jie Tang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.14040v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-14040/"}
{"id": "arxiv:2508.14104", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5034c06e63607a1279668964e91c14ed762a95b6e3953337ef8b8b63c4fd57d", "submitted_at": "2025-08-17T07:31:11Z", "updated_at": "2025-08-17T07:31:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.14104v1"}, "authors": ["Yutong Bian", "Xianhao Lin", "Yupeng Xie", "Tianyang Liu", "Mingchen Zhuge", "Siyuan Lu", "Haoming Tang", "Jinlin Wang", "Jiayi Zhang", "Jiaqi Chen", "Xiangru Tang", "Yongxin Ni", "Sirui Hong", "Chenglin Wu"], "canonical_source_url": "https://arxiv.org/abs/2508.14104", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-14104/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5034c06e63607a1279668964e91c14ed762a95b6e3953337ef8b8b63c4fd57d", "submitted_at": "2025-08-17T07:31:11Z", "updated_at": "2025-08-17T07:31:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.14104v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2508.14104"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.14104"}], "name": "You Don't Know Until You Click:Automated GUI Testing for Production-Ready Software Evaluation", "publication": null, "published_at": "2025-08-17", "record_sha256": "eb0f0a37108dfe64d5799fd5573ef8028a9d378f3ce2b7b545bffda181f0cf0c", "relations": [], "resource_id": "arxiv:2508.14104", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.14104v1", "topics": ["evaluation", "automation", "observability"], "updated_at": "2025-08-17"}, "text": "You Don't Know Until You Click:Automated GUI Testing for Production-Ready Software Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nRealDevWorld evaluates generated software repositories with 194 development tasks and AppEvalPilot, an agent judge that operates the resulting GUIs. The evaluation examines functional behavior, visual fidelity, and runtime responses to produce task-specific diagnostic feedback.\nPlatforms: unspecified\nTopics: evaluation, automation, observability\nPublished: 2025-08-17\nAuthors: Yutong Bian, Xianhao Lin, Yupeng Xie, Tianyang Liu, Mingchen Zhuge, Siyuan Lu, Haoming Tang, Jinlin Wang, Jiayi Zhang, Jiaqi Chen, Xiangru Tang, Yongxin Ni, Sirui Hong, Chenglin Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.14104v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-14104/"}
{"id": "arxiv:2508.15144", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "629421d60bb8beeba3f087854ebaddc5461f9af0e5fb0bc572d0c20394ba9dcd", "submitted_at": "2025-08-21T00:39:12Z", "updated_at": "2025-09-01T05:38:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15144v2"}, "authors": ["Jiabo Ye", "Xi Zhang", "Haiyang Xu", "Haowei Liu", "Junyang Wang", "Zhaoqing Zhu", "Ziwei Zheng", "Feiyu Gao", "Junjie Cao", "Zhengxi Lu", "Jitong Liao", "Qi Zheng", "Fei Huang", "Jingren Zhou", "Ming Yan"], "canonical_source_url": "https://arxiv.org/abs/2508.15144", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-15144/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "629421d60bb8beeba3f087854ebaddc5461f9af0e5fb0bc572d0c20394ba9dcd", "submitted_at": "2025-08-21T00:39:12Z", "updated_at": "2025-09-01T05:38:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15144v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2508.15144"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.15144"}], "name": "Mobile-Agent-v3: Fundamental Agents for GUI Automation", "publication": null, "published_at": "2025-08-21", "record_sha256": "af0bca4b46a767e94fd430a26f43cd883bfbf6f9270673c8200830fb5b6b8912", "relations": [], "resource_id": "arxiv:2508.15144", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.15144v2", "topics": ["data-generation", "planning", "mobile-navigation"], "updated_at": "2025-09-01"}, "text": "Mobile-Agent-v3: Fundamental Agents for GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-Owl models and Mobile-Agent-v3 orchestration, using environment-generated trajectories to support GUI capabilities.\nPlatforms: cross-platform\nTopics: data-generation, planning, mobile-navigation\nPublished: 2025-08-21\nAuthors: Jiabo Ye, Xi Zhang, Haiyang Xu, Haowei Liu, Junyang Wang, Zhaoqing Zhu, Ziwei Zheng, Feiyu Gao, Junjie Cao, Zhengxi Lu, Jitong Liao, Qi Zheng, Fei Huang, Jingren Zhou, Ming Yan\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.15144v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-15144/"}
{"id": "arxiv:2508.15832", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c8336a707723227a9d651a3e9347e45907f28ed7421b577761ed57f7d70593d", "submitted_at": "2025-08-18T21:58:43Z", "updated_at": "2026-04-21T00:38:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15832v2"}, "authors": ["Xianren Zhang", "Shreyas Prasad", "Di Wang", "Qiuhai Zeng", "Suhang Wang", "Wenbo Yan", "Mat Hans"], "canonical_source_url": "https://arxiv.org/abs/2508.15832", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-15832/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6c8336a707723227a9d651a3e9347e45907f28ed7421b577761ed57f7d70593d", "submitted_at": "2025-08-18T21:58:43Z", "updated_at": "2026-04-21T00:38:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.15832v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.15832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.15832"}], "name": "A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains", "publication": null, "published_at": "2025-08-18", "record_sha256": "3709a4157fd2e08826ff63ac6027dc072f75696e271f925876edf23e4ed07cbe", "relations": [], "resource_id": "arxiv:2508.15832", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.15832v2", "topics": ["evaluation", "safety-security", "data-generation", "web-navigation"], "updated_at": "2026-04-21"}, "text": "A Functionality-Grounded Benchmark for Evaluating Web Agents in E-commerce Domains\nResource type: paper\nReview status: curated\nScope: core\n\nAmazon-Bench generates functionality-specific web tasks from page content and interactive controls, covering operations beyond product search. Its evaluator checks task completion and unintended changes to account state, such as address or purchase settings.\nPlatforms: web\nTopics: evaluation, safety-security, data-generation, web-navigation\nPublished: 2025-08-18\nAuthors: Xianren Zhang, Shreyas Prasad, Di Wang, Qiuhai Zeng, Suhang Wang, Wenbo Yan, Mat Hans\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.15832v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-15832/"}
{"id": "arxiv:2508.16987", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67da98d98fec647de9b482fabd8c7861063c0fdb91f1ec303fe8df759ee0f69d", "submitted_at": "2025-08-23T11:02:59Z", "updated_at": "2025-08-23T11:02:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.16987v1"}, "authors": ["Tanvir Bhathal", "Asanshay Gupta"], "canonical_source_url": "https://arxiv.org/abs/2508.16987", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-16987/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67da98d98fec647de9b482fabd8c7861063c0fdb91f1ec303fe8df759ee0f69d", "submitted_at": "2025-08-23T11:02:59Z", "updated_at": "2025-08-23T11:02:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.16987v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.16987"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.16987"}], "name": "WebSight: A Vision-First Architecture for Robust Web Agents", "publication": null, "published_at": "2025-08-23", "record_sha256": "d3931400be2ad81caecb45f029683415e4017c6bfb771cab0de8791517a68eae", "relations": [], "resource_id": "arxiv:2508.16987", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.16987v1", "topics": ["web-navigation", "grounding", "planning", "memory"], "updated_at": "2025-08-23"}, "text": "WebSight: A Vision-First Architecture for Robust Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebSight uses a vision-language model trained on web-interface examples within a modular agent with planning, action, verification, and episodic memory components. It evaluates element clicking on Showdown Clicks and full browsing tasks on WebVoyager using visual observations.\nPlatforms: web\nTopics: web-navigation, grounding, planning, memory\nPublished: 2025-08-23\nAuthors: Tanvir Bhathal, Asanshay Gupta\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.16987v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-16987/"}
{"id": "arxiv:2508.18040", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d5fc98a2eb2232f237d71b34060df0acef89cfeaa4bc247580a4dda3223e0b", "submitted_at": "2025-08-25T13:57:02Z", "updated_at": "2025-08-25T13:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.18040v1"}, "authors": ["Xin Wang", "Zhiyao Cui", "Hao Li", "Ya Zeng", "Chenxu Wang", "Ruiqi Song", "Yihang Chen", "Kun Shao", "Qiaosheng Zhang", "Jinzhuo Liu", "Siyue Ren", "Shuyue Hu", "Zhen Wang"], "canonical_source_url": "https://arxiv.org/abs/2508.18040", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-18040/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57d5fc98a2eb2232f237d71b34060df0acef89cfeaa4bc247580a4dda3223e0b", "submitted_at": "2025-08-25T13:57:02Z", "updated_at": "2025-08-25T13:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.18040v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:18:36.598801+00:00", "role": "primary", "sha256": "3c3474b771af7bdd500e6aaa91b83d39ec0d455e59bb83df906e7a94700ad2b7", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2508.18040v1"}], "identifiers": {"arxiv": "2508.18040"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.18040"}], "name": "PerPilot: Personalizing VLM-based Mobile Agents via Memory and Exploration", "publication": null, "published_at": "2025-08-25", "record_sha256": "20726a702196bf0aad3eb41bd78f30f4c04cb4a0dbd3ccb81feb25484816f223", "relations": [], "resource_id": "arxiv:2508.18040", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.18040v1", "topics": ["memory", "planning", "human-in-the-loop", "mobile-navigation"], "updated_at": "2025-08-25"}, "text": "PerPilot: Personalizing VLM-based Mobile Agents via Memory and Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PerPilot and the PerInstruct dataset for completing user-specific mobile instructions through memory retrieval, app exploration, and clarification when missing context remains unresolved.\nPlatforms: unspecified\nTopics: memory, planning, human-in-the-loop, mobile-navigation\nPublished: 2025-08-25\nAuthors: Xin Wang, Zhiyao Cui, Hao Li, Ya Zeng, Chenxu Wang, Ruiqi Song, Yihang Chen, Kun Shao, Qiaosheng Zhang, Jinzhuo Liu, Siyue Ren, Shuyue Hu, Zhen Wang\nRecorded arXiv version: v1\n\nPersonalization depends on the validity of stored or inferred user information and retrieval access. This is instruction completion with interaction, not demonstrated permanent model-weight personalization. No specific operating system or reproduced result is asserted. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (The Framework of PerPilot; Personalization Perception; Personalization Completion; PerInstruct Data Construction); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2508.18040v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-18040/"}
{"id": "arxiv:2508.18689", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "02f4a76ae220caa289b322b622be6356086c5b04bc0f1598d0cd580c9ee25b6d", "submitted_at": "2025-08-26T05:23:24Z", "updated_at": "2025-08-27T04:25:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.18689v2"}, "authors": ["Yuyang Zhao", "Wentao Shi", "Fuli Feng", "Xiangnan He"], "canonical_source_url": "https://arxiv.org/abs/2508.18689", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-18689/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "02f4a76ae220caa289b322b622be6356086c5b04bc0f1598d0cd580c9ee25b6d", "submitted_at": "2025-08-26T05:23:24Z", "updated_at": "2025-08-27T04:25:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.18689v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.18689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.18689"}], "name": "AppAgent-Pro: A Proactive GUI Agent System for Multidomain Information Integration and User Assistance", "publication": null, "published_at": "2025-08-26", "record_sha256": "e10f71d31e4f45d34ed77e21f6742b91320fbec45fac9017bed9db13929a26fe", "relations": [], "resource_id": "arxiv:2508.18689", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.18689v2", "topics": ["planning", "workflow"], "updated_at": "2025-08-27"}, "text": "AppAgent-Pro: A Proactive GUI Agent System for Multidomain Information Integration and User Assistance\nResource type: paper\nReview status: curated\nScope: core\n\nProposes a proactive GUI assistant that gathers information across domains in response to user requests and anticipated information needs.\nPlatforms: android\nTopics: planning, workflow\nPublished: 2025-08-26\nAuthors: Yuyang Zhao, Wentao Shi, Fuli Feng, Xiangnan He\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.18689v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-18689/"}
{"id": "arxiv:2508.19493", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815c8c1fa5993d2ef58798abd43a2b8945c476c27c777d30c1cbfacadd87af84", "submitted_at": "2025-08-27T00:41:28Z", "updated_at": "2025-09-03T07:48:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19493v2"}, "authors": ["Zhixin Lin", "Jungang Li", "Shidong Pan", "Yibo Shi", "Yue Yao", "Dongliang Xu"], "canonical_source_url": "https://arxiv.org/abs/2508.19493", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-19493/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "815c8c1fa5993d2ef58798abd43a2b8945c476c27c777d30c1cbfacadd87af84", "submitted_at": "2025-08-27T00:41:28Z", "updated_at": "2025-09-03T07:48:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19493v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2508.19493"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.19493"}], "name": "Mind the Third Eye! Benchmarking Privacy Awareness in MLLM-powered Smartphone Agents", "publication": null, "published_at": "2025-08-27", "record_sha256": "a9bad715a56e88afbbb77755696939a7386a756c754258884c1db4626a69bdbc", "relations": [], "resource_id": "arxiv:2508.19493", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.19493v2", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2025-09-03"}, "text": "Mind the Third Eye! Benchmarking Privacy Awareness in MLLM-powered Smartphone Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SAPA-Bench with 7,138 smartphone scenarios to evaluate whether MLLM-powered mobile agents detect and protect sensitive personal information on screen.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2025-08-27\nAuthors: Zhixin Lin, Jungang Li, Shidong Pan, Yibo Shi, Yue Yao, Dongliang Xu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.19493v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-19493/"}
{"id": "arxiv:2508.19679", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "385ed1b6e60c27660f0cbf281b0f2bbde052691801f7faae99cc053dfbc0d127", "submitted_at": "2025-08-27T08:40:05Z", "updated_at": "2026-04-25T13:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19679v2"}, "authors": ["Qihang Ai", "Pi Bu", "Yue Cao", "Yingyao Wang", "Jihao Gu", "Jingxuan Xing", "Zekun Zhu", "Wei Jiang", "Zhicheng Zheng", "Jun Song", "Yuning Jiang"], "canonical_source_url": "https://arxiv.org/abs/2508.19679", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-19679/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "385ed1b6e60c27660f0cbf281b0f2bbde052691801f7faae99cc053dfbc0d127", "submitted_at": "2025-08-27T08:40:05Z", "updated_at": "2026-04-25T13:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2508.19679v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2508.19679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.19679"}], "name": "InquireMobile: Teaching VLM-based Mobile Agent to Request Human Assistance via Reinforcement Fine-Tuning", "publication": null, "published_at": "2025-08-27", "record_sha256": "6268a66efdf2b674d693a08cf786f54c5a673c149a8c30322764c69699097be8", "relations": [], "resource_id": "arxiv:2508.19679", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.19679v2", "topics": ["human-in-the-loop", "reinforcement-learning", "evaluation", "mobile-navigation"], "updated_at": "2026-04-25"}, "text": "InquireMobile: Teaching VLM-based Mobile Agent to Request Human Assistance via Reinforcement Fine-Tuning\nResource type: paper\nReview status: curated\nScope: core\n\nInquireMobile trains mobile agents to seek human confirmation at critical decisions using staged reinforcement fine-tuning and pre-action reasoning. InquireBench evaluates safe interaction and proactive inquiry, distinguishing asking for help from completing the underlying task.\nPlatforms: unspecified\nTopics: human-in-the-loop, reinforcement-learning, evaluation, mobile-navigation\nPublished: 2025-08-27\nAuthors: Qihang Ai, Pi Bu, Yue Cao, Yingyao Wang, Jihao Gu, Jingxuan Xing, Zekun Zhu, Wei Jiang, Zhicheng Zheng, Jun Song, Yuning Jiang\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.19679v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-19679/"}
{"id": "arxiv:2508.20096", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac401f8d5c56050e72839f3bffba4f2c78bbdf34a00cab05fe504f03efd96a19", "submitted_at": "2025-08-27T17:59:50Z", "updated_at": "2025-08-27T17:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.20096v1"}, "authors": ["Zeyi Sun", "Yuhang Cao", "Jianze Liang", "Qiushi Sun", "Ziyu Liu", "Zhixiong Zhang", "Yuhang Zang", "Xiaoyi Dong", "Kai Chen", "Dahua Lin", "Jiaqi Wang"], "canonical_source_url": "https://arxiv.org/abs/2508.20096", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-20096/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac401f8d5c56050e72839f3bffba4f2c78bbdf34a00cab05fe504f03efd96a19", "submitted_at": "2025-08-27T17:59:50Z", "updated_at": "2025-08-27T17:59:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.20096v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.20096"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.20096"}], "name": "CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning", "publication": null, "published_at": "2025-08-27", "record_sha256": "c13ce4a7e8be10ba6ccf9e94dab270bd0fa5f672f4494e75d81ac9578172c693", "relations": [], "resource_id": "arxiv:2508.20096", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.20096v1", "topics": ["reinforcement-learning", "planning", "science", "imitation-learning"], "updated_at": "2025-08-27"}, "text": "CODA: Coordinating the Cerebrum and Cerebellum for a Dual-Brain Computer Use Agent with Decoupled Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nCODA combines a generalist planner with a specialist GUI executor and trains application-specific planners through decoupled GRPO. Successful specialist trajectories then support supervised training of a consolidated planner, evaluated across four ScienceBoard applications.\nPlatforms: desktop\nTopics: reinforcement-learning, planning, science, imitation-learning\nPublished: 2025-08-27\nAuthors: Zeyi Sun, Yuhang Cao, Jianze Liang, Qiushi Sun, Ziyu Liu, Zhixiong Zhang, Yuhang Zang, Xiaoyi Dong, Kai Chen, Dahua Lin, Jiaqi Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.20096v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-20096/"}
{"id": "arxiv:2508.21456", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc6411807fd58daed4824e86c39377b24d9fd022ac19daceb836bfb382353cf", "submitted_at": "2025-08-29T09:39:00Z", "updated_at": "2025-08-29T09:39:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21456v1"}, "authors": ["Yi-Hao Peng", "Dingzeyu Li", "Jeffrey P. Bigham", "Amy Pavel"], "canonical_source_url": "https://arxiv.org/abs/2508.21456", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-21456/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc6411807fd58daed4824e86c39377b24d9fd022ac19daceb836bfb382353cf", "submitted_at": "2025-08-29T09:39:00Z", "updated_at": "2025-08-29T09:39:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21456v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:40Z", "role": "primary", "sha256": "713ecfafe72365a57604306ccaba68f94158fc9b1aa169b408945922cdd1418c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2508.21456v1"}], "identifiers": {"arxiv": "2508.21456"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.21456"}], "name": "Morae: Proactively Pausing UI Agents for User Choices", "publication": null, "published_at": "2025-08-29", "record_sha256": "bb8cc192f67a1625b4f025936a4b13aaafed54284dd69264fc6675eb8c99f8d9", "relations": [], "resource_id": "arxiv:2508.21456", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.21456v1", "topics": ["human-in-the-loop", "web-navigation", "evaluation"], "updated_at": "2025-08-29"}, "text": "Morae: Proactively Pausing UI Agents for User Choices\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Morae, a UI agent that detects decision points and pauses for user choices using interface code and screenshots. A study with blind and low-vision participants evaluates this mixed-initiative approach on real web tasks.\nPlatforms: web\nTopics: human-in-the-loop, web-navigation, evaluation\nPublished: 2025-08-29\nAuthors: Yi-Hao Peng, Dingzeyu Li, Jeffrey P. Bigham, Amy Pavel\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.21456v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-21456/"}
{"id": "arxiv:2508.21767", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2a14819c7d7e8258947c9db8cc89012f4d133eaced4e4b904c3bfd1761181e5", "submitted_at": "2025-08-29T16:40:57Z", "updated_at": "2025-08-29T16:40:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21767v1"}, "authors": ["Zhixiong Zeng", "Jing Huang", "Liming Zheng", "Wenkang Han", "Yufeng Zhong", "Lei Chen", "Longrong Yang", "Yingjie Chu", "Yuzhi He", "Lin Ma"], "canonical_source_url": "https://arxiv.org/abs/2508.21767", "citation_url": "https://awesomecua.github.io/resources/arxiv-2508-21767/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b2a14819c7d7e8258947c9db8cc89012f4d133eaced4e4b904c3bfd1761181e5", "submitted_at": "2025-08-29T16:40:57Z", "updated_at": "2025-08-29T16:40:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2508.21767v1"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2508.21767"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2508.21767"}], "name": "UItron: Foundational GUI Agent with Advanced Perception and Planning", "publication": null, "published_at": "2025-08-29", "record_sha256": "855644c4afa032d1dc74ebbf95d990fbd6e3e7afa4e6b525fd90f76c02d12de0", "relations": [], "resource_id": "arxiv:2508.21767", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2508.21767v1", "topics": ["data-generation", "reinforcement-learning", "planning", "grounding"], "updated_at": "2025-08-29"}, "text": "UItron: Foundational GUI Agent with Advanced Perception and Planning\nResource type: paper\nReview status: curated\nScope: core\n\nUItron combines GUI perception and planning data engineering with supervised training and curriculum reinforcement learning in connected mobile and PC environments. It also collects more than one million operation steps across 100 Chinese apps and builds offline and online evaluation environments.\nPlatforms: desktop, cross-platform\nTopics: data-generation, reinforcement-learning, planning, grounding\nPublished: 2025-08-29\nAuthors: Zhixiong Zeng, Jing Huang, Liming Zheng, Wenkang Han, Yufeng Zhong, Lei Chen, Longrong Yang, Yingjie Chu, Yuzhi He, Lin Ma\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2508.21767v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2508-21767/"}
{"id": "arxiv:2509.00366", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c22f63bb5de22dd922ba37ea45df9404def017fee492e8ade03c83ab8b4f9b58", "submitted_at": "2025-08-30T05:32:32Z", "updated_at": "2025-08-30T05:32:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00366v1"}, "authors": ["Ziyi Guan", "Jason Chun Lok Li", "Zhijian Hou", "Pingping Zhang", "Donglai Xu", "Yuzhi Zhao", "Mengyang Wu", "Jinpeng Chen", "Thanh-Toan Nguyen", "Pengfei Xian", "Wenao Ma", "Shengchao Qin", "Graziano Chesi", "Ngai Wong"], "canonical_source_url": "https://arxiv.org/abs/2509.00366", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-00366/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c22f63bb5de22dd922ba37ea45df9404def017fee492e8ade03c83ab8b4f9b58", "submitted_at": "2025-08-30T05:32:32Z", "updated_at": "2025-08-30T05:32:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00366v1"}}, "environments": ["android", "web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.00366"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.00366"}], "name": "KG-RAG: Enhancing GUI Agent Decision-Making via Knowledge Graph-Driven Retrieval-Augmented Generation", "publication": null, "published_at": "2025-08-30", "record_sha256": "f27b78dd396aa808dfe540d1aef59a46e1cd0410c23e4cdc976ba81e96941e0b", "relations": [], "resource_id": "arxiv:2509.00366", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.00366v1", "topics": ["memory", "planning", "mobile-navigation", "evaluation"], "updated_at": "2025-08-30"}, "text": "KG-RAG: Enhancing GUI Agent Decision-Making via Knowledge Graph-Driven Retrieval-Augmented Generation\nResource type: paper\nReview status: curated\nScope: core\n\nKG-RAG converts UI transition graphs into a searchable vector store and uses intent-guided retrieval to construct actionable navigation paths. It introduces KG-Android-Bench and KG-Harmony-Bench and evaluates transfer from mobile apps to web and desktop tasks.\nPlatforms: android, web, desktop, cross-platform\nTopics: memory, planning, mobile-navigation, evaluation\nPublished: 2025-08-30\nAuthors: Ziyi Guan, Jason Chun Lok Li, Zhijian Hou, Pingping Zhang, Donglai Xu, Yuzhi Zhao, Mengyang Wu, Jinpeng Chen, Thanh-Toan Nguyen, Pengfei Xian, Wenao Ma, Shengchao Qin, Graziano Chesi, Ngai Wong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.00366v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-00366/"}
{"id": "arxiv:2509.00531", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbf26f28382245db085fd48f15e905bfa31f1c289a9a7f0799d35440069b442a", "submitted_at": "2025-08-30T15:24:47Z", "updated_at": "2025-08-30T15:24:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00531v1"}, "authors": ["Cheng Zhang", "Erhu Feng", "Xi Zhao", "Yisheng Zhao", "Wangbo Gong", "Jiahui Sun", "Dong Du", "Zhichao Hua", "Yubin Xia", "Haibo Chen"], "canonical_source_url": "https://arxiv.org/abs/2509.00531", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-00531/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbf26f28382245db085fd48f15e905bfa31f1c289a9a7f0799d35440069b442a", "submitted_at": "2025-08-30T15:24:47Z", "updated_at": "2025-08-30T15:24:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.00531v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:43Z", "role": "primary", "sha256": "b61937e1f046c4158b92c2f9d34dddea52a757f0e49b79f30efd73db24484abe", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.00531v1"}], "identifiers": {"arxiv": "2509.00531"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.00531"}], "name": "MobiAgent: A Systematic Framework for Customizable Mobile Agents", "publication": null, "published_at": "2025-08-30", "record_sha256": "c97f7da4c0a3463fe824e278f09761f4fc276515fd33f56aa332f2fcb20030ba", "relations": [], "resource_id": "arxiv:2509.00531", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.00531v1", "topics": ["mobile-navigation", "efficiency", "data-generation"], "updated_at": "2025-08-30"}, "text": "MobiAgent: A Systematic Framework for Customizable Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents MobiAgent, combining MobiMind agent models, the AgentRR acceleration framework, and MobiFlow evaluation. It also introduces an AI-assisted data-collection pipeline for mobile GUI tasks.\nPlatforms: unspecified\nTopics: mobile-navigation, efficiency, data-generation\nPublished: 2025-08-30\nAuthors: Cheng Zhang, Erhu Feng, Xi Zhao, Yisheng Zhao, Wangbo Gong, Jiahui Sun, Dong Du, Zhichao Hua, Yubin Xia, Haibo Chen\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.00531v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-00531/"}
{"id": "arxiv:2509.01052", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bf86652aaeb252972f3010be38b75b78dfaa9a8063c480c80d0cea47397c2e7", "submitted_at": "2025-09-01T01:33:16Z", "updated_at": "2025-10-15T10:33:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01052v2"}, "authors": ["Jaewoo Ahn", "Junseo Kim", "Heeseung Yun", "Jaehyeon Son", "Dongmin Park", "Jaewoong Cho", "Gunhee Kim"], "canonical_source_url": "https://arxiv.org/abs/2509.01052", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-01052/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bf86652aaeb252972f3010be38b75b78dfaa9a8063c480c80d0cea47397c2e7", "submitted_at": "2025-09-01T01:33:16Z", "updated_at": "2025-10-15T10:33:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01052v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.01052"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01052"}], "name": "FlashAdventure: A Benchmark for GUI Agents Solving Full Story Arcs in Diverse Adventure Games", "publication": null, "published_at": "2025-09-01", "record_sha256": "e76f4445da910a7d5e41cd270656c3510503f213f728b6664d8508ed94f0334b", "relations": [], "resource_id": "arxiv:2509.01052", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2509.01052v2", "topics": ["evaluation", "long-horizon", "memory", "planning"], "updated_at": "2025-10-15"}, "text": "FlashAdventure: A Benchmark for GUI Agents Solving Full Story Arcs in Diverse Adventure Games\nResource type: paper\nReview status: curated\nScope: adjacent\n\nFlashAdventure evaluates GUI agents on completing story arcs in 34 adventure games. It pairs an automated gameplay judge with COAST, a framework that keeps long-term clue memory to plan and complete sequential objectives.\nPlatforms: unspecified\nTopics: evaluation, long-horizon, memory, planning\nPublished: 2025-09-01\nAuthors: Jaewoo Ahn, Junseo Kim, Heeseung Yun, Jaehyeon Son, Dongmin Park, Jaewoong Cho, Gunhee Kim\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.01052v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-01052/"}
{"id": "arxiv:2509.01619", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f6ff68c59cede9cf59c039416a87ea39b926e3a3179b9f81af9354767befb93", "submitted_at": "2025-09-01T16:56:16Z", "updated_at": "2025-09-01T16:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.01619v1"}, "authors": ["Abhinav Kumar", "Jaechul Roh", "Ali Naseh", "Amir Houmansadr", "Eugene Bagdasarian"], "canonical_source_url": "https://arxiv.org/abs/2509.01619", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-01619/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f6ff68c59cede9cf59c039416a87ea39b926e3a3179b9f81af9354767befb93", "submitted_at": "2025-09-01T16:56:16Z", "updated_at": "2025-09-01T16:56:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.01619v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.01619"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01619"}], "name": "Throttling Web Agents Using Reasoning Gates", "publication": null, "published_at": "2025-09-01", "record_sha256": "def1b4457e22bd74123f8a1b00702f872196dd02f7ff29992d2f5f4f82711592", "relations": [], "resource_id": "arxiv:2509.01619", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2509.01619v1", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2025-09-01"}, "text": "Throttling Web Agents Using Reasoning Gates\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces reasoning gates that require agents to solve generated, verifiable puzzles before receiving access to resources. The throttling approach is deployed on a custom website and MCP servers and evaluated with web agents.\nPlatforms: web\nTopics: safety-security, evaluation, tool-integration\nPublished: 2025-09-01\nAuthors: Abhinav Kumar, Jaechul Roh, Ali Naseh, Amir Houmansadr, Eugene Bagdasarian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The contribution is an access-cost mechanism spanning websites and MCP servers, not a GUI navigation policy.\n\nPrimary source: https://arxiv.org/abs/2509.01619v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-01619/"}
{"id": "arxiv:2509.01720", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44c49412776ec9837a08d3d8347214495ffbda14a391c7fa67ad780a0eac9577", "submitted_at": "2025-09-01T18:55:27Z", "updated_at": "2025-11-13T01:26:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01720v2"}, "authors": ["Georgios Papoudakis", "Thomas Coste", "Jianye Hao", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2509.01720", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-01720/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "44c49412776ec9837a08d3d8347214495ffbda14a391c7fa67ad780a0eac9577", "submitted_at": "2025-09-01T18:55:27Z", "updated_at": "2025-11-13T01:26:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.01720v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2509.01720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.01720"}], "name": "Succeed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App Control", "publication": null, "published_at": "2025-09-01", "record_sha256": "8914f1f0c5b3d74813392414de05d3ab83053159d1e05eef5293bbc2fadf0124", "relations": [], "resource_id": "arxiv:2509.01720", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.01720v2", "topics": ["reinforcement-learning", "mobile-navigation", "efficiency"], "updated_at": "2025-11-13"}, "text": "Succeed or Learn Slowly: Sample Efficient Off-Policy Reinforcement Learning for Mobile App Control\nResource type: paper\nReview status: curated\nScope: core\n\nSucceed or Learn Slowly applies direct off-policy updates to successful interactions and more conservative regularized updates to negative examples. Successful Transition Replay prioritizes successful experience during policy learning, with evaluation on AndroidWorld.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation, efficiency\nPublished: 2025-09-01\nAuthors: Georgios Papoudakis, Thomas Coste, Jianye Hao, Jun Wang, Kun Shao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.01720v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-01720/"}
{"id": "arxiv:2509.02322", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "227630b3a257557262ed25eb21906a129eb35c6c5d2de7c5754f919dd281c4bd", "submitted_at": "2025-09-02T13:47:54Z", "updated_at": "2025-09-02T13:47:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.02322v1"}, "authors": ["Longrong Yang", "Zhixiong Zeng", "Yufeng Zhong", "Jing Huang", "Liming Zheng", "Lei Chen", "Haibo Qiu", "Zequn Qin", "Lin Ma", "Xi Li"], "canonical_source_url": "https://arxiv.org/abs/2509.02322", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-02322/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "227630b3a257557262ed25eb21906a129eb35c6c5d2de7c5754f919dd281c4bd", "submitted_at": "2025-09-02T13:47:54Z", "updated_at": "2025-09-02T13:47:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.02322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2509.02322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02322"}], "name": "OmniActor: A Generalist GUI and Embodied Agent for 2D&3D Worlds", "publication": null, "published_at": "2025-09-02", "record_sha256": "eccbe36a2c91c146c5e92ad308e48f5f106c7fc547e4d8973f1421894de41f29", "relations": [], "resource_id": "arxiv:2509.02322", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.02322v1", "topics": ["action-prediction", "data-generation", "imitation-learning"], "updated_at": "2025-09-02"}, "text": "OmniActor: A Generalist GUI and Embodied Agent for 2D&3D Worlds\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OmniActor, which shares shallow model layers while separating deeper parameters to train jointly on GUI and embodied interactions. A unified action representation and mixed training data support evaluation in both task families.\nPlatforms: unspecified\nTopics: action-prediction, data-generation, imitation-learning\nPublished: 2025-09-02\nAuthors: Longrong Yang, Zhixiong Zeng, Yufeng Zhong, Jing Huang, Liming Zheng, Lei Chen, Haibo Qiu, Zequn Qin, Lin Ma, Xi Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.02322v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-02322/"}
{"id": "arxiv:2509.02444", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49a887b12b3395821aa100a66e61aac808309ef947ae752475df3d2db9a4fcd8", "submitted_at": "2025-09-02T15:48:21Z", "updated_at": "2025-10-17T00:57:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02444v2"}, "authors": ["Jingru Fan", "Yufan Dang", "Jingyao Wu", "Huatao Li", "Runde Yang", "Xiyuan Yang", "Yuheng Wang", "Chen Qian"], "canonical_source_url": "https://arxiv.org/abs/2509.02444", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-02444/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49a887b12b3395821aa100a66e61aac808309ef947ae752475df3d2db9a4fcd8", "submitted_at": "2025-09-02T15:48:21Z", "updated_at": "2025-10-17T00:57:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02444v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:33.619660+00:00", "role": "primary", "sha256": "6c2591dad8893f87cffdea211584a3733dc2137dafa9416c1885d1b0126fd540", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2509.02444v2"}], "identifiers": {"arxiv": "2509.02444"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02444"}], "name": "AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent", "publication": null, "published_at": "2025-09-02", "record_sha256": "cdafaaca033edeae2ef6c04d1c444e466b3794b41ca8060c7365bb3c1cfa13a9", "relations": [], "resource_id": "arxiv:2509.02444", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.02444v2", "topics": ["planning", "grounding", "workflow", "mobile-navigation", "long-horizon"], "updated_at": "2025-10-17"}, "text": "AppCopilot: Toward General, Accurate, Long-Horizon, and Efficient Mobile Agent\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a mobile-agent stack combining region-based coordinate correction, hierarchical task planning, multi-agent voting, and reuse of successful interaction traces.\nPlatforms: unspecified\nTopics: planning, grounding, workflow, mobile-navigation, long-horizon\nPublished: 2025-09-02\nAuthors: Jingru Fan, Yufan Dang, Jingyao Wu, Huatao Li, Runde Yang, Xiyuan Yang, Yuheng Wang, Chen Qian\nRecorded arXiv version: v2\n\nTrace reuse assumes stable layouts and fixed tasks. The described function-call routes are user-selected, and coordinate calibration depends on valid detected widgets. No particular operating-system coverage, latency guarantee or reproduced result is asserted. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (4.2 OCR/OR Region Localization; 4.2.2 Coordinate Calibration; 4.3 Multi-Agent Voting; 5.2 Task Decomposition; 6.2 Experience Reuse; 6.3 Function Calling); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2509.02444v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-02444/"}
{"id": "arxiv:2509.02544", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "800e417f4a2f720beb436ddd2e7bc60f0dd2483b44ac007a72a5eab2c6001ff8", "submitted_at": "2025-09-02T17:44:45Z", "updated_at": "2025-09-05T14:59:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02544v2"}, "authors": ["Haoming Wang", "Haoyang Zou", "Huatong Song", "Jiazhan Feng", "Junjie Fang", "Junting Lu", "Longxiang Liu", "Qinyu Luo", "Shihao Liang", "Shijue Huang", "Wanjun Zhong", "Yining Ye", "Yujia Qin", "Yuwen Xiong", "Yuxin Song", "Zhiyong Wu", "Aoyan Li", "Bo Li", "Chen Dun", "Chong Liu", "Daoguang Zan", "Fuxing Leng", "Hanbin Wang", "Hao Yu", "Haobin Chen", "Hongyi Guo", "Jing Su", "Jingjia Huang", "Kai Shen", "Kaiyu Shi", "Lin Yan", "Peiyao Zhao", "Pengfei Liu", "Qinghao Ye", "Renjie Zheng", "Shulin Xin", "Wayne Xin Zhao", "Wen Heng", "Wenhao Huang", "Wenqian Wang", "Xiaobo Qin", "Yi Lin", "Youbin Wu", "Zehui Chen", "Zihao Wang", "Baoquan Zhong", "Xinchun Zhang", "Xujing Li", "Yuanfan Li", "Zhongkai Zhao", "Chengquan Jiang", "Faming Wu", "Haotian Zhou", "Jinlin Pang", "Li Han", "Qi Liu", "Qianli Ma", "Siyao Liu", "Songhua Cai", "Wenqi Fu", "Xin Liu", "Yaohui Wang", "Zhi Zhang", "Bo Zhou", "Guoliang Li", "Jiajun Shi", "Jiale Yang", "Jie Tang", "Li Li", "Qihua Han", "Taoran Lu", "Woyu Lin", "Xiaokang Tong", "Xinyao Li", "Yichi Zhang", "Yu Miao", "Zhengxuan Jiang", "Zili Li", "Ziyuan Zhao", "Chenxin Li", "Dehua Ma", "Feng Lin", "Ge Zhang", "Haihua Yang", "Hangyu Guo", "Hongda Zhu", "Jiaheng Liu", "Junda Du", "Kai Cai", "Kuanye Li", "Lichen Yuan", "Meilan Han", "Minchao Wang", "Shuyue Guo", "Tianhao Cheng", "Xiaobo Ma", "Xiaojun Xiao", "Xiaolong Huang", "Xinjie Chen", "Yidi Du", "Yilin Chen", "Yiwen Wang", "Zhaojian Li", "Zhenzhu Yang", "Zhiyuan Zeng", "Chaolin Jin", "Chen Li", "Hao Chen", "Haoli Chen", "Jian Chen", "Qinghao Zhao", "Guang Shi"], "canonical_source_url": "https://arxiv.org/abs/2509.02544", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-02544/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "800e417f4a2f720beb436ddd2e7bc60f0dd2483b44ac007a72a5eab2c6001ff8", "submitted_at": "2025-09-02T17:44:45Z", "updated_at": "2025-09-05T14:59:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.02544v2"}}, "environments": ["web", "desktop", "windows", "android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.02544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.02544"}], "name": "UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning", "publication": null, "published_at": "2025-09-02", "record_sha256": "6cca54f136affaa4bca32b5edf8d6d591ac9eb99fc9688c848b2b5d4c7f688f7", "relations": [], "resource_id": "arxiv:2509.02544", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.02544v2", "topics": ["reinforcement-learning", "data-generation", "sandbox", "tool-integration"], "updated_at": "2025-09-05"}, "text": "UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUI-TARS-2 combines scalable trajectory generation, stabilized multi-turn reinforcement learning, and a sandbox that supports GUI, filesystem, and terminal actions. The report evaluates the resulting agent across web, desktop, Android, and game tasks and analyzes training stability.\nPlatforms: web, desktop, windows, android, cross-platform\nTopics: reinforcement-learning, data-generation, sandbox, tool-integration\nPublished: 2025-09-02\nAuthors: Haoming Wang, Haoyang Zou, Huatong Song, Jiazhan Feng, Junjie Fang, Junting Lu, Longxiang Liu, Qinyu Luo, Shihao Liang, Shijue Huang, Wanjun Zhong, Yining Ye, Yujia Qin, Yuwen Xiong, Yuxin Song, Zhiyong Wu, Aoyan Li, Bo Li, Chen Dun, Chong Liu, Daoguang Zan, Fuxing Leng, Hanbin Wang, Hao Yu, Haobin Chen, Hongyi Guo, Jing Su, Jingjia Huang, Kai Shen, Kaiyu Shi, Lin Yan, Peiyao Zhao, Pengfei Liu, Qinghao Ye, Renjie Zheng, Shulin Xin, Wayne Xin Zhao, Wen Heng, Wenhao Huang, Wenqian Wang, Xiaobo Qin, Yi Lin, Youbin Wu, Zehui Chen, Zihao Wang, Baoquan Zhong, Xinchun Zhang, Xujing Li, Yuanfan Li, Zhongkai Zhao, Chengquan Jiang, Faming Wu, Haotian Zhou, Jinlin Pang, Li Han, Qi Liu, Qianli Ma, Siyao Liu, Songhua Cai, Wenqi Fu, Xin Liu, Yaohui Wang, Zhi Zhang, Bo Zhou, Guoliang Li, Jiajun Shi, Jiale Yang, Jie Tang, Li Li, Qihua Han, Taoran Lu, Woyu Lin, Xiaokang Tong, Xinyao Li, Yichi Zhang, Yu Miao, Zhengxuan Jiang, Zili Li, Ziyuan Zhao, Chenxin Li, Dehua Ma, Feng Lin, Ge Zhang, Haihua Yang, Hangyu Guo, Hongda Zhu, Jiaheng Liu, Junda Du, Kai Cai, Kuanye Li, Lichen Yuan, Meilan Han, Minchao Wang, Shuyue Guo, Tianhao Cheng, Xiaobo Ma, Xiaojun Xiao, Xiaolong Huang, Xinjie Chen, Yidi Du, Yilin Chen, Yiwen Wang, Zhaojian Li, Zhenzhu Yang, Zhiyuan Zeng, Chaolin Jin, Chen Li, Hao Chen, Haoli Chen, Jian Chen, Qinghao Zhao, Guang Shi\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.02544v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-02544/"}
{"id": "arxiv:2509.03891", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8b2db00f62afe350e55e2266b3627de9b0cf089178b0cf80eb13b06275eee15", "submitted_at": "2025-09-04T05:22:42Z", "updated_at": "2025-09-04T05:22:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.03891v1"}, "authors": ["Gowen Loo", "Chang Liu", "Qinghong Yin", "Xiang Chen", "Jiawei Chen", "Jingyuan Zhang", "Yu Tian"], "canonical_source_url": "https://arxiv.org/abs/2509.03891", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-03891/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8b2db00f62afe350e55e2266b3627de9b0cf089178b0cf80eb13b06275eee15", "submitted_at": "2025-09-04T05:22:42Z", "updated_at": "2025-09-04T05:22:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.03891v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2509.03891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.03891"}], "name": "MobileRAG: Enhancing Mobile Agent with Retrieval-Augmented Generation", "publication": null, "published_at": "2025-09-04", "record_sha256": "1fe82b01a5c731376b1dc478a570a324d1b679a0ff87500cf37c4a3066152a75", "relations": [], "resource_id": "arxiv:2509.03891", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.03891v1", "topics": ["memory", "mobile-navigation", "evaluation", "long-horizon"], "updated_at": "2025-09-04"}, "text": "MobileRAG: Enhancing Mobile Agent with Retrieval-Augmented Generation\nResource type: paper\nReview status: curated\nScope: core\n\nMobileRAG combines retrieval and memory components to support mobile tasks that need external knowledge or long action sequences. MobileRAG-Eval is introduced to study these requirements on complex real-world app tasks.\nPlatforms: unspecified\nTopics: memory, mobile-navigation, evaluation, long-horizon\nPublished: 2025-09-04\nAuthors: Gowen Loo, Chang Liu, Qinghong Yin, Xiang Chen, Jiawei Chen, Jingyuan Zhang, Yu Tian\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.03891v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-03891/"}
{"id": "arxiv:2509.04243", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "875e2452985bae9e67d7071121ab13f5205307ecc30bf33f237048a44e75b3bc", "submitted_at": "2025-09-04T14:17:01Z", "updated_at": "2025-09-04T14:17:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04243v1"}, "authors": ["Wanfu Wang", "Qipeng Huang", "Guangquan Xue", "Xiaobo Liang", "Juntao Li"], "canonical_source_url": "https://arxiv.org/abs/2509.04243", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-04243/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "875e2452985bae9e67d7071121ab13f5205307ecc30bf33f237048a44e75b3bc", "submitted_at": "2025-09-04T14:17:01Z", "updated_at": "2025-09-04T14:17:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04243v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.04243"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04243"}], "name": "Learning Active Perception via Self-Evolving Preference Optimization for GUI Grounding", "publication": null, "published_at": "2025-09-04", "record_sha256": "9676bd46729af566de0e72d482d4e594b9c8e35e16686ee35f8a61d57a622d96", "relations": [], "resource_id": "arxiv:2509.04243", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.04243v1", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": "2025-09-04"}, "text": "Learning Active Perception via Self-Evolving Preference Optimization for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nLASER builds preference data for multi-step GUI perception by combining Monte Carlo quality estimates with region-overlap evaluation. Its self-evolving training encourages relevant region selection and adapts the amount of reasoning before predicting screen coordinates.\nPlatforms: unspecified\nTopics: grounding, data-generation, screen-understanding\nPublished: 2025-09-04\nAuthors: Wanfu Wang, Qipeng Huang, Guangquan Xue, Xiaobo Liang, Juntao Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.04243v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-04243/"}
{"id": "arxiv:2509.04908", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f74081d31c98d18e10a5febc5de436a3788fc4156215e4be8fc54b1c53260f58", "submitted_at": "2025-09-05T08:24:12Z", "updated_at": "2025-09-05T08:24:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04908v1"}, "authors": ["Hongyi Jing", "Jiafu Chen", "Chen Rao", "Ziqiang Dang", "Jiajie Teng", "Tianyi Chu", "Juncheng Mo", "Shuo Fang", "Huaizhong Lin", "Rui Lv", "Chenguang Ma", "Lei Zhao"], "canonical_source_url": "https://arxiv.org/abs/2509.04908", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-04908/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f74081d31c98d18e10a5febc5de436a3788fc4156215e4be8fc54b1c53260f58", "submitted_at": "2025-09-05T08:24:12Z", "updated_at": "2025-09-05T08:24:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.04908v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:46Z", "role": "primary", "sha256": "e4304967a9a63ead71f00e1933965321540df913f534c0711cc18b9190560cde", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.04908v1"}], "identifiers": {"arxiv": "2509.04908"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.04908"}], "name": "SparkUI-Parser: Enhancing GUI Perception with Robust Grounding and Parsing", "publication": null, "published_at": "2025-09-05", "record_sha256": "3e59c5f6bc61a1dcf9a3932eb09ebeaca569bc43428f40741957a438057e2dc6", "relations": [], "resource_id": "arxiv:2509.04908", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.04908v1", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2025-09-05"}, "text": "SparkUI-Parser: Enhancing GUI Perception with Robust Grounding and Parsing\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SparkUI-Parser, which adds continuous coordinate decoding and a rejection mechanism for nonexistent targets to a multimodal model. It also presents ScreenParse for evaluating structural interface perception.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, evaluation\nPublished: 2025-09-05\nAuthors: Hongyi Jing, Jiafu Chen, Chen Rao, Ziqiang Dang, Jiajie Teng, Tianyi Chu, Juncheng Mo, Shuo Fang, Huaizhong Lin, Rui Lv, Chenguang Ma, Lei Zhao\nRecorded arXiv version: v1\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.04908v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-04908/"}
{"id": "arxiv:2509.06477", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b1e0e807f5f396bfd9108ab6cfacd48ed4780abfb792589da0b41d7d246c717", "submitted_at": "2025-09-08T09:43:48Z", "updated_at": "2026-04-15T11:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.06477v2"}, "authors": ["Pengxiang Zhao", "Guangyi Liu", "YaoZhen Liang", "Weiqing He", "Zhengxi Lu", "WenHao Wang", "Yuehao Huang", "Yuxiang Chai", "Zhaolu Kang", "Yaxuan Guo", "Hao Wang", "Kexin Zhang", "Liang Liu", "Yong Liu"], "canonical_source_url": "https://arxiv.org/abs/2509.06477", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-06477/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b1e0e807f5f396bfd9108ab6cfacd48ed4780abfb792589da0b41d7d246c717", "submitted_at": "2025-09-08T09:43:48Z", "updated_at": "2026-04-15T11:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.06477v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.06477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.06477"}], "name": "MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents", "publication": null, "published_at": "2025-09-08", "record_sha256": "ea45db7d0f759027fd97e3369fc198910723e3b09e389f0f11c6bed7b8b1c15c", "relations": [], "resource_id": "arxiv:2509.06477", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.06477v2", "topics": ["evaluation", "mobile-navigation", "tool-integration", "workflow"], "updated_at": "2026-04-15"}, "text": "MAS-Bench: A Unified Benchmark for Shortcut-Augmented Hybrid Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMAS-Bench evaluates mobile agents that combine GUI actions with APIs, deep links, and reusable scripts. Its 139 tasks across 11 apps test use of 88 predefined shortcuts as well as the ability to generate new reusable workflows.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, tool-integration, workflow\nPublished: 2025-09-08\nAuthors: Pengxiang Zhao, Guangyi Liu, YaoZhen Liang, Weiqing He, Zhengxi Lu, WenHao Wang, Yuehao Huang, Yuxiang Chai, Zhaolu Kang, Yaxuan Guo, Hao Wang, Kexin Zhang, Liang Liu, Yong Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.06477v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-06477/"}
{"id": "arxiv:2509.07553", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bd6b323a376d7606b8c5ec0615b992f47a388cbd3604191241f6a0a9b2cce6e", "submitted_at": "2025-09-09T09:46:01Z", "updated_at": "2026-04-03T02:36:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.07553v3"}, "authors": ["Zheng Wu", "Heyuan Huang", "Xingyu Lou", "Xiangmou Qu", "Pengzhou Cheng", "Zongru Wu", "Weiwen Liu", "Weinan Zhang", "Jun Wang", "Zhaoxiang Wang", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2509.07553", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-07553/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3bd6b323a376d7606b8c5ec0615b992f47a388cbd3604191241f6a0a9b2cce6e", "submitted_at": "2025-09-09T09:46:01Z", "updated_at": "2026-04-03T02:36:28Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.07553v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:50Z", "role": "primary", "sha256": "776192081826494b05077d71757f3006575ed50a8cc244004e2de7053104d7eb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.07553v3"}], "identifiers": {"arxiv": "2509.07553"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.07553"}], "name": "VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents", "publication": null, "published_at": "2025-09-09", "record_sha256": "5323e55bc00dfbc3a13cc075e975b8256b837066f3ddc67bf91515d5e4f3fe9c", "relations": [], "resource_id": "arxiv:2509.07553", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.07553v3", "topics": ["human-in-the-loop", "safety-security", "reinforcement-learning"], "updated_at": "2026-04-03"}, "text": "VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VeriOS-Agent, which learns when to execute GUI actions autonomously and when to query a user under untrustworthy conditions. Its staged training combines supervised learning and group-relative policy optimization.\nPlatforms: unspecified\nTopics: human-in-the-loop, safety-security, reinforcement-learning\nPublished: 2025-09-09\nAuthors: Zheng Wu, Heyuan Huang, Xingyu Lou, Xiangmou Qu, Pengzhou Cheng, Zongru Wu, Weiwen Liu, Weinan Zhang, Jun Wang, Zhaoxiang Wang, Zhuosheng Zhang\nRecorded arXiv version: v3\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.07553v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-07553/"}
{"id": "arxiv:2509.07764", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae582aceb31075c460a9e4a6dce9978574c3e0ada731b26c13738a4a4fafcc32", "submitted_at": "2025-09-09T13:59:00Z", "updated_at": "2025-09-09T13:59:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.07764v1"}, "authors": ["Haitao Hu", "Peng Chen", "Yanpeng Zhao", "Yuqi Chen"], "canonical_source_url": "https://arxiv.org/abs/2509.07764", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-07764/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae582aceb31075c460a9e4a6dce9978574c3e0ada731b26c13738a4a4fafcc32", "submitted_at": "2025-09-09T13:59:00Z", "updated_at": "2025-09-09T13:59:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.07764v1"}}, "environments": ["desktop", "linux", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:25:42Z", "role": "primary", "sha256": "7f303629ef6d9c7df128c6d12d472abd9bf238e0204515f18839249adef317b9", "supports": ["summary", "environments", "scope", "notes"], "url": "https://arxiv.org/html/2509.07764v1"}], "identifiers": {"arxiv": "2509.07764"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.07764"}], "name": "AgentSentinel: An End-to-End and Real-Time Security Defense Framework for Computer-Use Agents", "publication": null, "published_at": "2025-09-09", "record_sha256": "b1d3e890ff7298930ea19b120325346c409d854ad78f7f38dcddd9f8246a97b5", "relations": [], "resource_id": "arxiv:2509.07764", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.07764v1", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2025-09-09"}, "text": "AgentSentinel: An End-to-End and Real-Time Security Defense Framework for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AgentSentinel, which intercepts sensitive computer-agent operations and audits them using task context together with system traces. BadComputerUse evaluates this enforcement approach across attack scenarios involving tool and computer control.\nPlatforms: desktop, linux, web\nTopics: safety-security, observability, evaluation\nPublished: 2025-09-09\nAuthors: Haitao Hu, Peng Chen, Yanpeng Zhao, Yuqi Chen\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The official full paper confirms GUI/desktop interaction alongside Bash and file tools. Its abstract and body enumerate attack categories differently; no category count is promoted into this annotation.\n\nPrimary source: https://arxiv.org/abs/2509.07764v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-07764/"}
{"id": "arxiv:2509.10723", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247ed053a4c8f62106d9f8e9af0801d94265ed712f4ae9f095fd8aaccbe33c59", "submitted_at": "2025-09-12T22:26:31Z", "updated_at": "2025-09-12T22:26:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.10723v1"}, "authors": ["Jingyu Tang", "Chaoran Chen", "Jiawen Li", "Zhiping Zhang", "Bingcan Guo", "Ibrahim Khalilov", "Simret Araya Gebreegziabher", "Bingsheng Yao", "Dakuo Wang", "Yanfang Ye", "Tianshi Li", "Ziang Xiao", "Yaxing Yao", "Toby Jia-Jun Li"], "canonical_source_url": "https://arxiv.org/abs/2509.10723", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-10723/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247ed053a4c8f62106d9f8e9af0801d94265ed712f4ae9f095fd8aaccbe33c59", "submitted_at": "2025-09-12T22:26:31Z", "updated_at": "2025-09-12T22:26:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.10723v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.10723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.10723"}], "name": "Dark Patterns Meet GUI Agents: LLM Agent Susceptibility to Manipulative Interfaces and the Role of Human Oversight", "publication": null, "published_at": "2025-09-12", "record_sha256": "75d51476bda734bff51538e26a70de72bf15632c33c637dc0db7c788d2f77776", "relations": [], "resource_id": "arxiv:2509.10723", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.10723v1", "topics": ["human-in-the-loop", "safety-security", "evaluation"], "updated_at": "2025-09-12"}, "text": "Dark Patterns Meet GUI Agents: LLM Agent Susceptibility to Manipulative Interfaces and the Role of Human Oversight\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how GUI agents, people, and human-agent teams respond to manipulative interface designs. The two-phase evaluation examines failure modes and the benefits and costs of human oversight.\nPlatforms: unspecified\nTopics: human-in-the-loop, safety-security, evaluation\nPublished: 2025-09-12\nAuthors: Jingyu Tang, Chaoran Chen, Jiawen Li, Zhiping Zhang, Bingcan Guo, Ibrahim Khalilov, Simret Araya Gebreegziabher, Bingsheng Yao, Dakuo Wang, Yanfang Ye, Tianshi Li, Ziang Xiao, Yaxing Yao, Toby Jia-Jun Li\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.10723v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-10723/"}
{"id": "arxiv:2509.11250", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ab96edd8e52b3a490455b46878c92c23d15c0f9e105e71dbd6848e058da42ec", "submitted_at": "2025-09-14T12:47:54Z", "updated_at": "2026-08-24T12:04:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.11250v3"}, "authors": ["Yitong Zhang", "Ximo Li", "Liyi Cai", "Jia Li"], "canonical_source_url": "https://arxiv.org/abs/2509.11250", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-11250/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ab96edd8e52b3a490455b46878c92c23d15c0f9e105e71dbd6848e058da42ec", "submitted_at": "2025-09-14T12:47:54Z", "updated_at": "2026-08-24T12:04:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.11250v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.11250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11250"}], "name": "Environmental Injection Attacks against GUI Agents in Realistic Dynamic Environments", "publication": null, "published_at": "2025-09-14", "record_sha256": "3816ec7bfe2fb97aae688cbb8ac8701b78f5d4a99805b3167210480c39a64794", "relations": [], "resource_id": "arxiv:2509.11250", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.11250v3", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-08-24"}, "text": "Environmental Injection Attacks against GUI Agents in Realistic Dynamic Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Chameleon, an environmental-injection evaluation framework that varies webpage content and trigger context. It combines simulated page variation with attention-based optimization and tests GUI agents across realistic websites.\nPlatforms: web\nTopics: safety-security, robustness, evaluation\nPublished: 2025-09-14\nAuthors: Yitong Zhang, Ximo Li, Liyi Cai, Jia Li\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.11250v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-11250/"}
{"id": "arxiv:2509.11543", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e31638e531aff49fe59bef6e80b35069b1f851ea0cbf1901130557148df4783", "submitted_at": "2025-09-15T03:24:08Z", "updated_at": "2025-09-24T15:05:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11543v2"}, "authors": ["Zhengxi Lu", "Jiabo Ye", "Fei Tang", "Yongliang Shen", "Haiyang Xu", "Ziwei Zheng", "Weiming Lu", "Ming Yan", "Fei Huang", "Jun Xiao", "Yueting Zhuang"], "canonical_source_url": "https://arxiv.org/abs/2509.11543", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-11543/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e31638e531aff49fe59bef6e80b35069b1f851ea0cbf1901130557148df4783", "submitted_at": "2025-09-15T03:24:08Z", "updated_at": "2025-09-24T15:05:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11543v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:49Z", "role": "primary", "sha256": "c208e36ae75cef4b32cfcb2f8182dd4aaaa218e3119bb3d4a2c20a5832f8b1ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2506.10387%2C2506.14866%2C2506.20332%2C2507.00008%2C2507.02592%2C2507.03730%2C2507.04227%2C2507.05445%2C2507.05673%2C2507.05720%2C2507.05791%2C2507.14293%2C2507.14799%2C2507.15846%2C2507.16853%2C2507.19132%2C2507.19478%2C2507.21071%2C2507.22025%2C2507.22358%2C2507.23779%2C2508.00935%2C2508.01330%2C2508.03700%2C2508.03923%2C2508.04010%2C2508.04026%2C2508.04482%2C2508.04700%2C2508.05615%2C2508.05731%2C2508.09057%2C2508.09123%2C2508.13167%2C2508.14040%2C2508.15144%2C2508.18689%2C2508.19679%2C2508.20096%2C2509.02544%2C2509.03891%2C2509.06477%2C2509.10723%2C2509.11250%2C2509.11543%2C2509.15221%2C2509.15566%2C2509.15738%2C2509.18119%2C2510.01354&max_results=50"}], "identifiers": {"arxiv": "2509.11543"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11543"}], "name": "UI-S1: Advancing GUI Automation via Semi-online Reinforcement Learning", "publication": null, "published_at": "2025-09-15", "record_sha256": "715e9b5aa9f34959f0d43ce021e3d32b3d4fcd3dd3b916352ca45a7b136454a3", "relations": [], "resource_id": "arxiv:2509.11543", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.11543v2", "topics": ["reinforcement-learning", "evaluation", "mobile-navigation"], "updated_at": "2025-09-24"}, "text": "UI-S1: Advancing GUI Automation via Semi-online Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUI-S1 simulates multi-turn online learning over offline trajectories, using a patch module to recover divergences while retaining the model’s own outputs in context. It combines discounted future returns with step- and episode-level advantages and introduces a semi-online performance proxy.\nPlatforms: android\nTopics: reinforcement-learning, evaluation, mobile-navigation\nPublished: 2025-09-15\nAuthors: Zhengxi Lu, Jiabo Ye, Fei Tang, Yongliang Shen, Haiyang Xu, Ziwei Zheng, Weiming Lu, Ming Yan, Fei Huang, Jun Xiao, Yueting Zhuang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.11543v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-11543/"}
{"id": "arxiv:2509.11548", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75ffc21abebbc2c362798f3455b9e6240b58407572e5633d5171e5612cfae71b", "submitted_at": "2025-09-15T03:28:29Z", "updated_at": "2026-06-10T01:13:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11548v2"}, "authors": ["Weiming Li", "Yan Shao", "Jing Yang", "Yujing Lu", "Ling Zhong", "Yuhan Wang", "Min Yu", "Tongxiao Ruan", "Manni Duan"], "canonical_source_url": "https://arxiv.org/abs/2509.11548", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-11548/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75ffc21abebbc2c362798f3455b9e6240b58407572e5633d5171e5612cfae71b", "submitted_at": "2025-09-15T03:28:29Z", "updated_at": "2026-06-10T01:13:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.11548v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.11548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11548"}], "name": "How Auxiliary Reasoning Unleashes GUI Grounding in VLMs", "publication": null, "published_at": "2025-09-15", "record_sha256": "2e6e1bc6897ad3aa4fdf0104f8b19b828e2f0ff3a453a7c69ca2e49eb247c5a9", "relations": [], "resource_id": "arxiv:2509.11548", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.11548v2", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-06-10"}, "text": "How Auxiliary Reasoning Unleashes GUI Grounding in VLMs\nResource type: paper\nReview status: curated\nScope: core\n\nStudies the gap between latent visual grounding and explicit coordinate output in general vision-language models. Three zero-shot methods add axes, grids, or labeled intersections to GUI screenshots so models can express spatial locations without fine-tuning.\nPlatforms: unspecified\nTopics: grounding, training-free, screen-understanding\nPublished: 2025-09-15\nAuthors: Weiming Li, Yan Shao, Jing Yang, Yujing Lu, Ling Zhong, Yuhan Wang, Min Yu, Tongxiao Ruan, Manni Duan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.11548v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-11548/"}
{"id": "arxiv:2509.11939", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f37ebc10535f96d17d28990e608a1f2eb6cc004f40d58b3288c4a81d781160ea", "submitted_at": "2025-09-15T13:58:52Z", "updated_at": "2025-09-15T13:58:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.11939v1"}, "authors": ["Shuning Zhang", "Yutong Jiang", "Rongjun Ma", "Yuting Yang", "Mingyao Xu", "Zhixin Huang", "Xin Yi", "Hewu Li"], "canonical_source_url": "https://arxiv.org/abs/2509.11939", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-11939/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f37ebc10535f96d17d28990e608a1f2eb6cc004f40d58b3288c4a81d781160ea", "submitted_at": "2025-09-15T13:58:52Z", "updated_at": "2025-09-15T13:58:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.11939v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.11939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.11939"}], "name": "PrivWeb: Unobtrusive and Content-aware Privacy Protection For Web Agents", "publication": null, "published_at": "2025-09-15", "record_sha256": "284f28ea28e44ee3c1a98f4466b9656db20831c328f07f3f5d4fd924e533272b", "relations": [], "resource_id": "arxiv:2509.11939", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.11939v1", "topics": ["safety-security", "human-in-the-loop", "web-navigation"], "updated_at": "2025-09-15"}, "text": "PrivWeb: Unobtrusive and Content-aware Privacy Protection For Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PrivWeb, which uses a local language model to anonymize interface information according to user preferences. Adaptive notifications and selective pauses give users control over sensitive information access during web-agent tasks.\nPlatforms: web\nTopics: safety-security, human-in-the-loop, web-navigation\nPublished: 2025-09-15\nAuthors: Shuning Zhang, Yutong Jiang, Rongjun Ma, Yuting Yang, Mingyao Xu, Zhixin Huang, Xin Yi, Hewu Li\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.11939v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-11939/"}
{"id": "arxiv:2509.13615", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be8620b021bd9218c1143282f1c7a7ea06be2abe6fc0928a310cfc8ac3172a18", "submitted_at": "2025-09-17T01:14:14Z", "updated_at": "2026-03-18T08:12:54Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2509.13615v5"}, "authors": ["Zongru Wu", "Rui Mao", "Zhiyuan Tian", "Pengzhou Cheng", "Tianjie Ju", "Zheng Wu", "Lingzhong Dong", "Haiyue Sheng", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2509.13615", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-13615/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "be8620b021bd9218c1143282f1c7a7ea06be2abe6fc0928a310cfc8ac3172a18", "submitted_at": "2025-09-17T01:14:14Z", "updated_at": "2026-03-18T08:12:54Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2509.13615v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:56Z", "role": "primary", "sha256": "3429cc2fc7737050cac964c976293fc4793e52c5913636c1f67d04243fbefb6a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.20410%2C2508.20890%2C2508.21456%2C2509.00088%2C2509.00531%2C2509.00969%2C2509.01684%2C2509.03536%2C2509.04018%2C2509.04908%2C2509.05368%2C2509.05831%2C2509.06917%2C2509.07098%2C2509.07553%2C2509.08820%2C2509.10345%2C2509.10540%2C2509.11253%2C2509.12543%2C2509.12987%2C2509.13615%2C2509.14547%2C2509.14744%2C2509.15061%2C2509.15237%2C2509.17336%2C2509.19185%2C2509.19199%2C2509.21072%2C2509.21291%2C2509.21766%2C2509.21799%2C2509.22715%2C2509.22830%2C2509.23738%2C2509.24704%2C2509.25455%2C2509.25693%2C2510.00172%2C2510.00515%2C2510.02190%2C2510.04491%2C2510.05096%2C2510.05156%2C2510.05788%2C2510.07284%2C2510.07915%2C2510.08787%2C2510.09517&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:53Z", "role": "primary", "sha256": "28426488ebca15b74a693ea9fcb2b0ce3165d3baf97b76a4884b34a0d1997af4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2509.13615v5"}], "identifiers": {"arxiv": "2509.13615"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.13615"}], "name": "See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles", "publication": null, "published_at": "2025-09-17", "record_sha256": "374f49932f2b36a5c4f420d72bed1f580298d9f6654133c079d39b55def4d2bd", "relations": [], "resource_id": "arxiv:2509.13615", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.13615v5", "topics": ["state-tracking", "grounding", "evaluation"], "updated_at": "2026-03-18"}, "text": "See, Think, Act: Teaching Multimodal Agents to Effectively Interact with GUI by Identifying Toggles\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a benchmark for binary GUI-toggle instructions and State-aware Reasoning, which compares the observed control state with the requested state before choosing an action. It evaluates agents on toggle tasks and broader interactive benchmarks.\nPlatforms: unspecified\nTopics: state-tracking, grounding, evaluation\nPublished: 2025-09-17\nAuthors: Zongru Wu, Rui Mao, Zhiyuan Tian, Pengzhou Cheng, Tianjie Ju, Zheng Wu, Lingzhong Dong, Haiyue Sheng, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v5\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.13615v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-13615/"}
{"id": "arxiv:2509.13704", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c1a4019df77278943829e16b155771abb1c8ca30726044cf5969227e0437f5f", "submitted_at": "2025-09-17T05:14:11Z", "updated_at": "2025-09-17T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.13704v1"}, "authors": ["Liangtao Lin", "Zhaomeng Zhu", "Tianwei Zhang", "Yonggang Wen"], "canonical_source_url": "https://arxiv.org/abs/2509.13704", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-13704/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c1a4019df77278943829e16b155771abb1c8ca30726044cf5969227e0437f5f", "submitted_at": "2025-09-17T05:14:11Z", "updated_at": "2025-09-17T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.13704v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.13704"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.13704"}], "name": "InfraMind: A Novel Exploration-based GUI Agentic Framework for Mission-critical Industrial Management", "publication": null, "published_at": "2025-09-17", "record_sha256": "fef7f1490d27fef2502d722fa136831846cb2d908a212d5d2d530c82053072a6", "relations": [], "resource_id": "arxiv:2509.13704", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.13704v1", "topics": ["industry", "planning", "memory", "safety-security"], "updated_at": "2025-09-17"}, "text": "InfraMind: A Novel Exploration-based GUI Agentic Framework for Mission-critical Industrial Management\nResource type: paper\nReview status: curated\nScope: core\n\nInfraMind explores industrial management GUIs using virtual-machine snapshots and builds memory for planning and hierarchical state recognition. Its framework also studies knowledge distillation and safety checks on open-source and commercial data-center management platforms.\nPlatforms: unspecified\nTopics: industry, planning, memory, safety-security\nPublished: 2025-09-17\nAuthors: Liangtao Lin, Zhaomeng Zhu, Tianwei Zhang, Yonggang Wen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.13704v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-13704/"}
{"id": "arxiv:2509.14172", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc1e4f78f1d3ffdb1b67f7c0bb645561a76db19dfa5f442fa9aa0a0d4d1ae0de", "submitted_at": "2025-09-17T16:58:44Z", "updated_at": "2025-09-19T02:13:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.14172v2"}, "authors": ["Ziyuan Chen", "Zhenghui Zhao", "Zhangye Han", "Miancan Liu", "Xianhang Ye", "Yiqing Li", "Hongbo Min", "Jinkui Ren", "Xiantao Zhang", "Guitao Cao"], "canonical_source_url": "https://arxiv.org/abs/2509.14172", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-14172/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc1e4f78f1d3ffdb1b67f7c0bb645561a76db19dfa5f442fa9aa0a0d4d1ae0de", "submitted_at": "2025-09-17T16:58:44Z", "updated_at": "2025-09-19T02:13:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.14172v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.14172"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.14172"}], "name": "TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning", "publication": null, "published_at": "2025-09-17", "record_sha256": "19f9983d0b3b33c61709436c5a5499be343a6972a684c774aee0812985f1553f", "relations": [], "resource_id": "arxiv:2509.14172", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.14172v2", "topics": ["reinforcement-learning", "reward-models", "web-navigation"], "updated_at": "2025-09-19"}, "text": "TGPO: Tree-Guided Preference Optimization for Robust Web Agent Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nTGPO merges semantically equivalent web states into tree-structured trajectories for offline preference optimization. A process reward model scores subgoal progress, redundant behavior, and action validity, while dynamic weighting emphasizes consequential decisions.\nPlatforms: web\nTopics: reinforcement-learning, reward-models, web-navigation\nPublished: 2025-09-17\nAuthors: Ziyuan Chen, Zhenghui Zhao, Zhangye Han, Miancan Liu, Xianhang Ye, Yiqing Li, Hongbo Min, Jinkui Ren, Xiantao Zhang, Guitao Cao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.14172v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-14172/"}
{"id": "arxiv:2509.14382", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac93e8cfca06ff86fad06408051abaa6b6510a44d7ad9f2d34c478b11391a32f", "submitted_at": "2025-09-17T19:34:49Z", "updated_at": "2025-09-17T19:34:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.14382v1"}, "authors": ["Daniel Röder", "Akhil Juneja", "Roland Roller", "Sven Schmeier"], "canonical_source_url": "https://arxiv.org/abs/2509.14382", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-14382/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac93e8cfca06ff86fad06408051abaa6b6510a44d7ad9f2d34c478b11391a32f", "submitted_at": "2025-09-17T19:34:49Z", "updated_at": "2025-09-17T19:34:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.14382v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.14382"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.14382"}], "name": "Detecting Pipeline Failures through Fine-Grained Analysis of Web Agents", "publication": null, "published_at": "2025-09-17", "record_sha256": "c29e2eeb1ed50432bdfae0cb0302c0c502f986b028b9f916051b77ddbc2f2ce8", "relations": [], "resource_id": "arxiv:2509.14382", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.14382v1", "topics": ["evaluation", "observability", "web-navigation"], "updated_at": "2025-09-17"}, "text": "Detecting Pipeline Failures through Fine-Grained Analysis of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a modular framework for identifying errors at intermediate stages of a web-agent pipeline. A SeeAct and Mind2Web case study examines weaknesses that are obscured by final task-success metrics.\nPlatforms: web\nTopics: evaluation, observability, web-navigation\nPublished: 2025-09-17\nAuthors: Daniel Röder, Akhil Juneja, Roland Roller, Sven Schmeier\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.14382v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-14382/"}
{"id": "arxiv:2509.15221", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70d09da7d4990118b2b0745cbddf2bab25852ba3ec26625902fee94cdab37c8d", "submitted_at": "2025-09-18T17:59:22Z", "updated_at": "2025-09-19T05:29:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.15221v2"}, "authors": ["Zhaoyang Liu", "Jingjing Xie", "Zichen Ding", "Zehao Li", "Bowen Yang", "Zhenyu Wu", "Xuehui Wang", "Qiushi Sun", "Shi Liu", "Weiyun Wang", "Shenglong Ye", "Qingyun Li", "Xuan Dong", "Yue Yu", "Chenyu Lu", "YunXiang Mo", "Yao Yan", "Zeyue Tian", "Xiao Zhang", "Yuan Huang", "Yiqian Liu", "Weijie Su", "Gen Luo", "Xiangyu Yue", "Biqing Qi", "Kai Chen", "Bowen Zhou", "Yu Qiao", "Qifeng Chen", "Wenhai Wang"], "canonical_source_url": "https://arxiv.org/abs/2509.15221", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-15221/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70d09da7d4990118b2b0745cbddf2bab25852ba3ec26625902fee94cdab37c8d", "submitted_at": "2025-09-18T17:59:22Z", "updated_at": "2025-09-19T05:29:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.15221v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.15221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15221"}], "name": "ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data", "publication": null, "published_at": "2025-09-18", "record_sha256": "f59939b755decdfe86d34f2d49468ba9395282eb63346021094d846e07019d5f", "relations": [], "resource_id": "arxiv:2509.15221", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.15221v2", "topics": ["data-generation", "automation"], "updated_at": "2025-09-19"}, "text": "ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a cross-platform computer-use dataset and agents trained with a collection pipeline combining automated interaction and human input.\nPlatforms: cross-platform\nTopics: data-generation, automation\nPublished: 2025-09-18\nAuthors: Zhaoyang Liu, Jingjing Xie, Zichen Ding, Zehao Li, Bowen Yang, Zhenyu Wu, Xuehui Wang, Qiushi Sun, Shi Liu, Weiyun Wang, Shenglong Ye, Qingyun Li, Xuan Dong, Yue Yu, Chenyu Lu, YunXiang Mo, Yao Yan, Zeyue Tian, Xiao Zhang, Yuan Huang, Yiqian Liu, Weijie Su, Gen Luo, Xiangyu Yue, Biqing Qi, Kai Chen, Bowen Zhou, Yu Qiao, Qifeng Chen, Wenhai Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.15221v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-15221/"}
{"id": "arxiv:2509.15532", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e415074c6a4b4995350d6936d338682b9a4469f1cbed3a83e0ac44eee6f02bef", "submitted_at": "2025-09-19T02:34:50Z", "updated_at": "2025-09-19T02:34:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15532v1"}, "authors": ["Xianhang Ye", "Yiqing Li", "Wei Dai", "Miancan Liu", "Ziyuan Chen", "Zhangye Han", "Hongbo Min", "Jinkui Ren", "Xiantao Zhang", "Wen Yang", "Zhi Jin"], "canonical_source_url": "https://arxiv.org/abs/2509.15532", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-15532/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e415074c6a4b4995350d6936d338682b9a4469f1cbed3a83e0ac44eee6f02bef", "submitted_at": "2025-09-19T02:34:50Z", "updated_at": "2025-09-19T02:34:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15532v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.15532"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15532"}], "name": "GUI-ARP: Enhancing Grounding with Adaptive Region Perception for GUI Agents", "publication": null, "published_at": "2025-09-19", "record_sha256": "16c423240a1b91bbec1009286aabb32c673fd87b7b1c8297db60208786a0a26a", "relations": [], "resource_id": "arxiv:2509.15532", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.15532v1", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2025-09-19"}, "text": "GUI-ARP: Enhancing Grounding with Adaptive Region Perception for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-ARP uses visual attention to crop relevant screenshot regions and decides whether a target needs single-stage or multi-stage analysis. Supervised fine-tuning followed by GRPO trains this adaptive procedure for high-resolution GUI grounding.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, screen-understanding\nPublished: 2025-09-19\nAuthors: Xianhang Ye, Yiqing Li, Wei Dai, Miancan Liu, Ziyuan Chen, Zhangye Han, Hongbo Min, Jinkui Ren, Xiantao Zhang, Wen Yang, Zhi Jin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.15532v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-15532/"}
{"id": "arxiv:2509.15566", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d190eefae9309d87f4ea4780d3483ae0ebd834e0bf70e7f2af204c3788d87e7e", "submitted_at": "2025-09-19T04:03:44Z", "updated_at": "2025-10-27T11:34:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.15566v4"}, "authors": ["Shaojie Zhang", "Ruoceng Zhang", "Pei Fu", "Shaokang Wang", "Jiahui Yang", "Xin Du", "Shiqi Cui", "Bin Qin", "Ying Huang", "Zhenbo Luo", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2509.15566", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-15566/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d190eefae9309d87f4ea4780d3483ae0ebd834e0bf70e7f2af204c3788d87e7e", "submitted_at": "2025-09-19T04:03:44Z", "updated_at": "2025-10-27T11:34:58Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.15566v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.15566"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15566"}], "name": "BTL-UI: Blink-Think-Link Reasoning Model for GUI Agent", "publication": null, "published_at": "2025-09-19", "record_sha256": "73cca12ef2d3eda9273be14e4f7ae4c1dd61cfcf30dd248272fb53913c66bb7a", "relations": [], "resource_id": "arxiv:2509.15566", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.15566v4", "topics": ["grounding", "planning", "data-generation", "reinforcement-learning"], "updated_at": "2025-10-27"}, "text": "BTL-UI: Blink-Think-Link Reasoning Model for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nBTL-UI separates GUI interaction into screen-region attention, planning and executable-command generation. It adds an automated annotation pipeline and rule-based process/outcome rewards for training, with evaluation on both static understanding and dynamic interaction.\nPlatforms: unspecified\nTopics: grounding, planning, data-generation, reinforcement-learning\nPublished: 2025-09-19\nAuthors: Shaojie Zhang, Ruoceng Zhang, Pei Fu, Shaokang Wang, Jiahui Yang, Xin Du, Shiqi Cui, Bin Qin, Ying Huang, Zhenbo Luo, Jian Luan\nRecorded arXiv version: v4\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.15566v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-15566/"}
{"id": "arxiv:2509.15738", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "035048459d2b303721ef0738212829d1fa3314ccd0736b6649cc786c6e975ad4", "submitted_at": "2025-09-19T08:09:18Z", "updated_at": "2025-09-19T08:09:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15738v1"}, "authors": ["Musen Lin", "Minghao Liu", "Taoran Lu", "Lichen Yuan", "Yiwei Liu", "Haonan Xu", "Yu Miao", "Yuhao Chao", "Zhaojian Li"], "canonical_source_url": "https://arxiv.org/abs/2509.15738", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-15738/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "035048459d2b303721ef0738212829d1fa3314ccd0736b6649cc786c6e975ad4", "submitted_at": "2025-09-19T08:09:18Z", "updated_at": "2025-09-19T08:09:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.15738v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.15738"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.15738"}], "name": "GUI-ReWalk: Massive Data Generation for GUI Agent via Stochastic Exploration and Intent-Aware Reasoning", "publication": null, "published_at": "2025-09-19", "record_sha256": "09fdcf1b934b3cb4fef3b56e0305ecd7cb0e1c9126e95655d4ab5760e248f6db", "relations": [], "resource_id": "arxiv:2509.15738", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.15738v1", "topics": ["data-generation", "cross-app", "long-horizon", "planning"], "updated_at": "2025-09-19"}, "text": "GUI-ReWalk: Massive Data Generation for GUI Agent via Stochastic Exploration and Intent-Aware Reasoning\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-ReWalk begins with stochastic interface exploration and then infers goals to generate coherent action trajectories. It supports long-horizon workflows across applications and studies training on the resulting data using grounding and navigation benchmarks.\nPlatforms: android\nTopics: data-generation, cross-app, long-horizon, planning\nPublished: 2025-09-19\nAuthors: Musen Lin, Minghao Liu, Taoran Lu, Lichen Yuan, Yiwei Liu, Haonan Xu, Yu Miao, Yuhao Chao, Zhaojian Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.15738v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-15738/"}
{"id": "arxiv:2509.17328", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97b7c0690ba7deb7dbbdac6ac94dd82e3864c440cb4325bf143ad735919bb08b", "submitted_at": "2025-09-22T03:04:53Z", "updated_at": "2025-09-22T03:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17328v1"}, "authors": ["Hongxin Li", "Jingran Su", "Jingfan Chen", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2509.17328", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-17328/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97b7c0690ba7deb7dbbdac6ac94dd82e3864c440cb4325bf143ad735919bb08b", "submitted_at": "2025-09-22T03:04:53Z", "updated_at": "2025-09-22T03:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17328v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.17328"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.17328"}], "name": "UIPro: Unleashing Superior Interaction Capability For GUI Agents", "publication": null, "published_at": "2025-09-22", "record_sha256": "72ed039e0762cdecc3006f6683b170a3b006591a650c951f0196f24e6fc4e171", "relations": [], "resource_id": "arxiv:2509.17328", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.17328v1", "topics": ["data-generation", "action-prediction", "grounding", "imitation-learning"], "updated_at": "2025-09-22"}, "text": "UIPro: Unleashing Superior Interaction Capability For GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUIPro pretrains GUI understanding on 20.6 million tasks and then harmonizes interaction datasets through a unified action space. Continued fine-tuning uses the merged multi-platform data to train action prediction for generalist GUI operation.\nPlatforms: cross-platform\nTopics: data-generation, action-prediction, grounding, imitation-learning\nPublished: 2025-09-22\nAuthors: Hongxin Li, Jingran Su, Jingfan Chen, Zheng Ju, Yuntao Chen, Qing Li, Zhaoxiang Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.17328v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-17328/"}
{"id": "arxiv:2509.17917", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b85c9bae829bad343dd6f57c866ea100859fa0ad67f5c1a3ac9b3e53d617486", "submitted_at": "2025-09-22T15:40:31Z", "updated_at": "2025-09-22T15:40:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17917v1"}, "authors": ["Junyu Lu", "Songxin Zhang", "Zejian Xie", "Zhuoyang Song", "Jiaxing Zhang"], "canonical_source_url": "https://arxiv.org/abs/2509.17917", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-17917/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b85c9bae829bad343dd6f57c866ea100859fa0ad67f5c1a3ac9b3e53d617486", "submitted_at": "2025-09-22T15:40:31Z", "updated_at": "2025-09-22T15:40:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.17917v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.17917"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.17917"}], "name": "Orcust: Stepwise-Feedback Reinforcement Learning for GUI Agent", "publication": null, "published_at": "2025-09-22", "record_sha256": "f7ba1b2fb51bd6dc481656d791faa488d1afe9c6e6a3c5e01ec886a60641d81d", "relations": [], "resource_id": "arxiv:2509.17917", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.17917v1", "topics": ["reinforcement-learning", "reward-models", "data-generation", "sandbox"], "updated_at": "2025-09-22"}, "text": "Orcust: Stepwise-Feedback Reinforcement Learning for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nOrcust combines principle-constrained reward modeling with GUI trajectory collection in instrumented virtual machines. Environment-checkable and model-derived principles guide stepwise feedback for grounding, basic operations, and complete task execution.\nPlatforms: unspecified\nTopics: reinforcement-learning, reward-models, data-generation, sandbox\nPublished: 2025-09-22\nAuthors: Junyu Lu, Songxin Zhang, Zejian Xie, Zhuoyang Song, Jiaxing Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.17917v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-17917/"}
{"id": "arxiv:2509.18119", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86bce7c1a120e0e953a239f2a9b39a04a844e6f9710525245d643932ec3c3233", "submitted_at": "2025-09-10T13:09:27Z", "updated_at": "2025-10-24T17:13:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.18119v2"}, "authors": ["Yifan Xu", "Xiao Liu", "Xinghan Liu", "Jiaqi Fu", "Hanchen Zhang", "Bohao Jing", "Shudan Zhang", "Yuting Wang", "Wenyi Zhao", "Yuxiao Dong"], "canonical_source_url": "https://arxiv.org/abs/2509.18119", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-18119/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86bce7c1a120e0e953a239f2a9b39a04a844e6f9710525245d643932ec3c3233", "submitted_at": "2025-09-10T13:09:27Z", "updated_at": "2025-10-24T17:13:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.18119v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.18119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.18119"}], "name": "MobileRL: Online Agentic Reinforcement Learning for Mobile GUI Agents", "publication": null, "published_at": "2025-09-10", "record_sha256": "3f02e90c5e3e7cf83efbd8deb39f553a4260dd17dedd44e5896aa47d99a700fe", "relations": [], "resource_id": "arxiv:2509.18119", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.18119v2", "topics": ["reinforcement-learning", "mobile-navigation", "efficiency"], "updated_at": "2025-10-24"}, "text": "MobileRL: Online Agentic Reinforcement Learning for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMobileRL adapts GRPO to mobile GUI training with difficulty-dependent replay of successes and filtering of failure curricula. A shortest-path reward adjustment accounts for task length, with evaluations on AndroidWorld and AndroidLab.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation, efficiency\nPublished: 2025-09-10\nAuthors: Yifan Xu, Xiao Liu, Xinghan Liu, Jiaqi Fu, Hanchen Zhang, Bohao Jing, Shudan Zhang, Yuting Wang, Wenyi Zhao, Yuxiao Dong\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.18119v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-18119/"}
{"id": "arxiv:2509.20729", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "571eee6476c9e75ffae65cf98ec589e98fd3b953313cc6d2191bb48bd27303fe", "submitted_at": "2025-09-25T04:21:31Z", "updated_at": "2025-12-01T08:01:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.20729v2"}, "authors": ["Jiazheng Sun", "Ruimeng Yang", "Xu Han", "Jiayang Niu", "Mingxuan Li", "Te Yang", "Yongyong Lu", "Xin Peng"], "canonical_source_url": "https://arxiv.org/abs/2509.20729", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-20729/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "571eee6476c9e75ffae65cf98ec589e98fd3b953313cc6d2191bb48bd27303fe", "submitted_at": "2025-09-25T04:21:31Z", "updated_at": "2025-12-01T08:01:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.20729v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.20729"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.20729"}], "name": "Robust, Observable, and Evolvable Agentic Systems Engineering: A Principled Framework Validated via the Fairy GUI Agent", "publication": null, "published_at": "2025-09-25", "record_sha256": "2ca2108d2fc57d956c263ffb04c48b7633adad915c240f63cdcf11a7b2e0957a", "relations": [], "resource_id": "arxiv:2509.20729", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.20729v2", "topics": ["observability", "human-in-the-loop", "memory", "robustness"], "updated_at": "2025-12-01"}, "text": "Robust, Observable, and Evolvable Agentic Systems Engineering: A Principled Framework Validated via the Fairy GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nPresents an agent-engineering framework combining runtime goal clarification, an observable component architecture and evolving execution memory. The Fairy mobile GUI agent evaluates these mechanisms on ambiguous tasks and through controlled and user studies.\nPlatforms: unspecified\nTopics: observability, human-in-the-loop, memory, robustness\nPublished: 2025-09-25\nAuthors: Jiazheng Sun, Ruimeng Yang, Xu Han, Jiayang Niu, Mingxuan Li, Te Yang, Yongyong Lu, Xin Peng\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.20729v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-20729/"}
{"id": "arxiv:2509.21552", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83fa51d3642fa7786f7f6c44fcf5d5ab5876d7ae26e5f6184ddff4d5e2009181", "submitted_at": "2025-09-25T20:38:01Z", "updated_at": "2026-05-25T20:58:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21552v2"}, "authors": ["Yu Zhao", "Wei-Ning Chen", "Huseyin Atahan Inan", "Samuel Kessler", "Lu Wang", "Lukas Wutschitz", "Fangkai Yang", "Chaoyun Zhang", "Pasquale Minervini", "Saravan Rajmohan", "Robert Sim"], "canonical_source_url": "https://arxiv.org/abs/2509.21552", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-21552/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "83fa51d3642fa7786f7f6c44fcf5d5ab5876d7ae26e5f6184ddff4d5e2009181", "submitted_at": "2025-09-25T20:38:01Z", "updated_at": "2026-05-25T20:58:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21552v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.21552"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21552"}], "name": "Learning GUI Grounding with Spatial Reasoning from Visual Feedback", "publication": null, "published_at": "2025-09-25", "record_sha256": "96378a38a967ea5a0e00e4732c4e7d526aabea0d3b100473ef0bd5a7e755c8e3", "relations": [], "resource_id": "arxiv:2509.21552", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.21552v2", "topics": ["grounding", "reinforcement-learning", "action-prediction", "robustness"], "updated_at": "2026-05-25"}, "text": "Learning GUI Grounding with Spatial Reasoning from Visual Feedback\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Cursor treats GUI grounding as a sequence of cursor movements guided by visible feedback and movement history. Online reinforcement learning with trajectory rewards trains the model to refine the target location across multiple steps.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, action-prediction, robustness\nPublished: 2025-09-25\nAuthors: Yu Zhao, Wei-Ning Chen, Huseyin Atahan Inan, Samuel Kessler, Lu Wang, Lukas Wutschitz, Fangkai Yang, Chaoyun Zhang, Pasquale Minervini, Saravan Rajmohan, Robert Sim\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.21552v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-21552/"}
{"id": "arxiv:2509.21782", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fed6f081e0330c54a304c79e8a4183865aae48f0e944545029d3de9852f8028", "submitted_at": "2025-09-26T02:38:14Z", "updated_at": "2026-03-04T16:31:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21782v2"}, "authors": ["Junliang Liu", "Jingyu Xiao", "Wenxin Tang", "Zhixian Wang", "Zipeng Xie", "Wenxuan Wang", "Minrui Zhang", "Shuanghe Yu"], "canonical_source_url": "https://arxiv.org/abs/2509.21782", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-21782/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fed6f081e0330c54a304c79e8a4183865aae48f0e944545029d3de9852f8028", "submitted_at": "2025-09-26T02:38:14Z", "updated_at": "2026-03-04T16:31:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21782v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2509.21782"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21782"}], "name": "Benchmarking MLLM-based Web Understanding: Reasoning, Robustness and Safety", "publication": null, "published_at": "2025-09-26", "record_sha256": "1250dcccceb4f91e75ee54bf2be7586e4f731a71fce77da0ad809b0724fe2ccf", "relations": [], "resource_id": "arxiv:2509.21782", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.21782v2", "topics": ["evaluation", "screen-understanding", "safety-security", "robustness"], "updated_at": "2026-03-04"}, "text": "Benchmarking MLLM-based Web Understanding: Reasoning, Robustness and Safety\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates multimodal LLMs on web interface reasoning, visual perturbation robustness, and safety-critical web understanding tasks.\nPlatforms: web\nTopics: evaluation, screen-understanding, safety-security, robustness\nPublished: 2025-09-26\nAuthors: Junliang Liu, Jingyu Xiao, Wenxin Tang, Zhixian Wang, Zipeng Xie, Wenxuan Wang, Minrui Zhang, Shuanghe Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.21782v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-21782/"}
{"id": "arxiv:2509.21823", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c96f310fb0f28888c59dce71d08697b603e7a7562fdf99b51e3504b93e9e81b3", "submitted_at": "2025-09-26T03:29:36Z", "updated_at": "2026-04-15T09:19:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21823v2"}, "authors": ["Gaole Dai", "Shiqi Jiang", "Ting Cao", "Yuqing Yang", "Yuanchun Li", "Rui Tan", "Mo Li", "Lili Qiu"], "canonical_source_url": "https://arxiv.org/abs/2509.21823", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-21823/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c96f310fb0f28888c59dce71d08697b603e7a7562fdf99b51e3504b93e9e81b3", "submitted_at": "2025-09-26T03:29:36Z", "updated_at": "2026-04-15T09:19:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21823v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.21823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21823"}], "name": "ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor Collaboration", "publication": null, "published_at": "2025-09-26", "record_sha256": "21576a8e309a7dc869a35905c5e1c7f10b0083e083e57502dd07f6fb7826f92c", "relations": [], "resource_id": "arxiv:2509.21823", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.21823v2", "topics": ["reward-models", "evaluation", "planning"], "updated_at": "2026-04-15"}, "text": "ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor Collaboration\nResource type: paper\nReview status: curated\nScope: core\n\nProRe evaluates GUI task outcomes by having a reasoner assign targeted environment-probing tasks to specialist actors. The additional observations support reward judgments, with separate evaluation of judge accuracy and downstream agent task success.\nPlatforms: unspecified\nTopics: reward-models, evaluation, planning\nPublished: 2025-09-26\nAuthors: Gaole Dai, Shiqi Jiang, Ting Cao, Yuqing Yang, Yuanchun Li, Rui Tan, Mo Li, Lili Qiu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.21823v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-21823/"}
{"id": "arxiv:2509.21982", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d23b9f33282f77b9d4965078ca993bcf21163dc1d3210aaa4e554d4fc12b37c5", "submitted_at": "2025-09-26T07:05:01Z", "updated_at": "2026-04-13T03:14:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21982v2"}, "authors": ["Renqi Chen", "Zeyin Tao", "Jianming Guo", "Jingzhe Zhu", "Yiheng Peng", "Qingqing Sun", "Tianyi Zhang", "Shuai Chen"], "canonical_source_url": "https://arxiv.org/abs/2509.21982", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-21982/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d23b9f33282f77b9d4965078ca993bcf21163dc1d3210aaa4e554d4fc12b37c5", "submitted_at": "2025-09-26T07:05:01Z", "updated_at": "2026-04-13T03:14:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.21982v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.21982"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.21982"}], "name": "RISK: A Framework for GUI Agents in E-commerce Risk Management", "publication": null, "published_at": "2025-09-26", "record_sha256": "c0893ed6d68edd260ccfa49f1e8b57d5dc158aa2a4469d2a52a6b934dce337fb", "relations": [], "resource_id": "arxiv:2509.21982", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.21982v2", "topics": ["industry", "data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2026-04-13"}, "text": "RISK: A Framework for GUI Agents in E-commerce Risk Management\nResource type: paper\nReview status: curated\nScope: core\n\nRISK combines browser-interaction training data, a benchmark, and reinforcement fine-tuning for e-commerce risk-management workflows. Its trajectories and rewards distinguish individual actions from multi-step stateful tasks involving deeply embedded web content.\nPlatforms: web\nTopics: industry, data-generation, evaluation, reinforcement-learning\nPublished: 2025-09-26\nAuthors: Renqi Chen, Zeyin Tao, Jianming Guo, Jingzhe Zhu, Yiheng Peng, Qingqing Sun, Tianyi Zhang, Shuai Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.21982v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-21982/"}
{"id": "arxiv:2509.22256", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "922c1d9b60db7c17b9019a320a4cfc2dd4eac325fd946f024ca82d43785f5eba", "submitted_at": "2025-09-26T12:19:27Z", "updated_at": "2026-01-14T06:07:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.22256v4"}, "authors": ["Haochen Gong", "Chenxiao Li", "Rui Chang", "Wenbo Shen"], "canonical_source_url": "https://arxiv.org/abs/2509.22256", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-22256/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "922c1d9b60db7c17b9019a320a4cfc2dd4eac325fd946f024ca82d43785f5eba", "submitted_at": "2025-09-26T12:19:27Z", "updated_at": "2026-01-14T06:07:30Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2509.22256v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.22256"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.22256"}], "name": "Secure and Efficient Access Control for Computer-Use Agents via Context Space", "publication": null, "published_at": "2025-09-26", "record_sha256": "8d7d40ce5e8cb207ac4dff564649c8e575318f1916d4652d606c05952186d0a7", "relations": [], "resource_id": "arxiv:2509.22256", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.22256v4", "topics": ["safety-security", "tool-integration"], "updated_at": "2026-01-14"}, "text": "Secure and Efficient Access Control for Computer-Use Agents via Context Space\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CSAgent, which combines static access policies with user intent and context and enforces them through an operating-system service. A supporting toolchain constructs and refines policies for agents using API, CLI and GUI interfaces.\nPlatforms: unspecified\nTopics: safety-security, tool-integration\nPublished: 2025-09-26\nAuthors: Haochen Gong, Chenxiao Li, Rui Chang, Wenbo Shen\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.22256v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-22256/"}
{"id": "arxiv:2509.23263", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b5375e77f087367383bb5de96cec46d8de6550d9a44bff6965b9201b6718a05", "submitted_at": "2025-09-27T11:42:36Z", "updated_at": "2026-08-30T18:12:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.23263v3"}, "authors": ["Tao Xiong", "Xavier Hu", "Yurun Chen", "Yuhang Liu", "Changqiao Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2509.23263", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-23263/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b5375e77f087367383bb5de96cec46d8de6550d9a44bff6965b9201b6718a05", "submitted_at": "2025-09-27T11:42:36Z", "updated_at": "2026-08-30T18:12:14Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2509.23263v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2509.23263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.23263"}], "name": "GUI-PRA: Process Reward Agent for GUI Tasks", "publication": null, "published_at": "2025-09-27", "record_sha256": "106506e9e46508e15f49961bb5efbb074934863a82ffa61188abf7461e4c84c7", "relations": [], "resource_id": "arxiv:2509.23263", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.23263v3", "topics": ["reward-models", "evaluation", "screen-understanding"], "updated_at": "2026-08-30"}, "text": "GUI-PRA: Process Reward Agent for GUI Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-PRA evaluates GUI progress by generating task-specific criteria and actively inspecting visual evidence with tools. It tests offline judgment on OS-Critic Bench and the effects of this feedback on online mobile policies in AndroidWorld.\nPlatforms: android\nTopics: reward-models, evaluation, screen-understanding\nPublished: 2025-09-27\nAuthors: Tao Xiong, Xavier Hu, Yurun Chen, Yuhang Liu, Changqiao Wu, Pengzhi Gao, Wei Liu, Jian Luan, Shengyu Zhang\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.23263v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-23263/"}
{"id": "arxiv:2509.23866", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a96e774920489ef366ad462972a938863a3e714672e9c1f4b784d57eee34c4ca", "submitted_at": "2025-09-28T13:19:20Z", "updated_at": "2025-09-28T13:19:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.23866v1"}, "authors": ["Pengxiang Li", "Zechen Hu", "Zirui Shang", "Jingrong Wu", "Yang Liu", "Hui Liu", "Zhi Gao", "Chenrui Shi", "Bofei Zhang", "Zihao Zhang", "Xiaochuan Shi", "Zedong YU", "Yuwei Wu", "Xinxiao Wu", "Yunde Jia", "Liuyu Xiang", "Zhaofeng He", "Qing Li"], "canonical_source_url": "https://arxiv.org/abs/2509.23866", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-23866/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a96e774920489ef366ad462972a938863a3e714672e9c1f4b784d57eee34c4ca", "submitted_at": "2025-09-28T13:19:20Z", "updated_at": "2025-09-28T13:19:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.23866v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.23866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.23866"}], "name": "Efficient Multi-turn RL for GUI Agents via Decoupled Training and Adaptive Data Curation", "publication": null, "published_at": "2025-09-28", "record_sha256": "03dc4abd1a838cb6128f8eac37cef46fc1a382a2752b0cb2a5c6885f75a6e961", "relations": [], "resource_id": "arxiv:2509.23866", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.23866v1", "topics": ["reinforcement-learning", "efficiency", "data-generation", "sandbox"], "updated_at": "2025-09-28"}, "text": "Efficient Multi-turn RL for GUI Agents via Decoupled Training and Adaptive Data Curation\nResource type: paper\nReview status: curated\nScope: core\n\nDART separates GUI reinforcement learning into asynchronous environment, rollout, data-management, and training services. It curates experience by task difficulty, successful trajectories, and step uncertainty while correcting rollout-policy mismatch, with evaluation on OSWorld.\nPlatforms: desktop\nTopics: reinforcement-learning, efficiency, data-generation, sandbox\nPublished: 2025-09-28\nAuthors: Pengxiang Li, Zechen Hu, Zirui Shang, Jingrong Wu, Yang Liu, Hui Liu, Zhi Gao, Chenrui Shi, Bofei Zhang, Zihao Zhang, Xiaochuan Shi, Zedong YU, Yuwei Wu, Xinxiao Wu, Yunde Jia, Liuyu Xiang, Zhaofeng He, Qing Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.23866v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-23866/"}
{"id": "arxiv:2509.24133", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b42e784de8dd007c60d234b7331bd98b5a04c58c13696bf436c81c7eccee37f", "submitted_at": "2025-09-29T00:06:31Z", "updated_at": "2025-09-29T00:06:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24133v1"}, "authors": ["Zhecheng Li", "Guoxian Song", "Yiwei Wang", "Zhen Xiong", "Junsong Yuan", "Yujun Cai"], "canonical_source_url": "https://arxiv.org/abs/2509.24133", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-24133/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b42e784de8dd007c60d234b7331bd98b5a04c58c13696bf436c81c7eccee37f", "submitted_at": "2025-09-29T00:06:31Z", "updated_at": "2025-09-29T00:06:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24133v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.24133"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.24133"}], "name": "Generalist Scanner Meets Specialist Locator: A Synergistic Coarse-to-Fine Framework for Robust GUI Grounding", "publication": null, "published_at": "2025-09-29", "record_sha256": "95833e4f087b60e40190a418acc52b00aed4657c7a6d4d009a026ed0561934f3", "relations": [], "resource_id": "arxiv:2509.24133", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.24133v1", "topics": ["grounding", "planning", "evaluation"], "updated_at": "2025-09-29"}, "text": "Generalist Scanner Meets Specialist Locator: A Synergistic Coarse-to-Fine Framework for Robust GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nGMS pairs a general vision-language scanner that proposes interface regions with a specialist locator that predicts coordinates within them. Hierarchical coarse-to-fine search coordinates the two components for GUI grounding on ScreenSpot-Pro.\nPlatforms: unspecified\nTopics: grounding, planning, evaluation\nPublished: 2025-09-29\nAuthors: Zhecheng Li, Guoxian Song, Yiwei Wang, Zhen Xiong, Junsong Yuan, Yujun Cai\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.24133v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-24133/"}
{"id": "arxiv:2509.24183", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1c6fd3e84212bac4c4c8341ffb3478c97becce08f92dac89e5de77ad42cd07d", "submitted_at": "2025-09-29T02:04:20Z", "updated_at": "2025-09-29T02:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24183v1"}, "authors": ["Ran Xu", "Kaixin Ma", "Wenhao Yu", "Hongming Zhang", "Joyce C. Ho", "Carl Yang", "Dong Yu"], "canonical_source_url": "https://arxiv.org/abs/2509.24183", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-24183/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1c6fd3e84212bac4c4c8341ffb3478c97becce08f92dac89e5de77ad42cd07d", "submitted_at": "2025-09-29T02:04:20Z", "updated_at": "2025-09-29T02:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.24183v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:20:58+00:00", "role": "primary", "sha256": "fe9f70994c119a85cb8075461174e9699c423d79094e611f1ac95a51666600c1", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2509.24183v1"}], "identifiers": {"arxiv": "2509.24183"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.24183"}], "name": "Retrieval-augmented GUI Agents with Generative Guidelines", "publication": null, "published_at": "2025-09-29", "record_sha256": "a1108f3ae77f8fdb1c6fd1fb23c465a1c351cd5af01b46101a4a5c689b32a849", "relations": [], "resource_id": "arxiv:2509.24183", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.24183v1", "topics": ["memory", "planning", "imitation-learning", "web-navigation", "mobile-navigation"], "updated_at": "2025-09-29"}, "text": "Retrieval-augmented GUI Agents with Generative Guidelines\nResource type: paper\nReview status: curated\nScope: core\n\nRAG-GUI trains a small guidance adapter to filter and summarize retrieved web tutorials while keeping the acting agent fixed. Supervised and rejection-sampling stages prepare the adapter for inference-time assistance, tested on offline web/mobile tasks and interactive Android tasks.\nPlatforms: web, android\nTopics: memory, planning, imitation-learning, web-navigation, mobile-navigation\nPublished: 2025-09-29\nAuthors: Ran Xu, Kaixin Ma, Wenhao Yu, Hongming Zhang, Joyce C. Ho, Carl Yang, Dong Yu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.24183v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-24183/"}
{"id": "arxiv:2509.25047", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d41da9264d664abbeacccae889f9f14f403fc4edbd7453ed4ff276689592fe68", "submitted_at": "2025-09-29T17:00:02Z", "updated_at": "2025-09-29T17:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.25047v1"}, "authors": ["Ram Ramrakhya", "Andrew Szot", "Omar Attia", "Yuhao Yang", "Anh Nguyen", "Bogdan Mazoure", "Zhe Gan", "Harsh Agrawal", "Alexander Toshev"], "canonical_source_url": "https://arxiv.org/abs/2509.25047", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-25047/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d41da9264d664abbeacccae889f9f14f403fc4edbd7453ed4ff276689592fe68", "submitted_at": "2025-09-29T17:00:02Z", "updated_at": "2025-09-29T17:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.25047v1"}}, "environments": ["android", "linux", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2509.25047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.25047"}], "name": "Scaling Synthetic Task Generation for Agents via Exploration", "publication": null, "published_at": "2025-09-29", "record_sha256": "12156b95310309da37444c0c297857ed1ac2824ff9eee8bb02b72b1fbd14b7b1", "relations": [], "resource_id": "arxiv:2509.25047", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.25047v1", "topics": ["data-generation", "reinforcement-learning", "imitation-learning"], "updated_at": "2025-09-29"}, "text": "Scaling Synthetic Task Generation for Agents via Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nAutoPlay explores application states and functions to generate tasks, then uses a model executor and verifier to construct demonstrations. It produces 20,000 tasks across 20 Android apps and 10,000 tasks across 13 Ubuntu apps for supervised and reinforcement learning of UI agents.\nPlatforms: android, linux, desktop, cross-platform\nTopics: data-generation, reinforcement-learning, imitation-learning\nPublished: 2025-09-29\nAuthors: Ram Ramrakhya, Andrew Szot, Omar Attia, Yuhao Yang, Anh Nguyen, Bogdan Mazoure, Zhe Gan, Harsh Agrawal, Alexander Toshev\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.25047v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-25047/"}
{"id": "arxiv:2509.25140", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d984883b65895d02debd54edbd198f0aadf8bd295da869680c3ebbe08d2e9492", "submitted_at": "2025-09-29T17:51:03Z", "updated_at": "2026-03-16T20:49:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.25140v2"}, "authors": ["Siru Ouyang", "Jun Yan", "I-Hung Hsu", "Yanfei Chen", "Ke Jiang", "Zifeng Wang", "Rujun Han", "Long T. Le", "Samira Daruki", "Xiangru Tang", "Vishy Tirumalashetty", "George Lee", "Mahsan Rofouei", "Hangfei Lin", "Jiawei Han", "Chen-Yu Lee", "Tomas Pfister"], "canonical_source_url": "https://arxiv.org/abs/2509.25140", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-25140/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d984883b65895d02debd54edbd198f0aadf8bd295da869680c3ebbe08d2e9492", "submitted_at": "2025-09-29T17:51:03Z", "updated_at": "2026-03-16T20:49:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2509.25140v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2509.25140"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.25140"}], "name": "ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory", "publication": null, "published_at": "2025-09-29", "record_sha256": "6782932e6a20f19b80e8c073b400b3492360f3400051636417ceb9a8fa0aaea7", "relations": [], "resource_id": "arxiv:2509.25140", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.25140v2", "topics": ["memory", "self-reflection", "web-navigation"], "updated_at": "2026-03-16"}, "text": "ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memory\nResource type: paper\nReview status: curated\nScope: core\n\nDistills generalizable reasoning strategies from both successful and failed agent trajectories into a persistent memory bank for test-time retrieval.\nPlatforms: web\nTopics: memory, self-reflection, web-navigation\nPublished: 2025-09-29\nAuthors: Siru Ouyang, Jun Yan, I-Hung Hsu, Yanfei Chen, Ke Jiang, Zifeng Wang, Rujun Han, Long T. Le, Samira Daruki, Xiangru Tang, Vishy Tirumalashetty, George Lee, Mahsan Rofouei, Hangfei Lin, Jiawei Han, Chen-Yu Lee, Tomas Pfister\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.25140v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-25140/"}
{"id": "arxiv:2509.26506", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e22d26dfbda8530f32d4686f26d91ea1a8e8c5470e857f2e13803258a2f7aaf", "submitted_at": "2025-09-30T16:48:49Z", "updated_at": "2025-09-30T16:48:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26506v1"}, "authors": ["Yutong Dai", "Krithika Ramakrishnan", "Jing Gu", "Matthew Fernandez", "Yanqi Luo", "Viraj Prabhu", "Zhenyu Hu", "Silvio Savarese", "Caiming Xiong", "Zeyuan Chen", "Ran Xu"], "canonical_source_url": "https://arxiv.org/abs/2509.26506", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-26506/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e22d26dfbda8530f32d4686f26d91ea1a8e8c5470e857f2e13803258a2f7aaf", "submitted_at": "2025-09-30T16:48:49Z", "updated_at": "2025-09-30T16:48:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26506v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.26506"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.26506"}], "name": "SCUBA: Salesforce Computer Use Benchmark", "publication": null, "published_at": "2025-09-30", "record_sha256": "989a29bce4a4450aeec3c47b2ad675645671dd691ac1b4815e33f452038594e2", "relations": [], "resource_id": "arxiv:2509.26506", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.26506v1", "topics": ["evaluation", "industry", "workflow", "sandbox"], "updated_at": "2025-09-30"}, "text": "SCUBA: Salesforce Computer Use Benchmark\nResource type: paper\nReview status: curated\nScope: core\n\nSCUBA evaluates enterprise UI navigation and CRM workflows with 300 task instances derived from Salesforce user interviews. Salesforce sandbox execution and milestone-based assessment support comparisons between zero-shot agents and agents given demonstrations.\nPlatforms: unspecified\nTopics: evaluation, industry, workflow, sandbox\nPublished: 2025-09-30\nAuthors: Yutong Dai, Krithika Ramakrishnan, Jing Gu, Matthew Fernandez, Yanqi Luo, Viraj Prabhu, Zhenyu Hu, Silvio Savarese, Caiming Xiong, Zeyuan Chen, Ran Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.26506v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-26506/"}
{"id": "arxiv:2509.26539", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa35e6bcd86fbdb709652d3592776ed93aebe2a0ebb3595fed91478dc0e3e0a", "submitted_at": "2025-09-30T17:13:56Z", "updated_at": "2025-09-30T17:13:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26539v1"}, "authors": ["Zhen Yang", "Zi-Yi Dou", "Di Feng", "Forrest Huang", "Anh Nguyen", "Keen You", "Omar Attia", "Yuhao Yang", "Michael Feng", "Haotian Zhang", "Ram Ramrakhya", "Chao Jia", "Jeffrey Nichols", "Alexander Toshev", "Yinfei Yang", "Zhe Gan"], "canonical_source_url": "https://arxiv.org/abs/2509.26539", "citation_url": "https://awesomecua.github.io/resources/arxiv-2509-26539/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "daa35e6bcd86fbdb709652d3592776ed93aebe2a0ebb3595fed91478dc0e3e0a", "submitted_at": "2025-09-30T17:13:56Z", "updated_at": "2025-09-30T17:13:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2509.26539v1"}}, "environments": ["android", "web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2509.26539"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2509.26539"}], "name": "Ferret-UI Lite: Lessons from Building Small On-Device GUI Agents", "publication": null, "published_at": "2025-09-30", "record_sha256": "4a316abcb1256d8bcb5f2dcd3611ea137761c604f8ee7f5061d6f09035543fca", "relations": [], "resource_id": "arxiv:2509.26539", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2509.26539v1", "topics": ["grounding", "reinforcement-learning", "tool-integration", "efficiency"], "updated_at": "2025-09-30"}, "text": "Ferret-UI Lite: Lessons from Building Small On-Device GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nFerret-UI Lite develops a 3B GUI agent using mixed real and synthetic data, reasoning with visual tools at inference, and reinforcement learning. It evaluates grounding and navigation across mobile, web, and desktop settings, including AndroidWorld and OSWorld.\nPlatforms: android, web, desktop, cross-platform\nTopics: grounding, reinforcement-learning, tool-integration, efficiency\nPublished: 2025-09-30\nAuthors: Zhen Yang, Zi-Yi Dou, Di Feng, Forrest Huang, Anh Nguyen, Keen You, Omar Attia, Yuhao Yang, Michael Feng, Haotian Zhang, Ram Ramrakhya, Chao Jia, Jeffrey Nichols, Alexander Toshev, Yinfei Yang, Zhe Gan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2509.26539v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2509-26539/"}
{"id": "arxiv:2510.00413", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "277315ecc1a341c0a511bce84b87190fc0aeeb2306666ea8d4746aa8e9b39984", "submitted_at": "2025-10-01T01:48:39Z", "updated_at": "2025-10-04T15:23:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00413v2"}, "authors": ["Zikang Liu", "Junyi Li", "Wayne Xin Zhao", "Dawei Gao", "Yaliang Li", "Ji-rong Wen"], "canonical_source_url": "https://arxiv.org/abs/2510.00413", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-00413/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "277315ecc1a341c0a511bce84b87190fc0aeeb2306666ea8d4746aa8e9b39984", "submitted_at": "2025-10-01T01:48:39Z", "updated_at": "2025-10-04T15:23:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00413v2"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.00413"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.00413"}], "name": "PAL-UI: Planning with Active Look-back for Vision-Based GUI Agents", "publication": null, "published_at": "2025-10-01", "record_sha256": "c515712260afa789c862cb87570efe97280621a5e089caeef06dc77fdf5c9892", "relations": [], "resource_id": "arxiv:2510.00413", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.00413v2", "topics": ["memory", "planning", "interaction-history", "mobile-navigation"], "updated_at": "2025-10-04"}, "text": "PAL-UI: Planning with Active Look-back for Vision-Based GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPAL-UI combines observation/action summaries with selective retrieval of historical screenshots during planning. It trains mobile navigation models from step-level supervision and evaluates transfer to web navigation without additional training.\nPlatforms: web, cross-platform\nTopics: memory, planning, interaction-history, mobile-navigation\nPublished: 2025-10-01\nAuthors: Zikang Liu, Junyi Li, Wayne Xin Zhao, Dawei Gao, Yaliang Li, Ji-rong Wen\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.00413v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-00413/"}
{"id": "arxiv:2510.00536", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e07a5a8f9c0aa39643bb919c2d637435669318420223d7d72b0ef3d733308852", "submitted_at": "2025-10-01T05:37:54Z", "updated_at": "2026-10-07T23:23:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00536v2"}, "authors": ["Kung-Hsiang Huang", "Haoyi Qiu", "Yutong Dai", "Caiming Xiong", "Chien-Sheng Wu"], "canonical_source_url": "https://arxiv.org/abs/2510.00536", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-00536/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e07a5a8f9c0aa39643bb919c2d637435669318420223d7d72b0ef3d733308852", "submitted_at": "2025-10-01T05:37:54Z", "updated_at": "2026-10-07T23:23:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.00536v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.00536"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.00536"}], "name": "GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal Awareness", "publication": null, "published_at": "2025-10-01", "record_sha256": "20efd24bf1016eb721a8ab56665f7024865fd17c53933d5d2f33be2b08af528c", "relations": [], "resource_id": "arxiv:2510.00536", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.00536v2", "topics": ["efficiency", "interaction-history", "training-free", "screen-understanding"], "updated_at": "2026-10-07"}, "text": "GUI-KV: Efficient GUI Agents via KV Cache with Spatio-Temporal Awareness\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-KV compresses vision-language agents’ attention caches using visual-token saliency and redundancy between screenshot frames. The method requires no retraining and evaluates how cache budgets affect GUI action accuracy and computational cost.\nPlatforms: unspecified\nTopics: efficiency, interaction-history, training-free, screen-understanding\nPublished: 2025-10-01\nAuthors: Kung-Hsiang Huang, Haoyi Qiu, Yutong Dai, Caiming Xiong, Chien-Sheng Wu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.00536v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-00536/"}
{"id": "arxiv:2510.01354", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e997e87268cfe66869bf19fdc8445e5487c980736a63cbe76f779bd1f142c396", "submitted_at": "2025-10-01T18:34:06Z", "updated_at": "2026-09-23T16:44:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.01354v2"}, "authors": ["Yinuo Liu", "Xilong Wang", "Ruohan Xu", "Yuqi Jia", "Neil Zhenqiang Gong"], "canonical_source_url": "https://arxiv.org/abs/2510.01354", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-01354/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e997e87268cfe66869bf19fdc8445e5487c980736a63cbe76f779bd1f142c396", "submitted_at": "2025-10-01T18:34:06Z", "updated_at": "2026-09-23T16:44:27Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.01354v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.01354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01354"}], "name": "WAInjectBench: Benchmarking Prompt Injection Detections for Web Agents", "publication": null, "published_at": "2025-10-01", "record_sha256": "b2e1f363140c5d4a3b98b3696546b1397cfca223ac3ecbef84d7ee898fa51630", "relations": [], "resource_id": "arxiv:2510.01354", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.01354v2", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-09-23"}, "text": "WAInjectBench: Benchmarking Prompt Injection Detections for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WAInjectBench with a threat-model taxonomy and malicious and benign text-image datasets for web-agent prompt-injection detection. It compares detector families across explicit and less visible forms of injected content.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2025-10-01\nAuthors: Yinuo Liu, Xilong Wang, Ruohan Xu, Yuqi Jia, Neil Zhenqiang Gong\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The reported contribution evaluates attack detection; end-to-end prevention of browser-agent compromise is not inferred.\n\nPrimary source: https://arxiv.org/abs/2510.01354v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-01354/"}
{"id": "arxiv:2510.01524", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "23794419031fe219fc342c041d666045080b274926dc22ab101a65581b4e4efd", "submitted_at": "2025-10-01T23:41:47Z", "updated_at": "2025-10-01T23:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01524v1"}, "authors": ["Viraj Prabhu", "Yutong Dai", "Matthew Fernandez", "Jing Gu", "Krithika Ramakrishnan", "Yanqi Luo", "Silvio Savarese", "Caiming Xiong", "Junnan Li", "Zeyuan Chen", "Ran Xu"], "canonical_source_url": "https://arxiv.org/abs/2510.01524", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-01524/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "23794419031fe219fc342c041d666045080b274926dc22ab101a65581b4e4efd", "submitted_at": "2025-10-01T23:41:47Z", "updated_at": "2025-10-01T23:41:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01524v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:27:51+00:00", "role": "primary", "sha256": "8466eed10f7f630a68fa742ff9de6b5ba052c497ff421acd4fbee529093d3a58", "supports": ["summary", "scope", "environments"], "url": "https://arxiv.org/html/2510.01524v1"}], "identifiers": {"arxiv": "2510.01524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01524"}], "name": "WALT: Web Agents that Learn Tools", "publication": null, "published_at": "2025-10-01", "record_sha256": "d5b5a1c157499f6886f5f26d26236d3efcb67fbcc2eaedf3dbfa83a194519df0", "relations": [], "resource_id": "arxiv:2510.01524", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.01524v1", "topics": ["web-navigation", "tool-integration", "efficiency", "automation"], "updated_at": "2025-10-01"}, "text": "WALT: Web Agents that Learn Tools\nResource type: paper\nReview status: curated\nScope: core\n\nWALT explores website functionality and converts demonstrated interactions into validated, reusable tools combining DOM actions, URL navigation, and agentic fallbacks. It evaluates this expanded browser action space on WebArena and VisualWebArena.\nPlatforms: web\nTopics: web-navigation, tool-integration, efficiency, automation\nPublished: 2025-10-01\nAuthors: Viraj Prabhu, Yutong Dai, Matthew Fernandez, Jing Gu, Krithika Ramakrishnan, Yanqi Luo, Silvio Savarese, Caiming Xiong, Junnan Li, Zeyuan Chen, Ran Xu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.01524v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-01524/"}
{"id": "arxiv:2510.01670", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a6426fa5eeb3dfafef30e47dd46a6a43dff3e24bf0ff485b7d10ab9efe133f94", "submitted_at": "2025-10-02T04:52:15Z", "updated_at": "2025-10-02T04:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01670v1"}, "authors": ["Erfan Shayegani", "Keegan Hines", "Yue Dong", "Nael Abu-Ghazaleh", "Roman Lutz", "Spencer Whitehead", "Vidhisha Balachandran", "Besmira Nushi", "Vibhav Vineet"], "canonical_source_url": "https://arxiv.org/abs/2510.01670", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-01670/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a6426fa5eeb3dfafef30e47dd46a6a43dff3e24bf0ff485b7d10ab9efe133f94", "submitted_at": "2025-10-02T04:52:15Z", "updated_at": "2025-10-02T04:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.01670v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.01670"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.01670"}], "name": "Just Do It!? Computer-Use Agents Exhibit Blind Goal-Directedness", "publication": null, "published_at": "2025-10-02", "record_sha256": "baf4d29d08a629cd8bd3ff0aaf4bea22d8d153f4b64d8fceb4054da06f5b2bce", "relations": [], "resource_id": "arxiv:2510.01670", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.01670v1", "topics": ["evaluation", "safety-security", "robustness"], "updated_at": "2025-10-02"}, "text": "Just Do It!? Computer-Use Agents Exhibit Blind Goal-Directedness\nResource type: paper\nReview status: curated\nScope: core\n\nBLIND-ACT tests whether computer-use agents persist with ambiguous, contradictory, or infeasible goals without adequate contextual reasoning. Its OSWorld-based tasks and judged trajectories examine these behaviors and the effect of mitigation prompts.\nPlatforms: desktop\nTopics: evaluation, safety-security, robustness\nPublished: 2025-10-02\nAuthors: Erfan Shayegani, Keegan Hines, Yue Dong, Nael Abu-Ghazaleh, Roman Lutz, Spencer Whitehead, Vidhisha Balachandran, Besmira Nushi, Vibhav Vineet\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.01670v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-01670/"}
{"id": "arxiv:2510.02250", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29c00ba4713da26f912fec8046d498b4fd10cf71a3109f66e92fa23f7cebcf44", "submitted_at": "2025-10-02T17:37:08Z", "updated_at": "2026-02-03T23:48:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02250v2"}, "authors": ["Gonzalo Gonzalez-Pumariega", "Vincent Tu", "Chih-Lun Lee", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2510.02250", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-02250/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29c00ba4713da26f912fec8046d498b4fd10cf71a3109f66e92fa23f7cebcf44", "submitted_at": "2025-10-02T17:37:08Z", "updated_at": "2026-02-03T23:48:45Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02250v2"}}, "environments": ["desktop", "windows", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.02250"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.02250"}], "name": "Scaling Agents for Computer Use", "publication": null, "published_at": "2025-10-02", "record_sha256": "bdf20ea55b57f7679361ee59782ff409633242cd32f56654dd9fd937c4440c19", "relations": [], "resource_id": "arxiv:2510.02250", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.02250v2", "topics": ["evaluation", "reward-models", "long-horizon"], "updated_at": "2026-02-03"}, "text": "Scaling Agents for Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Behavior Judge, which converts execution trajectories into behavior narratives and compares multiple candidate rollouts. Experiments examine this selection approach on OSWorld, WindowsAgentArena and AndroidWorld.\nPlatforms: desktop, windows, android\nTopics: evaluation, reward-models, long-horizon\nPublished: 2025-10-02\nAuthors: Gonzalo Gonzalez-Pumariega, Vincent Tu, Chih-Lun Lee, Jiachen Yang, Ang Li, Xin Eric Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.02250v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-02250/"}
{"id": "arxiv:2510.02418", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7906e06c2efb8778a43ec774821f90947e70cec0b78d6d43ac8901f7f0562332", "submitted_at": "2025-10-02T15:22:21Z", "updated_at": "2025-10-07T15:12:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02418v2"}, "authors": ["Sagnik Anupam", "Davis Brown", "Shuo Li", "Eric Wong", "Hamed Hassani", "Osbert Bastani"], "canonical_source_url": "https://arxiv.org/abs/2510.02418", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-02418/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7906e06c2efb8778a43ec774821f90947e70cec0b78d6d43ac8901f7f0562332", "submitted_at": "2025-10-02T15:22:21Z", "updated_at": "2025-10-07T15:12:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.02418v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2510.02418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.02418"}], "name": "BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks", "publication": null, "published_at": "2025-10-02", "record_sha256": "7ebcf2f716ef75611837be9c0f2ba06fe581634dbd8cc1aa0af4270e7a370710", "relations": [], "resource_id": "arxiv:2510.02418", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.02418v2", "topics": ["evaluation", "web-navigation", "human-in-the-loop"], "updated_at": "2025-10-07"}, "text": "BrowserArena: Evaluating LLM Agents on Real-World Web Navigation Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces BrowserArena, a live-web platform for head-to-head agent comparisons on user-submitted tasks. Step-level human annotations support targeted analysis of navigation, pop-up handling and CAPTCHA-related failures.\nPlatforms: web\nTopics: evaluation, web-navigation, human-in-the-loop\nPublished: 2025-10-02\nAuthors: Sagnik Anupam, Davis Brown, Shuo Li, Eric Wong, Hamed Hassani, Osbert Bastani\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.02418v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-02418/"}
{"id": "arxiv:2510.03230", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0c7eedadb9a321b222d4081271d6c979a25e1f64714cea8cf2ff088c5e61df", "submitted_at": "2025-10-03T17:59:34Z", "updated_at": "2025-10-03T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03230v1"}, "authors": ["Suyuchen Wang", "Tianyu Zhang", "Ahmed Masry", "Christopher Pal", "Spandana Gella", "Bang Liu", "Perouz Taslakian"], "canonical_source_url": "https://arxiv.org/abs/2510.03230", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-03230/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f0c7eedadb9a321b222d4081271d6c979a25e1f64714cea8cf2ff088c5e61df", "submitted_at": "2025-10-03T17:59:34Z", "updated_at": "2025-10-03T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03230v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.03230"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.03230"}], "name": "Improving GUI Grounding with Explicit Position-to-Coordinate Mapping", "publication": null, "published_at": "2025-10-03", "record_sha256": "0b4e644d85dfcb5d58c488fc51e5e07c0659401c23ca202338329f9bf7856d8c", "relations": [], "resource_id": "arxiv:2510.03230", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.03230v1", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2025-10-03"}, "text": "Improving GUI Grounding with Explicit Position-to-Coordinate Mapping\nResource type: paper\nReview status: curated\nScope: core\n\nThis work addresses GUI localization errors caused by mismatches between visual patch positions and output pixel coordinates. RULER tokens provide explicit coordinate references, while Interleaved MRoPE balances width and height encoding; evaluation uses the ScreenSpot benchmark family.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, robustness\nPublished: 2025-10-03\nAuthors: Suyuchen Wang, Tianyu Zhang, Ahmed Masry, Christopher Pal, Spandana Gella, Bang Liu, Perouz Taslakian\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.03230v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-03230/"}
{"id": "arxiv:2510.03285", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88e25f6d876dc6ad11219e3faa0e7d95f4abb39a9aba32213896225aff64ec42", "submitted_at": "2025-09-28T20:51:05Z", "updated_at": "2025-09-28T20:51:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03285v1"}, "authors": ["Su Kara", "Fazle Faisal", "Suman Nath"], "canonical_source_url": "https://arxiv.org/abs/2510.03285", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-03285/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88e25f6d876dc6ad11219e3faa0e7d95f4abb39a9aba32213896225aff64ec42", "submitted_at": "2025-09-28T20:51:05Z", "updated_at": "2025-09-28T20:51:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.03285v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.03285"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.03285"}], "name": "WAREX: Web Agent Reliability Evaluation on Existing Benchmarks", "publication": null, "published_at": "2025-09-28", "record_sha256": "b8cf745c30a8f1c461a785fc9741e3b0f0cdf4fc68637974b7f7d86c477a7002", "relations": [], "resource_id": "arxiv:2510.03285", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.03285v1", "topics": ["evaluation", "robustness", "web-navigation"], "updated_at": "2025-09-28"}, "text": "WAREX: Web Agent Reliability Evaluation on Existing Benchmarks\nResource type: paper\nReview status: curated\nScope: core\n\nWAREX augments existing web-agent benchmarks with reliability challenges arising from unstable systems and changing or disruptive website behavior. Evaluation on WebArena, WebVoyager and REAL examines how these disturbances affect task completion.\nPlatforms: web\nTopics: evaluation, robustness, web-navigation\nPublished: 2025-09-28\nAuthors: Su Kara, Fazle Faisal, Suman Nath\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.03285v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-03285/"}
{"id": "arxiv:2510.04039", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "798fc5ca4fe483a1235f0979b0c641e9c1732660780c13340c5221319a1b28fc", "submitted_at": "2025-10-05T05:15:45Z", "updated_at": "2025-10-05T05:15:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.04039v1"}, "authors": ["Bin Lei", "Nuo Xu", "Ali Payani", "Mingyi Hong", "Chunhua Liao", "Yu Cao", "Caiwen Ding"], "canonical_source_url": "https://arxiv.org/abs/2510.04039", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-04039/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "798fc5ca4fe483a1235f0979b0c641e9c1732660780c13340c5221319a1b28fc", "submitted_at": "2025-10-05T05:15:45Z", "updated_at": "2025-10-05T05:15:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.04039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2510.04039"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04039"}], "name": "\\textsc{GUI-Spotlight}: Adaptive Iterative Focus Refinement for Enhanced GUI Visual Grounding", "publication": null, "published_at": "2025-10-05", "record_sha256": "9d1d0421ced06da2f8a41372d2f5d569992bda996eb8ef653c7d5a1ea9c97c6e", "relations": [], "resource_id": "arxiv:2510.04039", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.04039v1", "topics": ["grounding", "planning", "tool-integration"], "updated_at": "2025-10-05"}, "text": "\\textsc{GUI-Spotlight}: Adaptive Iterative Focus Refinement for Enhanced GUI Visual Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Spotlight trains a grounding model to reason over images and call tools that iteratively narrow its visual focus. The paper evaluates this refinement method on ScreenSpot-Pro after training on 18,500 samples.\nPlatforms: unspecified\nTopics: grounding, planning, tool-integration\nPublished: 2025-10-05\nAuthors: Bin Lei, Nuo Xu, Ali Payani, Mingyi Hong, Chunhua Liao, Yu Cao, Caiwen Ding\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.04039v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-04039/"}
{"id": "arxiv:2510.04373", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d4bd517ada9303e765827b3badfa2ee432bc5511ff4ee0cc42d937735e1c8de", "submitted_at": "2025-10-05T21:34:42Z", "updated_at": "2026-02-20T19:34:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04373v2"}, "authors": ["Hadi Nekoei", "Aman Jaiswal", "Patrice Bechard", "Oleh Shliazhko", "Orlando Marquez Ayala", "Mathieu Reymond", "Massimo Caccia", "Alexandre Drouin", "Sarath Chandar", "Alexandre Lacoste"], "canonical_source_url": "https://arxiv.org/abs/2510.04373", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-04373/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d4bd517ada9303e765827b3badfa2ee432bc5511ff4ee0cc42d937735e1c8de", "submitted_at": "2025-10-05T21:34:42Z", "updated_at": "2026-02-20T19:34:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04373v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.04373"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04373"}], "name": "JEF-Hinter: Leveraging Offline Knowledge for Improving Web Agents Adaptation", "publication": null, "published_at": "2025-10-05", "record_sha256": "6eb137daaa7969664ecf3752a597a745549bfd35194e3e425c5016882ee74052", "relations": [], "resource_id": "arxiv:2510.04373", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.04373v2", "topics": ["memory", "interaction-history", "web-navigation"], "updated_at": "2026-02-20"}, "text": "JEF-Hinter: Leveraging Offline Knowledge for Improving Web Agents Adaptation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces JEF-Hinter, which distills successful and failed trajectories into compact episodic hints and retrieves them for the current agent state. Web benchmark experiments examine guidance from decisive steps rather than full demonstration replay.\nPlatforms: web\nTopics: memory, interaction-history, web-navigation\nPublished: 2025-10-05\nAuthors: Hadi Nekoei, Aman Jaiswal, Patrice Bechard, Oleh Shliazhko, Orlando Marquez Ayala, Mathieu Reymond, Massimo Caccia, Alexandre Drouin, Sarath Chandar, Alexandre Lacoste\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.04373v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-04373/"}
{"id": "arxiv:2510.04607", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c22547bf82eb2fb87e8bf6a3846d25c17cac296763726ed9a25faf88bdca33c", "submitted_at": "2025-10-06T09:14:58Z", "updated_at": "2026-03-25T11:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04607v2"}, "authors": ["Yuan Wang", "Mingyu Li", "Haibo Chen"], "canonical_source_url": "https://arxiv.org/abs/2510.04607", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-04607/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c22547bf82eb2fb87e8bf6a3846d25c17cac296763726ed9a25faf88bdca33c", "submitted_at": "2025-10-06T09:14:58Z", "updated_at": "2026-03-25T11:29:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.04607v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.04607"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.04607"}], "name": "From Imperative to Declarative: Towards LLM-friendly OS Interfaces for Boosted Computer-Use Agents", "publication": null, "published_at": "2025-10-06", "record_sha256": "77e8b50af4afc925ab9c98ccce14171aa455c2d0d854d5374a146948e133c23b", "relations": [], "resource_id": "arxiv:2510.04607", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.04607v2", "topics": ["automation", "tool-integration", "planning"], "updated_at": "2026-03-25"}, "text": "From Imperative to Declarative: Towards LLM-friendly OS Interfaces for Boosted Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a Declarative Model Interface that maps existing GUIs to access, state and observation primitives while handling low-level interactions. Evaluation uses Word, PowerPoint and Excel on Windows with a GUI-agent baseline.\nPlatforms: desktop, windows\nTopics: automation, tool-integration, planning\nPublished: 2025-10-06\nAuthors: Yuan Wang, Mingyu Li, Haibo Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.04607v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-04607/"}
{"id": "arxiv:2510.06042", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bf0a561460c080b14da6fd13f6d92781e01e52b08dc691fe019e70b88e44c86", "submitted_at": "2025-10-07T15:36:04Z", "updated_at": "2026-09-27T17:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.06042v3"}, "authors": ["Shang Ma", "Xusheng Xiao", "Yanfang Ye"], "canonical_source_url": "https://arxiv.org/abs/2510.06042", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-06042/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bf0a561460c080b14da6fd13f6d92781e01e52b08dc691fe019e70b88e44c86", "submitted_at": "2025-10-07T15:36:04Z", "updated_at": "2026-09-27T17:25:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.06042v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:10.946679+00:00", "role": "primary", "sha256": "594b5261dd2039777d36c46cdb8525d973ad57a154b902ff1a171bcbed2c64cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:29.861666+00:00", "role": "primary", "sha256": "5e2cb2f053983140904ec15c097d060ff18315c6c3c956bbcacbec26921709c5", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2510.06042v3"}], "identifiers": {"arxiv": "2510.06042"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:10.946679+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.06042"}], "name": "Agent+P: Guiding UI Agents via Symbolic Planning", "publication": null, "published_at": "2025-10-07", "record_sha256": "78d77b88bd896524308d92e64c3eb37dd5f6dccd040eb19b7a6001d16f521658", "relations": [], "resource_id": "arxiv:2510.06042", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.06042v3", "topics": ["planning", "long-horizon", "mobile-navigation"], "updated_at": "2026-09-27"}, "text": "Agent+P: Guiding UI Agents via Symbolic Planning\nResource type: paper\nReview status: curated\nScope: core\n\nGuides UI agents with symbolic planning over a UI transition graph initialized from app-code analysis and updated through online exploration, with evaluation on AndroidWorld.\nPlatforms: android\nTopics: planning, long-horizon, mobile-navigation\nPublished: 2025-10-07\nAuthors: Shang Ma, Xusheng Xiao, Yanfang Ye\nRecorded arXiv version: v3\n\nPlanning guarantees are relative to the modeled graph and selected target. They do not establish complete graphs, correct grounding or faithful goal translation. The described design depends on static source analysis; no experiments were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.3 UI Automation to Classical Planning; 4.1 UTG Builder; 4.2 Node Selector; 4.3 Plan Generator; 4.4 UI Explorer; 5.1 Datasets); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2510.06042v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-06042/"}
{"id": "arxiv:2510.07809", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e80999148931b7485d51ef87887d06240f813747a9739282ed0daf53a912f9f6", "submitted_at": "2025-10-09T05:34:57Z", "updated_at": "2026-04-09T01:41:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2510.07809v4"}, "authors": ["Renhua Ding", "Xiao Yang", "Zhengwei Fang", "Jun Luo", "Kun He", "Jun Zhu"], "canonical_source_url": "https://arxiv.org/abs/2510.07809", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-07809/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e80999148931b7485d51ef87887d06240f813747a9739282ed0daf53a912f9f6", "submitted_at": "2025-10-09T05:34:57Z", "updated_at": "2026-04-09T01:41:48Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2510.07809v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2510.07809"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.07809"}], "name": "Invisible to Humans, Triggered by Agents: Stealthy Jailbreak Attacks on Mobile Vision-Language Agents", "publication": null, "published_at": "2025-10-09", "record_sha256": "dd51c65fb1ea57916c7d271c87592acee81ccc4bb1fa77edd2bb11abfdf9e870", "relations": [], "resource_id": "arxiv:2510.07809", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.07809v4", "topics": ["safety-security", "mobile-navigation", "robustness"], "updated_at": "2026-04-09"}, "text": "Invisible to Humans, Triggered by Agents: Stealthy Jailbreak Attacks on Mobile Vision-Language Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses the near-zero contact-touch signals of automated mobile agents to expose malicious visual content selectively during agent interactions. The paper combines this agent-only perceptual injection with HG-IDA* prompt optimization to study unauthorized cross-app actions.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, robustness\nPublished: 2025-10-09\nAuthors: Renhua Ding, Xiao Yang, Zhengwei Fang, Jun Luo, Kun He, Jun Zhu\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.07809v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-07809/"}
{"id": "arxiv:2510.07988", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9460a0926936d3db23c36c6c0e069d664265a8a9513c08c1a04f5a49970d6567", "submitted_at": "2025-10-09T09:22:05Z", "updated_at": "2025-10-09T09:22:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.07988v1"}, "authors": ["Haitao Jia", "Ming He", "Zimo Yin", "Likang Wu", "Jianping Fan", "Jitao Sang"], "canonical_source_url": "https://arxiv.org/abs/2510.07988", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-07988/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9460a0926936d3db23c36c6c0e069d664265a8a9513c08c1a04f5a49970d6567", "submitted_at": "2025-10-09T09:22:05Z", "updated_at": "2025-10-09T09:22:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.07988v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.07988"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.07988"}], "name": "ReInAgent: A Context-Aware GUI Agent Enabling Human-in-the-Loop Mobile Task Navigation", "publication": null, "published_at": "2025-10-09", "record_sha256": "b54a4f19baeec627a4c515599edd03ba6d640cd508ddbc815b6ba6fde47fa8fa", "relations": [], "resource_id": "arxiv:2510.07988", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.07988v1", "topics": ["human-in-the-loop", "memory", "mobile-navigation"], "updated_at": "2025-10-09"}, "text": "ReInAgent: A Context-Aware GUI Agent Enabling Human-in-the-Loop Mobile Task Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ReInAgent, coordinating information management, planning, and reflection around a shared memory for mobile tasks. It proactively asks users for information and checks consistency when instructions are ambiguous or change during execution.\nPlatforms: unspecified\nTopics: human-in-the-loop, memory, mobile-navigation\nPublished: 2025-10-09\nAuthors: Haitao Jia, Ming He, Zimo Yin, Likang Wu, Jianping Fan, Jitao Sang\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.07988v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-07988/"}
{"id": "arxiv:2510.09038", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c41dfbef8b76bbdccd7c435694cee8788b116db278122f64a22e7498bd5a5a", "submitted_at": "2025-10-10T06:16:45Z", "updated_at": "2025-10-10T06:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.09038v1"}, "authors": ["Wenyi Wu", "Kun Zhou", "Ruoxin Yuan", "Vivian Yu", "Stephen Wang", "Zhiting Hu", "Biwei Huang"], "canonical_source_url": "https://arxiv.org/abs/2510.09038", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-09038/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c41dfbef8b76bbdccd7c435694cee8788b116db278122f64a22e7498bd5a5a", "submitted_at": "2025-10-10T06:16:45Z", "updated_at": "2025-10-10T06:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.09038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.09038"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.09038"}], "name": "Auto-scaling Continuous Memory for GUI Agent", "publication": null, "published_at": "2025-10-10", "record_sha256": "843f527fae6f9d6a8df57ee0463b10e7e0086c92babd98bce148042780784758", "relations": [], "resource_id": "arxiv:2510.09038", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.09038v1", "topics": ["memory", "interaction-history", "efficiency"], "updated_at": "2025-10-10"}, "text": "Auto-scaling Continuous Memory for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nEncodes GUI trajectories into compact continuous embeddings that can be retrieved directly into an agent’s input. An automated environment-discovery and rollout pipeline expands this memory for long tasks and unfamiliar interfaces.\nPlatforms: unspecified\nTopics: memory, interaction-history, efficiency\nPublished: 2025-10-10\nAuthors: Wenyi Wu, Kun Zhou, Ruoxin Yuan, Vivian Yu, Stephen Wang, Zhiting Hu, Biwei Huang\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.09038v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-09038/"}
{"id": "arxiv:2510.09872", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af3fb75d012ebcf4ad082ab7b83b0553596bbbb8598751d0a70c86d929a22ee", "submitted_at": "2025-10-10T21:20:51Z", "updated_at": "2026-05-18T22:06:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.09872v2"}, "authors": ["Sanjari Srivastava", "Gang Li", "Cheng Chang", "Rishu Garg", "Manpreet Kaur", "Charlene Y. Lee", "Yuezhang Li", "Yining Mao", "Ignacio Cases", "Yanan Xie", "Peng Qi"], "canonical_source_url": "https://arxiv.org/abs/2510.09872", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-09872/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3af3fb75d012ebcf4ad082ab7b83b0553596bbbb8598751d0a70c86d929a22ee", "submitted_at": "2025-10-10T21:20:51Z", "updated_at": "2026-05-18T22:06:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.09872v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2510.09872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.09872"}], "name": "WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions", "publication": null, "published_at": "2025-10-10", "record_sha256": "c711bfa8b681930af8fbd4ed6791138e7a8438cbdb1286cf0d188c7a9436dbd4", "relations": [], "resource_id": "arxiv:2510.09872", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.09872v2", "topics": ["evaluation", "web-navigation", "sandbox", "reinforcement-learning"], "updated_at": "2026-05-18"}, "text": "WARC-Bench: Web Archive Based Benchmark for GUI Subtask Executions\nResource type: paper\nReview status: curated\nScope: core\n\nWARC-Bench packages dynamic web pages as replayable archives and evaluates short navigation subtasks involving components such as date pickers and scrollable containers. It also studies supervised training followed by reinforcement learning with verifiable rewards on these browser interactions.\nPlatforms: web\nTopics: evaluation, web-navigation, sandbox, reinforcement-learning\nPublished: 2025-10-10\nAuthors: Sanjari Srivastava, Gang Li, Cheng Chang, Rishu Garg, Manpreet Kaur, Charlene Y. Lee, Yuezhang Li, Yining Mao, Ignacio Cases, Yanan Xie, Peng Qi\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.09872v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-09872/"}
{"id": "arxiv:2510.10073", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c9f7c91539df587a4f08888820731aef5bb3fca2cfb9e742443330aaf1c35f", "submitted_at": "2025-10-11T07:18:12Z", "updated_at": "2026-04-14T02:53:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10073v2"}, "authors": ["Zonghao Ying", "Yangguang Shao", "Jianle Gan", "Gan Xu", "Wenxin Zhang", "Quanchen Zou", "Junzheng Shi", "Zhenfei Yin", "Mingchuan Zhang", "Aishan Liu", "Xianglong Liu"], "canonical_source_url": "https://arxiv.org/abs/2510.10073", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-10073/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43c9f7c91539df587a4f08888820731aef5bb3fca2cfb9e742443330aaf1c35f", "submitted_at": "2025-10-11T07:18:12Z", "updated_at": "2026-04-14T02:53:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10073v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.10073"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.10073"}], "name": "SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents", "publication": null, "published_at": "2025-10-11", "record_sha256": "86c53e7dcbb0a8ab29a4573d9472ac6f73481ceb3c7230da31a172d77307d3cb", "relations": [], "resource_id": "arxiv:2510.10073", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.10073v2", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-04-14"}, "text": "SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SecureWebArena with simulated interactive websites, an attack-vector taxonomy and labeled agent trajectories. Its evaluation separates failures in reasoning, behavioral traces and task outcomes across vision-language agent families.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2025-10-11\nAuthors: Zonghao Ying, Yangguang Shao, Jianle Gan, Gan Xu, Wenxin Zhang, Quanchen Zou, Junzheng Shi, Zhenfei Yin, Mingchuan Zhang, Aishan Liu, Xianglong Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.10073v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-10073/"}
{"id": "arxiv:2510.10666", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe982af6e7285c21e13dcd4565255b501b401b188da64069d9cfe2fab7891f88", "submitted_at": "2025-10-12T15:43:37Z", "updated_at": "2025-10-14T08:54:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10666v2"}, "authors": ["Tao Yu", "Zhengbo Zhang", "Zhiheng Lyu", "Junhao Gong", "Hongzhu Yi", "Xinming Wang", "Yuxuan Zhou", "Jiabing Yang", "Ping Nie", "Yan Huang", "Wenhu Chen"], "canonical_source_url": "https://arxiv.org/abs/2510.10666", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-10666/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe982af6e7285c21e13dcd4565255b501b401b188da64069d9cfe2fab7891f88", "submitted_at": "2025-10-12T15:43:37Z", "updated_at": "2025-10-14T08:54:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.10666v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.10666"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.10666"}], "name": "BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions", "publication": null, "published_at": "2025-10-12", "record_sha256": "9e23d4a16c7bb5554e518e5189b8e2239bcb3ab2bcbe50a1638b6d08d537fdbc", "relations": [], "resource_id": "arxiv:2510.10666", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.10666v2", "topics": ["web-navigation", "imitation-learning", "memory"], "updated_at": "2025-10-14"}, "text": "BrowserAgent: Building Web Agents with Human-Inspired Web Browsing Actions\nResource type: paper\nReview status: curated\nScope: core\n\nTrains BrowserAgent to use Playwright-based scrolling, clicking and typing directly on webpages for information-seeking tasks. Supervised and rejection fine-tuning are paired with explicit memory for conclusions carried across steps.\nPlatforms: web\nTopics: web-navigation, imitation-learning, memory\nPublished: 2025-10-12\nAuthors: Tao Yu, Zhengbo Zhang, Zhiheng Lyu, Junhao Gong, Hongzhu Yi, Xinming Wang, Yuxuan Zhou, Jiabing Yang, Ping Nie, Yan Huang, Wenhu Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.10666v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-10666/"}
{"id": "arxiv:2510.11035", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6152a61f81c435e76a432bc10086827c0cf0d2f44026bac03d6b80ebae8ca77", "submitted_at": "2025-10-13T06:12:03Z", "updated_at": "2026-02-23T08:31:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11035v2"}, "authors": ["Longjie Guo", "Chenjie Yuan", "Mingyuan Zhong", "Robert Wolfe", "Ruican Zhong", "Yue Xu", "Bingbing Wen", "Hua Shen", "Lucy Lu Wang", "Alexis Hiniker"], "canonical_source_url": "https://arxiv.org/abs/2510.11035", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-11035/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6152a61f81c435e76a432bc10086827c0cf0d2f44026bac03d6b80ebae8ca77", "submitted_at": "2025-10-13T06:12:03Z", "updated_at": "2026-02-23T08:31:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11035v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.11035"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.11035"}], "name": "SusBench: An Online Benchmark for Evaluating Dark Pattern Susceptibility of Computer-Use Agents", "publication": null, "published_at": "2025-10-13", "record_sha256": "485894554bbb537aba39683c8d01bd59e644e07c965fc03c431d1aec2f3d4dd8", "relations": [], "resource_id": "arxiv:2510.11035", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.11035v2", "topics": ["ui-design", "safety-security", "evaluation"], "updated_at": "2026-02-23"}, "text": "SusBench: An Online Benchmark for Evaluating Dark Pattern Susceptibility of Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SusBench, an online benchmark that inserts dark-pattern designs into consumer websites for agent evaluation. A human study checks the realism of the interventions, and agent trials compare susceptibility across pattern types.\nPlatforms: web\nTopics: ui-design, safety-security, evaluation\nPublished: 2025-10-13\nAuthors: Longjie Guo, Chenjie Yuan, Mingyuan Zhong, Robert Wolfe, Ruican Zhong, Yue Xu, Bingbing Wen, Hua Shen, Lucy Lu Wang, Alexis Hiniker\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.11035v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-11035/"}
{"id": "arxiv:2510.11892", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c37dae372655e8ef556163744e47ed42db3533175ca7d1db7560033774ed2a0", "submitted_at": "2025-10-13T19:52:04Z", "updated_at": "2026-03-09T15:25:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11892v2"}, "authors": ["Kai Mei", "Jiang Guo", "Shuaichen Chang", "Mingwen Dong", "Dongkyu Lee", "Xing Niu", "Jiarong Jiang"], "canonical_source_url": "https://arxiv.org/abs/2510.11892", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-11892/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c37dae372655e8ef556163744e47ed42db3533175ca7d1db7560033774ed2a0", "submitted_at": "2025-10-13T19:52:04Z", "updated_at": "2026-03-09T15:25:16Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.11892v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.11892"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.11892"}], "name": "R-WoM: Retrieval-augmented World Model For Computer-use Agents", "publication": null, "published_at": "2025-10-13", "record_sha256": "6133db8184e65ce2ff63538215321dfe45a70359afb323c97f49055ef2ed6613", "relations": [], "resource_id": "arxiv:2510.11892", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.11892v2", "topics": ["world-models", "memory", "planning", "evaluation"], "updated_at": "2026-03-09"}, "text": "R-WoM: Retrieval-augmented World Model For Computer-use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nR-WoM evaluates how well world models predict nearby GUI states, align complete procedures, and recognize milestone transitions. It retrieves external tutorials to guide simulated future interactions, with experiments on subsets of OSWorld and WebArena.\nPlatforms: desktop, web\nTopics: world-models, memory, planning, evaluation\nPublished: 2025-10-13\nAuthors: Kai Mei, Jiang Guo, Shuaichen Chang, Mingwen Dong, Dongkyu Lee, Xing Niu, Jiarong Jiang\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.11892v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-11892/"}
{"id": "arxiv:2510.12200", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92db1e568c3c512aa00e2363bbdc9869cf834d638c331b45d4753d29a282004", "submitted_at": "2025-10-14T06:52:15Z", "updated_at": "2025-10-14T06:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.12200v1"}, "authors": ["Xiaoxue Ren", "Penghao Jiang", "Kaixin Li", "Zhiyong Huang", "Xiaoning Du", "Jiaojiao Jiang", "Zhenchang Xing", "Jiamou Sun", "Terry Yue Zhuo"], "canonical_source_url": "https://arxiv.org/abs/2510.12200", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-12200/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92db1e568c3c512aa00e2363bbdc9869cf834d638c331b45d4753d29a282004", "submitted_at": "2025-10-14T06:52:15Z", "updated_at": "2025-10-14T06:52:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.12200v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.12200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.12200"}], "name": "HackWorld: Evaluating Computer-Use Agents on Exploiting Web Application Vulnerabilities", "publication": null, "published_at": "2025-10-14", "record_sha256": "8cf2403357efc7f15799a39571685454bc8b3d6f468a5bfa3ddaae6ba7394875", "relations": [], "resource_id": "arxiv:2510.12200", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.12200v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-10-14"}, "text": "HackWorld: Evaluating Computer-Use Agents on Exploiting Web Application Vulnerabilities\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces HackWorld for evaluating visual computer-use agents in a capture-the-flag collection of web applications. The benchmark examines interface navigation and multistep planning during vulnerability discovery and exploitation tasks.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2025-10-14\nAuthors: Xiaoxue Ren, Penghao Jiang, Kaixin Li, Zhiyong Huang, Xiaoning Du, Jiaojiao Jiang, Zhenchang Xing, Jiamou Sun, Terry Yue Zhuo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.12200v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-12200/"}
{"id": "arxiv:2510.13543", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bedde658a28c57551c27563fa22cb037839dddd4e5eabe1c66afa27523d0d406", "submitted_at": "2025-10-15T13:39:13Z", "updated_at": "2025-10-15T13:39:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.13543v1"}, "authors": ["Avihay Cohen"], "canonical_source_url": "https://arxiv.org/abs/2510.13543", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-13543/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bedde658a28c57551c27563fa22cb037839dddd4e5eabe1c66afa27523d0d406", "submitted_at": "2025-10-15T13:39:13Z", "updated_at": "2025-10-15T13:39:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.13543v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2510.13543"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.13543"}], "name": "In-Browser LLM-Guided Fuzzing for Real-Time Prompt Injection Testing in Agentic AI Browsers", "publication": null, "published_at": "2025-10-15", "record_sha256": "2a7c155b1537eb78a574bfa6103b2e5132e6e78af82d8c4115f540c0111857db", "relations": [], "resource_id": "arxiv:2510.13543", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.13543v1", "topics": ["safety-security", "automation"], "updated_at": "2025-10-15"}, "text": "In-Browser LLM-Guided Fuzzing for Real-Time Prompt Injection Testing in Agentic AI Browsers\nResource type: paper\nReview status: curated\nScope: core\n\nProposes an in-browser fuzzing framework that uses language-model guidance to test indirect prompt-injection vulnerabilities in agentic browsers. The work focuses on unwanted actions triggered by webpage instructions under the browser agent's user privileges.\nPlatforms: web\nTopics: safety-security, automation\nPublished: 2025-10-15\nAuthors: Avihay Cohen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The short abstract supplies the framework and threat setting but no quantitative evaluation details.\n\nPrimary source: https://arxiv.org/abs/2510.13543v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-13543/"}
{"id": "arxiv:2510.14388", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95fd3e1286ed49d569b05644d77f822ea9747cfb4769cb7cb93f39c9271cf57e", "submitted_at": "2025-10-16T07:38:21Z", "updated_at": "2026-02-06T02:04:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.14388v2"}, "authors": ["Zhe Wu", "Hongjin Lu", "Junliang Xing", "Changhao Zhang", "Yuxuan Li", "Yin Zhu", "Yuhao Yang", "Yuheng Jing", "Kai Li", "Kun Shao", "Jianye Hao", "Jun Wang", "Yuanchun Shi"], "canonical_source_url": "https://arxiv.org/abs/2510.14388", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-14388/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95fd3e1286ed49d569b05644d77f822ea9747cfb4769cb7cb93f39c9271cf57e", "submitted_at": "2025-10-16T07:38:21Z", "updated_at": "2026-02-06T02:04:25Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.14388v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T07:57:59Z", "role": "primary", "sha256": "240e3d6528900f4a3e87f12f5e55cb7654a4d25d4478e291e9bbf725e7013118", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.10587%2C2510.10909%2C2510.11144%2C2510.12798%2C2510.14388%2C2510.14513%2C2510.14969%2C2510.16156%2C2510.17017%2C2510.17800%2C2510.19207%2C2510.19247%2C2510.19386%2C2510.19864%2C2510.20333%2C2510.21652%2C2510.22443%2C2510.22936%2C2510.23730%2C2510.23883%2C2510.24702%2C2510.25726%2C2510.26328%2C2510.27135%2C2510.27410%2C2511.00153%2C2511.02650%2C2511.02687%2C2511.03690%2C2511.04220%2C2511.05797%2C2511.06251%2C2511.08319%2C2511.08487%2C2511.10110%2C2511.13725%2C2511.14210%2C2511.14945%2C2511.15705%2C2511.16842%2C2511.17649%2C2511.17775%2C2511.19496%2C2511.19516%2C2511.20636%2C2511.21990%2C2511.22074%2C2512.04302%2C2512.04529%2C2512.06726&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:56Z", "role": "primary", "sha256": "fecd42f8ca75572410c354c892734cb8eefbd3fa622064f6bc6db0f535cf5607", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/abs/2510.14388v2"}], "identifiers": {"arxiv": "2510.14388"}, "kind": "paper", "last_verified_at": "2026-10-17T11:20:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.14388"}], "name": "Hi-Agent: Hierarchical Vision-Language Agents for Mobile Device Control", "publication": null, "published_at": "2025-10-16", "record_sha256": "3344fd12c3e6571278fc1332dca84c9b7eff17b559968f51b6252deb496ef18a", "relations": [], "resource_id": "arxiv:2510.14388", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.14388v2", "topics": ["planning", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-02-06"}, "text": "Hi-Agent: Hierarchical Vision-Language Agents for Mobile Device Control\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Hi-Agent, jointly training a high-level reasoning model and a low-level mobile action model. Execution feedback supplies a foresight advantage for optimizing subgoal decisions, with evaluations on Android control benchmarks.\nPlatforms: android\nTopics: planning, reinforcement-learning, mobile-navigation\nPublished: 2025-10-16\nAuthors: Zhe Wu, Hongjin Lu, Junliang Xing, Changhao Zhang, Yuxuan Li, Yin Zhu, Yuhao Yang, Yuheng Jing, Kai Li, Kun Shao, Jianye Hao, Jun Wang, Yuanchun Shi\nRecorded arXiv version: v2\n\nReviewed against the exact-version official arXiv abstract page; its complete title, ordered author list, visible version, and normalized abstract SHA match the retained snapshot. The original API response body was not locally available. Experimental results and artifact availability have not been independently reproduced. This Hi-Agent has a distinct author team and method from the earlier HiAgent working-memory paper; the shared project-name prefix is not treated as an identity match.\n\nPrimary source: https://arxiv.org/abs/2510.14388v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-14388/"}
{"id": "arxiv:2510.14621", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc45dc62068c971cd831fc3db64bd3d8995d4e5fd60e9c83953476d4fb24f7f3", "submitted_at": "2025-10-16T12:30:05Z", "updated_at": "2025-10-16T12:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.14621v1"}, "authors": ["Yuanyi Song", "Heyuan Huang", "Qiqiang Lin", "Yin Zhao", "Xiangmou Qu", "Jun Wang", "Xingyu Lou", "Weiwen Liu", "Zhuosheng Zhang", "Jun Wang", "Yong Yu", "Weinan Zhang", "Zhaoxiang Wang"], "canonical_source_url": "https://arxiv.org/abs/2510.14621", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-14621/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fc45dc62068c971cd831fc3db64bd3d8995d4e5fd60e9c83953476d4fb24f7f3", "submitted_at": "2025-10-16T12:30:05Z", "updated_at": "2025-10-16T12:30:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.14621v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2510.14621"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.14621"}], "name": "ColorBench: Benchmarking Mobile Agents with Graph-Structured Framework for Complex Long-Horizon Tasks", "publication": null, "published_at": "2025-10-16", "record_sha256": "0f1aec7008db05dfd76151d84cdb66e5d1e62b9246d423532a1db71702650c84", "relations": [], "resource_id": "arxiv:2510.14621", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.14621v1", "topics": ["evaluation", "mobile-navigation", "long-horizon", "state-tracking"], "updated_at": "2025-10-16"}, "text": "ColorBench: Benchmarking Mobile Agents with Graph-Structured Framework for Complex Long-Horizon Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nColorBench represents mobile tasks as graphs of states collected from real devices, providing a quasi-dynamic environment with multiple valid paths and error paths. It evaluates single-app and cross-app tasks using completion, subtask progress, and atomic diagnostics.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, long-horizon, state-tracking\nPublished: 2025-10-16\nAuthors: Yuanyi Song, Heyuan Huang, Qiqiang Lin, Yin Zhao, Xiangmou Qu, Jun Wang, Xingyu Lou, Weiwen Liu, Zhuosheng Zhang, Jun Wang, Yong Yu, Weinan Zhang, Zhaoxiang Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.14621v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-14621/"}
{"id": "arxiv:2510.15455", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8584797a1c12158b20da04c30e3a0f577c63f02f9e906a051579959de2cae480", "submitted_at": "2025-10-17T09:11:05Z", "updated_at": "2025-10-17T09:11:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.15455v1"}, "authors": ["Gucongcong Fan", "Chaoyue Niu", "Chengfei Lyu", "Fan Wu", "Guihai Chen"], "canonical_source_url": "https://arxiv.org/abs/2510.15455", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-15455/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8584797a1c12158b20da04c30e3a0f577c63f02f9e906a051579959de2cae480", "submitted_at": "2025-10-17T09:11:05Z", "updated_at": "2025-10-17T09:11:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.15455v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:52Z", "role": "primary", "sha256": "0b730323b2ce85c39aa7d534665901a480ff06540c73f1f80832fcc0735b8789", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2310.08588%2C2402.09205%2C2406.05720%2C2407.21075%2C2410.07166%2C2503.03743%2C2508.04037%2C2508.10833%2C2508.11360%2C2508.13024%2C2508.13186%2C2508.14104%2C2509.01619%2C2509.01720%2C2509.02322%2C2509.04198%2C2509.07764%2C2509.21823%2C2509.22256%2C2509.23263%2C2509.23866%2C2509.25047%2C2509.26539%2C2510.00413%2C2510.00536%2C2510.01524%2C2510.02204%2C2510.02250%2C2510.02418%2C2510.03204%2C2510.03612%2C2510.04039%2C2510.04373%2C2510.04607%2C2510.06223%2C2510.06607%2C2510.09872%2C2510.10666%2C2510.11035%2C2510.11221%2C2510.11892%2C2510.13543%2C2510.15455%2C2510.15863%2C2510.16252%2C2510.18113%2C2510.18314%2C2510.19205%2C2510.22732%2C2511.06101&max_results=50"}], "identifiers": {"arxiv": "2510.15455"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.15455"}], "name": "CORE: Reducing UI Exposure in Mobile Agents via Collaboration Between Cloud and Local LLMs", "publication": null, "published_at": "2025-10-17", "record_sha256": "df6aae799505a0c7364b9791505d7368607a5d58ad9ffd4cf4df9d542b08dac7", "relations": [], "resource_id": "arxiv:2510.15455", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.15455v1", "topics": ["safety-security", "mobile-navigation", "efficiency"], "updated_at": "2025-10-17"}, "text": "CORE: Reducing UI Exposure in Mobile Agents via Collaboration Between Cloud and Local LLMs\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CORE, which divides mobile UI hierarchies into semantic blocks and shares planning between local and cloud language models. Local ranking limits the screen information sent for cloud-side element selection, with a multi-round mechanism for correcting inadequate context.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, efficiency\nPublished: 2025-10-17\nAuthors: Gucongcong Fan, Chaoyue Niu, Chengfei Lyu, Fan Wu, Guihai Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. XML screen hierarchies do not by themselves establish Android deployment.\n\nPrimary source: https://arxiv.org/abs/2510.15455v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-15455/"}
{"id": "arxiv:2510.16051", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e4377fd6b510455afa1eb04c64e8e0b551e89d3b54ce31f241f283233985e1c", "submitted_at": "2025-10-16T19:03:45Z", "updated_at": "2026-03-24T14:39:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16051v2"}, "authors": ["Sofiya Garkot", "Maksym Shamrai", "Ivan Synytsia", "Mariya Hirna"], "canonical_source_url": "https://arxiv.org/abs/2510.16051", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-16051/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6e4377fd6b510455afa1eb04c64e8e0b551e89d3b54ce31f241f283233985e1c", "submitted_at": "2025-10-16T19:03:45Z", "updated_at": "2026-03-24T14:39:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16051v2"}}, "environments": ["macos", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2510.16051"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.16051"}], "name": "GUIrilla: A Scalable Framework for Automated Desktop UI Exploration", "publication": null, "published_at": "2025-10-16", "record_sha256": "1fa5b6e7908f66bbe7eef92605d5f674db6acf232e10d043952513f14fa197c1", "relations": [], "resource_id": "arxiv:2510.16051", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.16051v2", "topics": ["data-generation", "screen-understanding", "recording"], "updated_at": "2026-03-24"}, "text": "GUIrilla: A Scalable Framework for Automated Desktop UI Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nGUIrilla systematically explores macOS interfaces to collect interaction traces and accessibility metadata for downstream GUI research. It organizes these observations into hierarchical MacApp Trees and describes an accessibility-driven collection library.\nPlatforms: macos, desktop\nTopics: data-generation, screen-understanding, recording\nPublished: 2025-10-16\nAuthors: Sofiya Garkot, Maksym Shamrai, Ivan Synytsia, Mariya Hirna\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.16051v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-16051/"}
{"id": "arxiv:2510.16252", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9a1f8ebdf70ff2fa7b0c807160cd583c0f017a536457ff5915e589fca5fa751", "submitted_at": "2025-10-17T22:54:33Z", "updated_at": "2026-05-17T17:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16252v2"}, "authors": ["Yuxuan Lu", "Ziyi Wang", "Jing Huang", "Hui Liu", "Jiri Gesi", "Yan Han", "Shihan Fu", "Tianqi Zheng", "Xianfeng Tang", "Chen Luo", "Yisi Sang", "Jin Lai", "Dakuo Wang"], "canonical_source_url": "https://arxiv.org/abs/2510.16252", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-16252/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9a1f8ebdf70ff2fa7b0c807160cd583c0f017a536457ff5915e589fca5fa751", "submitted_at": "2025-10-17T22:54:33Z", "updated_at": "2026-05-17T17:26:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.16252v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.16252"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.16252"}], "name": "WEBSERV: A Full-Stack and RL-Ready Web Environment for Training Web Agents at Scale", "publication": null, "published_at": "2025-10-17", "record_sha256": "4cd1acbbf31f14fe8b41c942604ecef0dfb429bb32ded0704d1e0e14fe6d1a30", "relations": [], "resource_id": "arxiv:2510.16252", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.16252v2", "topics": ["sandbox", "reinforcement-learning", "efficiency", "web-navigation"], "updated_at": "2026-05-17"}, "text": "WEBSERV: A Full-Stack and RL-Ready Web Environment for Training Web Agents at Scale\nResource type: paper\nReview status: curated\nScope: core\n\nWebServ combines copy-on-write container environments with a compact DOM-derived browser interface for training web agents. Network-aware waiting supports action execution in single-page applications, with evaluations and reinforcement-learning experiments on WebArena-Lite.\nPlatforms: web\nTopics: sandbox, reinforcement-learning, efficiency, web-navigation\nPublished: 2025-10-17\nAuthors: Yuxuan Lu, Ziyi Wang, Jing Huang, Hui Liu, Jiri Gesi, Yan Han, Shihan Fu, Tianqi Zheng, Xianfeng Tang, Chen Luo, Yisi Sang, Jin Lai, Dakuo Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.16252v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-16252/"}
{"id": "arxiv:2510.17790", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33569f6ccca9c7e2675796191124424e76636828179c300d35da01ad9ccef1ce", "submitted_at": "2025-10-20T17:48:26Z", "updated_at": "2026-05-26T11:29:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.17790v3"}, "authors": ["Yuhao Yang", "Zhen Yang", "Zi-Yi Dou", "Anh Nguyen", "Keen You", "Omar Attia", "Andrew Szot", "Michael Feng", "Ram Ramrakhya", "Alexander Toshev", "Chao Huang", "Yinfei Yang", "Zhe Gan"], "canonical_source_url": "https://arxiv.org/abs/2510.17790", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-17790/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33569f6ccca9c7e2675796191124424e76636828179c300d35da01ad9ccef1ce", "submitted_at": "2025-10-20T17:48:26Z", "updated_at": "2026-05-26T11:29:37Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.17790v3"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.17790"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.17790"}], "name": "UltraCUA: A Foundation Model for Computer Use Agents with Hybrid Action", "publication": null, "published_at": "2025-10-20", "record_sha256": "734bef5566aec1f7a34f25a66fae13a72d37af491d2ea2ee4d8ca840cdf8c217", "relations": [], "resource_id": "arxiv:2510.17790", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.17790v3", "topics": ["tool-integration", "reinforcement-learning", "data-generation"], "updated_at": "2026-05-26"}, "text": "UltraCUA: A Foundation Model for Computer Use Agents with Hybrid Action\nResource type: paper\nReview status: curated\nScope: core\n\nUltraCUA trains a shared policy for primitive GUI actions and high-level tool calls using synthesized tasks and hybrid trajectories. Supervised fine-tuning followed by online reinforcement learning is evaluated on OSWorld and WindowsAgentArena.\nPlatforms: desktop, windows\nTopics: tool-integration, reinforcement-learning, data-generation\nPublished: 2025-10-20\nAuthors: Yuhao Yang, Zhen Yang, Zi-Yi Dou, Anh Nguyen, Keen You, Omar Attia, Andrew Szot, Michael Feng, Ram Ramrakhya, Alexander Toshev, Chao Huang, Yinfei Yang, Zhe Gan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.17790v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-17790/"}
{"id": "arxiv:2510.18113", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db3c25ba55221ca3d7c45ce4243a4a37b56a698909c0ea3eaaabdcd37388a1ec", "submitted_at": "2025-10-20T21:26:26Z", "updated_at": "2025-10-20T21:26:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18113v1"}, "authors": ["Devin Ersoy", "Brandon Lee", "Ananth Shreekumar", "Arjun Arunasalam", "Muhammad Ibrahim", "Antonio Bianchi", "Z. Berkay Celik"], "canonical_source_url": "https://arxiv.org/abs/2510.18113", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-18113/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db3c25ba55221ca3d7c45ce4243a4a37b56a698909c0ea3eaaabdcd37388a1ec", "submitted_at": "2025-10-20T21:26:26Z", "updated_at": "2025-10-20T21:26:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18113v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.18113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.18113"}], "name": "Investigating the Impact of Dark Patterns on LLM-Based Web Agents", "publication": null, "published_at": "2025-10-20", "record_sha256": "27d2cb7fa871c35593bb795ba005c7c05bdcd0c1114e2db49d59b87c3c10461a", "relations": [], "resource_id": "arxiv:2510.18113", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.18113v1", "topics": ["safety-security", "ui-design", "evaluation"], "updated_at": "2025-10-20"}, "text": "Investigating the Impact of Dark Patterns on LLM-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents LiteAgent for recording web-agent interactions and TrickyArena for enabling or disabling dark patterns in controlled websites. Experiments compare how individual patterns, combined patterns and visual or HTML changes affect agent decisions.\nPlatforms: web\nTopics: safety-security, ui-design, evaluation\nPublished: 2025-10-20\nAuthors: Devin Ersoy, Brandon Lee, Ananth Shreekumar, Arjun Arunasalam, Muhammad Ibrahim, Antonio Bianchi, Z. Berkay Celik\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.18113v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-18113/"}
{"id": "arxiv:2510.18488", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33b21a63d51148f2b5c30478c302e3df9b92e03e2e106a1c68224f2c835a8608", "submitted_at": "2025-10-21T10:11:33Z", "updated_at": "2025-10-21T10:11:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18488v1"}, "authors": ["Ho Fai Leung", "Xiaoyan Xi", "Fei Zuo"], "canonical_source_url": "https://arxiv.org/abs/2510.18488", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-18488/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33b21a63d51148f2b5c30478c302e3df9b92e03e2e106a1c68224f2c835a8608", "submitted_at": "2025-10-21T10:11:33Z", "updated_at": "2025-10-21T10:11:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.18488v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.18488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.18488"}], "name": "AndroidControl-Curated: Revealing the True Potential of GUI Agents through Benchmark Purification", "publication": null, "published_at": "2025-10-21", "record_sha256": "4bfca229ede1e97a38c66c77ba01ad1b1741cb3f06c426204ecd8f6c7aba3f54", "relations": [], "resource_id": "arxiv:2510.18488", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.18488v1", "topics": ["evaluation", "data-generation", "grounding"], "updated_at": "2025-10-21"}, "text": "AndroidControl-Curated: Revealing the True Potential of GUI Agents through Benchmark Purification\nResource type: paper\nReview status: curated\nScope: core\n\nAndroidControl-Curated audits ambiguities and factual errors in AndroidControl and produces a refined evaluation dataset. The paper also studies post-training a small Magma-R1 model on a curated subset to examine the effect of data quality on GUI action assessment.\nPlatforms: android\nTopics: evaluation, data-generation, grounding\nPublished: 2025-10-21\nAuthors: Ho Fai Leung, Xiaoyan Xi, Fei Zuo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.18488v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-18488/"}
{"id": "arxiv:2510.19205", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d404ff922ef38e57481c01a5e3108bd38457d88f83d7adf6d3c3c7b5ddf124e2", "submitted_at": "2025-10-22T03:29:25Z", "updated_at": "2025-10-22T03:29:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19205v1"}, "authors": ["Yaoyao Qian", "Yuanli Wang", "Jinda Zhang", "Yun Zong", "Meixu Chen", "Hanhan Zhou", "Jindan Huang", "Yifan Zeng", "Xinyu Hu", "Chan Hee Song", "Danqing Zhang"], "canonical_source_url": "https://arxiv.org/abs/2510.19205", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-19205/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d404ff922ef38e57481c01a5e3108bd38457d88f83d7adf6d3c3c7b5ddf124e2", "submitted_at": "2025-10-22T03:29:25Z", "updated_at": "2025-10-22T03:29:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19205v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.19205"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19205"}], "name": "WebGraphEval: Multi-Turn Trajectory Evaluation for Web Agents using Graph Representation", "publication": null, "published_at": "2025-10-22", "record_sha256": "bed516cea6840c98271ef1749dc606c480c7ac8dd0c69d5d197762e828a79588", "relations": [], "resource_id": "arxiv:2510.19205", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.19205v1", "topics": ["evaluation", "observability", "interaction-history"], "updated_at": "2025-10-22"}, "text": "WebGraphEval: Multi-Turn Trajectory Evaluation for Web Agents using Graph Representation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebGraphEval, which merges multiple agents' trajectories into a weighted graph of canonical actions. Structural analyses examine recurring behavior, inefficient paths and important decision points beyond a final success score.\nPlatforms: web\nTopics: evaluation, observability, interaction-history\nPublished: 2025-10-22\nAuthors: Yaoyao Qian, Yuanli Wang, Jinda Zhang, Yun Zong, Meixu Chen, Hanhan Zhou, Jindan Huang, Yifan Zeng, Xinyu Hu, Chan Hee Song, Danqing Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.19205v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-19205/"}
{"id": "arxiv:2510.19488", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebb064e1a9c3e2255677d9b4cd9ce33d2a97dd26a379c321721d576e7c5d6399", "submitted_at": "2025-10-22T11:25:48Z", "updated_at": "2025-10-22T11:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19488v1"}, "authors": ["Dunjie Lu", "Yiheng Xu", "Junli Wang", "Haoyuan Wu", "Xinyuan Wang", "Zekun Wang", "Junlin Yang", "Hongjin Su", "Jixuan Chen", "Junda Chen", "Yuchen Mao", "Jingren Zhou", "Junyang Lin", "Binyuan Hui", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2510.19488", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-19488/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ebb064e1a9c3e2255677d9b4cd9ce33d2a97dd26a379c321721d576e7c5d6399", "submitted_at": "2025-10-22T11:25:48Z", "updated_at": "2025-10-22T11:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.19488v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.19488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19488"}], "name": "VideoAgentTrek: Computer Use Pretraining from Unlabeled Videos", "publication": null, "published_at": "2025-10-22", "record_sha256": "9caf00c76df46f3b40e228e9cbcd1b4394b0d1e8b47a5153de01a4c41ba15b74", "relations": [], "resource_id": "arxiv:2510.19488", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.19488v1", "topics": ["data-generation", "video"], "updated_at": "2025-10-22"}, "text": "VideoAgentTrek: Computer Use Pretraining from Unlabeled Videos\nResource type: paper\nReview status: curated\nScope: core\n\nExtracts structured GUI actions from screen-recorded tutorials using inverse dynamics, then uses the inferred trajectories for computer-use training.\nPlatforms: desktop, web\nTopics: data-generation, video\nPublished: 2025-10-22\nAuthors: Dunjie Lu, Yiheng Xu, Junli Wang, Haoyuan Wu, Xinyuan Wang, Zekun Wang, Junlin Yang, Hongjin Su, Jixuan Chen, Junda Chen, Yuchen Mao, Jingren Zhou, Junyang Lin, Binyuan Hui, Tao Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.19488v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-19488/"}
{"id": "arxiv:2510.19949", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de25494af1bc2ceaa760a21da51803aaf89d9d2d1961887cfd0023fb87f28f63", "submitted_at": "2025-10-22T18:21:52Z", "updated_at": "2025-10-24T11:52:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.19949v2"}, "authors": ["Mathieu Andreux", "Märt Bakler", "Yanael Barbier", "Hamza Benchekroun", "Emilien Biré", "Antoine Bonnet", "Riaz Bordie", "Nathan Bout", "Matthias Brunel", "Aleix Cambray", "Pierre-Louis Cedoz", "Antoine Chassang", "Gautier Cloix", "Ethan Connelly", "Alexandra Constantinou", "Ramzi De Coster", "Hubert de la Jonquiere", "Aurélien Delfosse", "Maxime Delpit", "Alexis Deprez", "Augustin Derupti", "Mathieu Diaz", "Shannon D'Souza", "Julie Dujardin", "Abai Edmund", "Michael Eickenberg", "Armand Fatalot", "Wissem Felissi", "Isaac Herring", "Xavier Koegler", "Erwan Le Jumeau de Kergaradec", "Aurélien Lac", "Maxime Langevin", "Corentin Lauverjat", "Antonio Loison", "Avshalom Manevich", "Axel Moyal", "Axel Nguyen Kerbel", "Marinela Parovic", "Julien Revelle", "Guillaume Richard", "Mats Richter", "Ronan Riochet", "María Santos", "Romain Savidan", "Laurent Sifre", "Maxime Theillard", "Marc Thibault", "Ivan Valentini", "Tony Wu", "Laura Yie", "Kai Yuan", "Jevgenij Zubovskij"], "canonical_source_url": "https://arxiv.org/abs/2510.19949", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-19949/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de25494af1bc2ceaa760a21da51803aaf89d9d2d1961887cfd0023fb87f28f63", "submitted_at": "2025-10-22T18:21:52Z", "updated_at": "2025-10-24T11:52:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.19949v2"}}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.19949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.19949"}], "name": "Surfer 2: The Next Generation of Cross-Platform Computer Use Agents", "publication": null, "published_at": "2025-10-22", "record_sha256": "7bcf331db74af2fe2590c7b9f4917038aac689443e326c8ab4464972486fbb0e", "relations": [], "resource_id": "arxiv:2510.19949", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.19949v2", "topics": ["planning", "memory", "long-horizon", "robustness"], "updated_at": "2025-10-24"}, "text": "Surfer 2: The Next Generation of Cross-Platform Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSurfer2 combines hierarchical context management with separate planning and execution, self-verification, and recovery for visual computer use. It evaluates the framework across web, desktop, and Android tasks without task-specific fine-tuning.\nPlatforms: web, desktop, android\nTopics: planning, memory, long-horizon, robustness\nPublished: 2025-10-22\nAuthors: Mathieu Andreux, Märt Bakler, Yanael Barbier, Hamza Benchekroun, Emilien Biré, Antoine Bonnet, Riaz Bordie, Nathan Bout, Matthias Brunel, Aleix Cambray, Pierre-Louis Cedoz, Antoine Chassang, Gautier Cloix, Ethan Connelly, Alexandra Constantinou, Ramzi De Coster, Hubert de la Jonquiere, Aurélien Delfosse, Maxime Delpit, Alexis Deprez, Augustin Derupti, Mathieu Diaz, Shannon D'Souza, Julie Dujardin, Abai Edmund, Michael Eickenberg, Armand Fatalot, Wissem Felissi, Isaac Herring, Xavier Koegler, Erwan Le Jumeau de Kergaradec, Aurélien Lac, Maxime Langevin, Corentin Lauverjat, Antonio Loison, Avshalom Manevich, Axel Moyal, Axel Nguyen Kerbel, Marinela Parovic, Julien Revelle, Guillaume Richard, Mats Richter, Ronan Riochet, María Santos, Romain Savidan, Laurent Sifre, Maxime Theillard, Marc Thibault, Ivan Valentini, Tony Wu, Laura Yie, Kai Yuan, Jevgenij Zubovskij\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.19949v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-19949/"}
{"id": "arxiv:2510.20286", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3378dd9187f8c8df983d97b10b2ca5b64b2e2332760a734a09db8ce34a398a0b", "submitted_at": "2025-10-23T07:18:32Z", "updated_at": "2025-10-23T07:18:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.20286v1"}, "authors": ["Liangyu Chen", "Hanzhang Zhou", "Chenglin Cai", "Jianan Zhang", "Panrong Tong", "Quyu Kong", "Xu Zhang", "Chen Liu", "Yuqi Liu", "Wenxuan Wang", "Yue Wang", "Qin Jin", "Steven Hoi"], "canonical_source_url": "https://arxiv.org/abs/2510.20286", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-20286/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3378dd9187f8c8df983d97b10b2ca5b64b2e2332760a734a09db8ce34a398a0b", "submitted_at": "2025-10-23T07:18:32Z", "updated_at": "2025-10-23T07:18:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.20286v1"}}, "environments": ["cross-platform", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.20286"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.20286"}], "name": "UI-Ins: Enhancing GUI Grounding with Multi-Perspective Instruction-as-Reasoning", "publication": null, "published_at": "2025-10-23", "record_sha256": "91dd6804687fa365cdd27b2a6f4b904708acf5b4be037d35e21b25293c46f1e2", "relations": [], "resource_id": "arxiv:2510.20286", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.20286v1", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2025-10-23"}, "text": "UI-Ins: Enhancing GUI Grounding with Multi-Perspective Instruction-as-Reasoning\nResource type: paper\nReview status: curated\nScope: core\n\nTreats alternative instruction formulations as reasoning paths and trains GUI-grounding models to select and combine useful perspectives.\nPlatforms: cross-platform, android\nTopics: grounding, reinforcement-learning\nPublished: 2025-10-23\nAuthors: Liangyu Chen, Hanzhang Zhou, Chenglin Cai, Jianan Zhang, Panrong Tong, Quyu Kong, Xu Zhang, Chen Liu, Yuqi Liu, Wenxuan Wang, Yue Wang, Qin Jin, Steven Hoi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.20286v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-20286/"}
{"id": "arxiv:2510.22009", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09bd73dff01641593f31f7d40effd60df42e3d9460ffa010f76c00a32be3593b", "submitted_at": "2025-10-24T20:23:12Z", "updated_at": "2026-10-05T01:17:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.22009v3"}, "authors": ["Yangqin Jiang", "Chao Huang"], "canonical_source_url": "https://arxiv.org/abs/2510.22009", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-22009/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09bd73dff01641593f31f7d40effd60df42e3d9460ffa010f76c00a32be3593b", "submitted_at": "2025-10-24T20:23:12Z", "updated_at": "2026-10-05T01:17:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.22009v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2510.22009"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.22009"}], "name": "OpenPhone: Mobile Agentic Foundation Models", "publication": null, "published_at": "2025-10-24", "record_sha256": "5dfed4b534cb0d10e84002fe8036ef42b050e59613502b775a0734e538a3456d", "relations": [], "resource_id": "arxiv:2510.22009", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.22009v3", "topics": ["reinforcement-learning", "memory", "efficiency", "mobile-navigation"], "updated_at": "2026-10-05"}, "text": "OpenPhone: Mobile Agentic Foundation Models\nResource type: paper\nReview status: curated\nScope: core\n\nOpenPhone trains a compact mobile policy with supervised learning followed by GRPO on synthetic GUI data, incorporating reasoning and memory management. Its device-cloud design handles routine work locally and delegates complex subtasks, with online evaluation on AndroidLab.\nPlatforms: android\nTopics: reinforcement-learning, memory, efficiency, mobile-navigation\nPublished: 2025-10-24\nAuthors: Yangqin Jiang, Chao Huang\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.22009v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-22009/"}
{"id": "arxiv:2510.24168", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d80312fbd8f57cdc53cc10e03f2c68e2152d06945e5089967dceda50c02b9d4b", "submitted_at": "2025-10-28T08:19:58Z", "updated_at": "2026-04-14T04:34:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24168v3"}, "authors": ["Weihua Cheng", "Junming Liu", "Yifei Sun", "Botian Shi", "Yirong Chen", "Ding Wang"], "canonical_source_url": "https://arxiv.org/abs/2510.24168", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-24168/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d80312fbd8f57cdc53cc10e03f2c68e2152d06945e5089967dceda50c02b9d4b", "submitted_at": "2025-10-28T08:19:58Z", "updated_at": "2026-04-14T04:34:20Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24168v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.24168"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24168"}], "name": "MGA: Memory-Driven GUI Agent for Observation-Centric Interaction", "publication": null, "published_at": "2025-10-28", "record_sha256": "31c309e39a663ec7a77b1accca3d5e46b1ff17491b3f1267ab88050a8fb56302", "relations": [], "resource_id": "arxiv:2510.24168", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.24168v3", "topics": ["memory", "state-tracking", "long-horizon"], "updated_at": "2026-04-14"}, "text": "MGA: Memory-Driven GUI Agent for Observation-Centric Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MGA, which separates task-agnostic screen-state reading from decision making and connects steps through structured state memory. Validated state deltas replace full trajectory replay, evaluated on OSWorld and application tasks.\nPlatforms: desktop\nTopics: memory, state-tracking, long-horizon\nPublished: 2025-10-28\nAuthors: Weihua Cheng, Junming Liu, Yifei Sun, Botian Shi, Yirong Chen, Ding Wang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.24168v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-24168/"}
{"id": "arxiv:2510.24411", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e3fde732c9e760437f2c60d585f7ba0d2ed13bbe9e051838844a21174159099", "submitted_at": "2025-10-28T13:22:39Z", "updated_at": "2026-07-26T01:16:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24411v3"}, "authors": ["Qiushi Sun", "Mukai Li", "Zhoumianze Liu", "Zhihui Xie", "Fangzhi Xu", "Zhangyue Yin", "Kanzhi Cheng", "Zehao Li", "Zichen Ding", "Qi Liu", "Zhiyong Wu", "Zhuosheng Zhang", "Ben Kao", "Lingpeng Kong"], "canonical_source_url": "https://arxiv.org/abs/2510.24411", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-24411/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e3fde732c9e760437f2c60d585f7ba0d2ed13bbe9e051838844a21174159099", "submitted_at": "2025-10-28T13:22:39Z", "updated_at": "2026-07-26T01:16:06Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.24411v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.24411"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24411"}], "name": "OS-Sentinel: Towards Safety-Enhanced Mobile GUI Agents via Hybrid Validation in Realistic Workflows", "publication": null, "published_at": "2025-10-28", "record_sha256": "92f89e2b07cc71772293baa4b65808dc6bc2d05f9cb29b52b98e3d5c92e8e884", "relations": [], "resource_id": "arxiv:2510.24411", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.24411v3", "topics": ["safety-security", "evaluation", "mobile-navigation"], "updated_at": "2026-07-26"}, "text": "OS-Sentinel: Towards Safety-Enhanced Mobile GUI Agents via Hybrid Validation in Realistic Workflows\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MobileRisk-Live and OS-Sentinel, combining a formal verifier for system-level violations with a multimodal judge for contextual risks in mobile GUI workflows.\nPlatforms: unspecified\nTopics: safety-security, evaluation, mobile-navigation\nPublished: 2025-10-28\nAuthors: Qiushi Sun, Mukai Li, Zhoumianze Liu, Zhihui Xie, Fangzhi Xu, Zhangyue Yin, Kanzhi Cheng, Zehao Li, Zichen Ding, Qi Liu, Zhiyong Wu, Zhuosheng Zhang, Ben Kao, Lingpeng Kong\nRecorded arXiv version: v3\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.24411v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-24411/"}
{"id": "arxiv:2510.24563", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d115356f723746ce95c3e40fed933786a0747b3b560c8f73a6b7af05ca22d24f", "submitted_at": "2025-10-28T15:56:36Z", "updated_at": "2025-11-11T04:10:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.24563v2"}, "authors": ["Hongrui Jia", "Jitong Liao", "Xi Zhang", "Haiyang Xu", "Tianbao Xie", "Chaoya Jiang", "Ming Yan", "Si Liu", "Wei Ye", "Fei Huang"], "canonical_source_url": "https://arxiv.org/abs/2510.24563", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-24563/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d115356f723746ce95c3e40fed933786a0747b3b560c8f73a6b7af05ca22d24f", "submitted_at": "2025-10-28T15:56:36Z", "updated_at": "2025-11-11T04:10:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.24563v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.24563"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.24563"}], "name": "OSWorld-MCP: Benchmarking MCP Tool Invocation In Computer-Use Agents", "publication": null, "published_at": "2025-10-28", "record_sha256": "751513dbec35b4ab271afdc2db349ebdcf060796cabc3c740ca0d5d342e43667", "relations": [], "resource_id": "arxiv:2510.24563", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.24563v2", "topics": ["evaluation", "tool-integration", "planning"], "updated_at": "2025-11-11"}, "text": "OSWorld-MCP: Benchmarking MCP Tool Invocation In Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOSWorld-MCP evaluates how agents choose between GUI operations and MCP tool calls in computer tasks. It combines generated and curated tools, with 158 manually validated tools across seven applications, to examine tool use alongside task completion.\nPlatforms: desktop\nTopics: evaluation, tool-integration, planning\nPublished: 2025-10-28\nAuthors: Hongrui Jia, Jitong Liao, Xi Zhang, Haiyang Xu, Tianbao Xie, Chaoya Jiang, Ming Yan, Si Liu, Wei Ye, Fei Huang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.24563v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-24563/"}
{"id": "arxiv:2510.27140", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61b611712fff54af25aae660f2353fa6c3aaa764b0767831ebf2278c120d9dea", "submitted_at": "2025-10-31T03:35:59Z", "updated_at": "2026-07-07T02:18:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.27140v3"}, "authors": ["Chenghao Du", "Quanfeng Huang", "Tingxuan Tang", "Zihao Wang", "Adwait Nadkarni", "Yue Xiao"], "canonical_source_url": "https://arxiv.org/abs/2510.27140", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-27140/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61b611712fff54af25aae660f2353fa6c3aaa764b0767831ebf2278c120d9dea", "submitted_at": "2025-10-31T03:35:59Z", "updated_at": "2026-07-07T02:18:26Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2510.27140v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2510.27140"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27140"}], "name": "Measuring the Security of Mobile LLM Agents under Adversarial Prompts from Untrusted Third-Party Channels", "publication": null, "published_at": "2025-10-31", "record_sha256": "bb74de8c942fc88b62778c78ec897af606b65f2e085699bee9afff5364993b69", "relations": [], "resource_id": "arxiv:2510.27140", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.27140v3", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-07-07"}, "text": "Measuring the Security of Mobile LLM Agents under Adversarial Prompts from Untrusted Third-Party Channels\nResource type: paper\nReview status: curated\nScope: core\n\nMeasures how untrusted third-party notifications, messages, and app content can hijack mobile LLM agents into executing unauthorized cross-app actions.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2025-10-31\nAuthors: Chenghao Du, Quanfeng Huang, Tingxuan Tang, Zihao Wang, Adwait Nadkarni, Yue Xiao\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.27140v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-27140/"}
{"id": "arxiv:2510.27210", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "283b1a1de2a4e1d4ec8175f0c15af66c10097a7f5ad9fd2793e93b888526ff7e", "submitted_at": "2025-10-31T06:10:57Z", "updated_at": "2025-10-31T06:10:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.27210v1"}, "authors": ["Tao Liu", "Chongyu Wang", "Rongjie Li", "Yingchen Yu", "Xuming He", "Bai Song"], "canonical_source_url": "https://arxiv.org/abs/2510.27210", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-27210/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "283b1a1de2a4e1d4ec8175f0c15af66c10097a7f5ad9fd2793e93b888526ff7e", "submitted_at": "2025-10-31T06:10:57Z", "updated_at": "2025-10-31T06:10:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2510.27210v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2510.27210"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27210"}], "name": "GUI-Rise: Structured Reasoning and History Summarization for GUI Navigation", "publication": null, "published_at": "2025-10-31", "record_sha256": "65421bca6c0802ec09e05682d18c425ad05389bd2bd070bb0e423da30a608594", "relations": [], "resource_id": "arxiv:2510.27210", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.27210v1", "topics": ["planning", "interaction-history", "reinforcement-learning"], "updated_at": "2025-10-31"}, "text": "GUI-Rise: Structured Reasoning and History Summarization for GUI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nCombines structured progress-and-decision reasoning with history summarization to improve cross-domain generalization in GUI navigation agents.\nPlatforms: unspecified\nTopics: planning, interaction-history, reinforcement-learning\nPublished: 2025-10-31\nAuthors: Tao Liu, Chongyu Wang, Rongjie Li, Yingchen Yu, Xuming He, Bai Song\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.27210v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-27210/"}
{"id": "arxiv:2510.27266", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6869e86d7cd4242ba8dcf942bb1cef3c157dff4f0ef37f7138ff5025d082c7b8", "submitted_at": "2025-10-31T08:07:02Z", "updated_at": "2026-05-27T02:32:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.27266v2"}, "authors": ["Shaojie Zhang", "Pei Fu", "Ruoceng Zhang", "Jiahui Yang", "Anan Du", "Xiuwen Xi", "Shaokang Wang", "Ying Huang", "Bin Qin", "Zhenbo Luo", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2510.27266", "citation_url": "https://awesomecua.github.io/resources/arxiv-2510-27266/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6869e86d7cd4242ba8dcf942bb1cef3c157dff4f0ef37f7138ff5025d082c7b8", "submitted_at": "2025-10-31T08:07:02Z", "updated_at": "2026-05-27T02:32:58Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2510.27266v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2510.27266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2510.27266"}], "name": "Enhancing Trustworthy GUI Grounding via Self-Critiqued Reinforcement Learning", "publication": null, "published_at": "2025-10-31", "record_sha256": "12045bdad3fd452dc892303a8885d359f0b0ef22453e3ec80a0d20f309e230fd", "relations": [], "resource_id": "arxiv:2510.27266", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2510.27266v2", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-05-27"}, "text": "Enhancing Trustworthy GUI Grounding via Self-Critiqued Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nHyperClick trains GUI grounders to produce a click and an accompanying confidence estimate using rewards for correctness and confidence alignment. Its evaluation examines localization, calibration, and abstention behavior.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, robustness, evaluation\nPublished: 2025-10-31\nAuthors: Shaojie Zhang, Pei Fu, Ruoceng Zhang, Jiahui Yang, Anan Du, Xiuwen Xi, Shaokang Wang, Ying Huang, Bin Qin, Zhenbo Luo, Jian Luan\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2510.27266v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2510-27266/"}
{"id": "arxiv:2511.00810", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c2224c58c8fbac7a68f3757391f5e6a4f62c32fa42aae86bf9821c0755b402b1", "submitted_at": "2025-11-02T05:34:21Z", "updated_at": "2026-06-30T10:43:39Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2511.00810v4"}, "authors": ["Shijie Zhou", "Viet Dac Lai", "Hao Tan", "Jihyung Kil", "Wanrong Zhu", "Changyou Chen", "Ruiyi Zhang"], "canonical_source_url": "https://arxiv.org/abs/2511.00810", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-00810/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c2224c58c8fbac7a68f3757391f5e6a4f62c32fa42aae86bf9821c0755b402b1", "submitted_at": "2025-11-02T05:34:21Z", "updated_at": "2026-06-30T10:43:39Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2511.00810v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.00810"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.00810"}], "name": "GUI-AIMA: Aligning Intrinsic Multimodal Attention with a Context Anchor for GUI Grounding", "publication": null, "published_at": "2025-11-02", "record_sha256": "2d495da3b9908359468d42a89c8498e4d772a7fde58282530488a014249df3aa", "relations": [], "resource_id": "arxiv:2511.00810", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.00810v4", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-06-30"}, "text": "GUI-AIMA: Aligning Intrinsic Multimodal Attention with a Context Anchor for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-AIMA, which supervises instruction-relevant attention over visual patches instead of directly training coordinate text generation. Patch-level grounding can be combined with an optional zoom stage and is evaluated across GUI localization benchmarks.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, efficiency\nPublished: 2025-11-02\nAuthors: Shijie Zhou, Viet Dac Lai, Hao Tan, Jihyung Kil, Wanrong Zhu, Changyou Chen, Ruiyi Zhang\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.00810v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-00810/"}
{"id": "arxiv:2511.04137", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a91413c1e80869860d411402ab49f6e5968787b3bca83f0a9066a0e9c2f75d40", "submitted_at": "2025-11-06T07:29:02Z", "updated_at": "2025-11-06T07:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04137v1"}, "authors": ["Yujian Liu", "Ze Wang", "Hao Chen", "Ximeng Sun", "Xiaodong Yu", "Jialian Wu", "Jiang Liu", "Emad Barsoum", "Zicheng Liu", "Shiyu Chang"], "canonical_source_url": "https://arxiv.org/abs/2511.04137", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-04137/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a91413c1e80869860d411402ab49f6e5968787b3bca83f0a9066a0e9c2f75d40", "submitted_at": "2025-11-06T07:29:02Z", "updated_at": "2025-11-06T07:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.04137v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.04137"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.04137"}], "name": "Learning from Online Videos at Inference Time for Computer-Use Agents", "publication": null, "published_at": "2025-11-06", "record_sha256": "18c707513c0ed630b878631ece0536cdc88d7ea7bfda8021b70675526cf8767e", "relations": [], "resource_id": "arxiv:2511.04137", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.04137v1", "topics": ["video", "memory", "planning", "data-generation"], "updated_at": "2025-11-06"}, "text": "Learning from Online Videos at Inference Time for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis work retrieves online video tutorials and converts them into structured GUI action demonstrations with short objectives. A two-stage selection process supplies a relevant local trajectory as in-context guidance at each execution step.\nPlatforms: unspecified\nTopics: video, memory, planning, data-generation\nPublished: 2025-11-06\nAuthors: Yujian Liu, Ze Wang, Hao Chen, Ximeng Sun, Xiaodong Yu, Jialian Wu, Jiang Liu, Emad Barsoum, Zicheng Liu, Shiyu Chang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.04137v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-04137/"}
{"id": "arxiv:2511.04307", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b063a4d2b3a552aa05a153745978473a7d165f2c6b4b97e4c91eff11b7b0f80", "submitted_at": "2025-11-06T12:19:02Z", "updated_at": "2025-11-10T12:27:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.04307v2"}, "authors": ["Jian Mu", "Chaoyun Zhang", "Chiming Ni", "Lu Wang", "Bo Qiao", "Kartik Mathur", "Qianhui Wu", "Yuhang Xie", "Xiaojun Ma", "Mengyu Zhou", "Si Qin", "Liqun Li", "Yu Kang", "Minghua Ma", "Qingwei Lin", "Saravan Rajmohan", "Dongmei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2511.04307", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-04307/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6b063a4d2b3a552aa05a153745978473a7d165f2c6b4b97e4c91eff11b7b0f80", "submitted_at": "2025-11-06T12:19:02Z", "updated_at": "2025-11-10T12:27:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.04307v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.04307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.04307"}], "name": "GUI-360$^\\circ$: A Comprehensive Dataset and Benchmark for Computer-Using Agents", "publication": null, "published_at": "2025-11-06", "record_sha256": "59d68483c46b1b274cdc395ca6fa2ed14ae20b9d1515cbbebc9d303d3da2578c", "relations": [], "resource_id": "arxiv:2511.04307", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.04307v2", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-10"}, "text": "GUI-360$^\\circ$: A Comprehensive Dataset and Benchmark for Computer-Using Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCollects Windows application trajectories with screenshots and action traces for grounding, screen parsing, and action prediction.\nPlatforms: desktop, windows\nTopics: data-generation, grounding, evaluation\nPublished: 2025-11-06\nAuthors: Jian Mu, Chaoyun Zhang, Chiming Ni, Lu Wang, Bo Qiao, Kartik Mathur, Qianhui Wu, Yuhang Xie, Xiaojun Ma, Mengyu Zhou, Si Qin, Liqun Li, Yu Kang, Minghua Ma, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang\nRecorded arXiv version: v2\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.04307v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-04307/"}
{"id": "arxiv:2511.06101", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7420d51203de06ed3afe6f2695b8612ec36dc072765a4f2394f366f9e6bf8838", "submitted_at": "2025-11-08T18:45:33Z", "updated_at": "2026-04-13T06:03:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.06101v3"}, "authors": ["Zhaoyang Wang", "Yiming Liang", "Xuchao Zhang", "Qianhui Wu", "Siwei Han", "Anson Bastos", "Rujia Wang", "Chetan Bansal", "Baolin Peng", "Jianfeng Gao", "Saravan Rajmohan", "Huaxiu Yao"], "canonical_source_url": "https://arxiv.org/abs/2511.06101", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-06101/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7420d51203de06ed3afe6f2695b8612ec36dc072765a4f2394f366f9e6bf8838", "submitted_at": "2025-11-08T18:45:33Z", "updated_at": "2026-04-13T06:03:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.06101v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.06101"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.06101"}], "name": "SynthAgent: Adapting Web Agents with Synthetic Supervision", "publication": null, "published_at": "2025-11-08", "record_sha256": "00d524387232da417c785fbadc5aef54907ae649d444a1529b96e7948e2c9276", "relations": [], "resource_id": "arxiv:2511.06101", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.06101v3", "topics": ["data-generation", "imitation-learning", "web-navigation"], "updated_at": "2026-04-13"}, "text": "SynthAgent: Adapting Web Agents with Synthetic Supervision\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SynthAgent, which generates tasks by exploring website elements and refines them when observations reveal conflicts. A later trajectory-refinement stage removes misaligned behavior before fine-tuning agents for the target websites.\nPlatforms: web\nTopics: data-generation, imitation-learning, web-navigation\nPublished: 2025-11-08\nAuthors: Zhaoyang Wang, Yiming Liang, Xuchao Zhang, Qianhui Wu, Siwei Han, Anson Bastos, Rujia Wang, Chetan Bansal, Baolin Peng, Jianfeng Gao, Saravan Rajmohan, Huaxiu Yao\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.06101v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-06101/"}
{"id": "arxiv:2511.06417", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a5f43873a47e5b47250cedd3d860abe2dc5ec9b85fe657444bf2b7378f3e8f7", "submitted_at": "2025-11-09T15:13:45Z", "updated_at": "2025-11-09T15:13:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.06417v1"}, "authors": ["Xiangwu Guo", "Difei Gao", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2511.06417", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-06417/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a5f43873a47e5b47250cedd3d860abe2dc5ec9b85fe657444bf2b7378f3e8f7", "submitted_at": "2025-11-09T15:13:45Z", "updated_at": "2025-11-09T15:13:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.06417v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.06417"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.06417"}], "name": "AUTO-Explorer: Automated Data Collection for GUI Agent", "publication": null, "published_at": "2025-11-09", "record_sha256": "f32200f1cadc263c3cb353f982c5213be453ea5a9b890440f616de3a57b4f1e4", "relations": [], "resource_id": "arxiv:2511.06417", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.06417v1", "topics": ["data-generation", "grounding", "evaluation"], "updated_at": "2025-11-09"}, "text": "AUTO-Explorer: Automated Data Collection for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Auto-Explorer, an automated GUI exploration method that records software states for training-data collection. UIXplore evaluates exploration coverage, and collected data is used to fine-tune and test GUI element grounding.\nPlatforms: unspecified\nTopics: data-generation, grounding, evaluation\nPublished: 2025-11-09\nAuthors: Xiangwu Guo, Difei Gao, Mike Zheng Shou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.06417v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-06417/"}
{"id": "arxiv:2511.07332", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f2b80f294ddf5e3331dcb47a836864032f88f2fdbe35588bfef000cd9ceddc8", "submitted_at": "2025-11-10T17:35:21Z", "updated_at": "2026-06-09T21:30:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.07332v2"}, "authors": ["Aarash Feizi", "Shravan Nayak", "Xiangru Jian", "Kevin Qinghong Lin", "Kaixin Li", "Rabiul Awal", "Xing Han Lù", "Johan Obando-Ceron", "Juan A. Rodriguez", "Nicolas Chapados", "David Vazquez", "Adriana Romero-Soriano", "Reihaneh Rabbany", "Perouz Taslakian", "Christopher Pal", "Spandana Gella", "Sai Rajeswar"], "canonical_source_url": "https://arxiv.org/abs/2511.07332", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-07332/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f2b80f294ddf5e3331dcb47a836864032f88f2fdbe35588bfef000cd9ceddc8", "submitted_at": "2025-11-10T17:35:21Z", "updated_at": "2026-06-09T21:30:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.07332v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.07332"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.07332"}], "name": "Grounding Computer Use Agents on Human Demonstrations", "publication": null, "published_at": "2025-11-10", "record_sha256": "5fcca76c183311c62ba5bc00c13fdbfa74bebec2f7e30bb6f67149356944521c", "relations": [], "resource_id": "arxiv:2511.07332", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.07332v2", "topics": ["grounding", "data-generation", "evaluation", "reinforcement-learning"], "updated_at": "2026-06-09"}, "text": "Grounding Computer Use Agents on Human Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nGroundCUA turns expert desktop demonstrations into densely annotated screenshots and grounding instructions. The GroundNext model family uses this supervision and subsequent reinforcement learning, with both grounding benchmarks and an OSWorld agent configuration used for evaluation.\nPlatforms: desktop\nTopics: grounding, data-generation, evaluation, reinforcement-learning\nPublished: 2025-11-10\nAuthors: Aarash Feizi, Shravan Nayak, Xiangru Jian, Kevin Qinghong Lin, Kaixin Li, Rabiul Awal, Xing Han Lù, Johan Obando-Ceron, Juan A. Rodriguez, Nicolas Chapados, David Vazquez, Adriana Romero-Soriano, Reihaneh Rabbany, Perouz Taslakian, Christopher Pal, Spandana Gella, Sai Rajeswar\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.07332v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-07332/"}
{"id": "arxiv:2511.09127", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0f61f09ce52b1e66fa81479483ea419aaa16e6b27ce331d3c1039927075fd91", "submitted_at": "2025-11-12T09:06:25Z", "updated_at": "2025-11-12T09:06:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09127v1"}, "authors": ["Ziwei Wang", "Leyang Yang", "Xiaoxuan Tang", "Sheng Zhou", "Dajun Chen", "Wei Jiang", "Yong Li"], "canonical_source_url": "https://arxiv.org/abs/2511.09127", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-09127/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0f61f09ce52b1e66fa81479483ea419aaa16e6b27ce331d3c1039927075fd91", "submitted_at": "2025-11-12T09:06:25Z", "updated_at": "2025-11-12T09:06:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09127v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.09127"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.09127"}], "name": "History-Aware Reasoning for GUI Agents", "publication": null, "published_at": "2025-11-12", "record_sha256": "3bb6aad63fe956826e1f3f7dc76bc7244cb3c27bc2a6adc16b9e9c18d6e367b1", "relations": [], "resource_id": "arxiv:2511.09127", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.09127v1", "topics": ["interaction-history", "memory", "self-reflection", "reinforcement-learning"], "updated_at": "2025-11-12"}, "text": "History-Aware Reasoning for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nHAR-GUI combines episodic reasoning and short-term memory to use earlier interactions when selecting GUI actions. Its training constructs reflective scenarios and correction guidance and applies a hybrid reinforcement-learning reward to the 3B policy.\nPlatforms: unspecified\nTopics: interaction-history, memory, self-reflection, reinforcement-learning\nPublished: 2025-11-12\nAuthors: Ziwei Wang, Leyang Yang, Xiaoxuan Tang, Sheng Zhou, Dajun Chen, Wei Jiang, Yong Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.09127v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-09127/"}
{"id": "arxiv:2511.09157", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7f5f0fe1ca373272cda4bbeca3a465e1a7eaec829e8a59bbecbe6049dc7b4a", "submitted_at": "2025-11-12T09:49:31Z", "updated_at": "2025-11-12T09:49:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09157v1"}, "authors": ["Leyang Yang", "Ziwei Wang", "Xiaoxuan Tang", "Sheng Zhou", "Dajun Chen", "Wei Jiang", "Yong Li"], "canonical_source_url": "https://arxiv.org/abs/2511.09157", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-09157/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7f5f0fe1ca373272cda4bbeca3a465e1a7eaec829e8a59bbecbe6049dc7b4a", "submitted_at": "2025-11-12T09:49:31Z", "updated_at": "2025-11-12T09:49:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.09157v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.09157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.09157"}], "name": "ProBench: Benchmarking GUI Agents with Accurate Process Information", "publication": null, "published_at": "2025-11-12", "record_sha256": "8906803a4591e200f1765a51a2ac33ff9d35f62b6f68d292c5eaf2b798922afc", "relations": [], "resource_id": "arxiv:2511.09157", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.09157v1", "topics": ["evaluation", "mobile-navigation", "observability"], "updated_at": "2025-11-12"}, "text": "ProBench: Benchmarking GUI Agents with Accurate Process Information\nResource type: paper\nReview status: curated\nScope: core\n\nProBench extends mobile GUI evaluation with tasks whose correctness depends on the interaction process as well as the final state. Its ProcessProvider supplies process information for assessing agent behavior and analyzing failure modes.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, observability\nPublished: 2025-11-12\nAuthors: Leyang Yang, Ziwei Wang, Xiaoxuan Tang, Sheng Zhou, Dajun Chen, Wei Jiang, Yong Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.09157v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-09157/"}
{"id": "arxiv:2511.10705", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101d38ed439c41ebfa826e990697b502b9126728956295d25b0ea9d539ff23e3", "submitted_at": "2025-11-13T03:41:02Z", "updated_at": "2025-11-13T03:41:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.10705v1"}, "authors": ["Yuan Zhao", "Hualei Zhu", "Tingyu Jiang", "Shen Li", "Xiaohang Xu", "Hao Henry Wang"], "canonical_source_url": "https://arxiv.org/abs/2511.10705", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-10705/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "101d38ed439c41ebfa826e990697b502b9126728956295d25b0ea9d539ff23e3", "submitted_at": "2025-11-13T03:41:02Z", "updated_at": "2025-11-13T03:41:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.10705v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.10705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.10705"}], "name": "Co-EPG: A Framework for Co-Evolution of Planning and Grounding in Autonomous GUI Agents", "publication": null, "published_at": "2025-11-13", "record_sha256": "192047a92def0132be2dc21f7c3872d0225a43f1fa0e1611797bea05a9344325", "relations": [], "resource_id": "arxiv:2511.10705", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.10705v1", "topics": ["grounding", "planning", "reinforcement-learning", "data-generation"], "updated_at": "2025-11-13"}, "text": "Co-EPG: A Framework for Co-Evolution of Planning and Grounding in Autonomous GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCo-EPG alternates reinforcement learning for a planner with updates to a grounding model using generated training examples. The evolving grounder provides feedback to the planner, and evaluation uses Multimodal Mind2Web and AndroidControl.\nPlatforms: web, android\nTopics: grounding, planning, reinforcement-learning, data-generation\nPublished: 2025-11-13\nAuthors: Yuan Zhao, Hualei Zhu, Tingyu Jiang, Shen Li, Xiaohang Xu, Hao Henry Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.10705v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-10705/"}
{"id": "arxiv:2511.11672", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755d41575ae8ee5d07be0022c8f078c742da4628f12ecd1cad984269441021cf", "submitted_at": "2025-11-11T20:03:38Z", "updated_at": "2026-04-02T02:42:36Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2511.11672v5"}, "authors": ["Zengyi Qin", "Jinyuan Chen", "Yunze Man", "Shengcao Cao", "Ziqi Pang", "Zhuoyuan Wang", "Han Fang", "Ling Zhu", "Zixin Xie", "Zibu Wei", "Tianshu Ran", "Haoran Geng", "Ray Pan", "Qizhen Sun", "Zachary Bright", "Yuyang Cai", "Chongye Yang", "Jiace Zhao", "Tianrui Liu", "Han Cao", "Yeyang Zhou", "Rui Wang", "Song Wang", "Xiang Ren", "Bo Zhang", "Yutong Ban", "Pieter Abbeel", "Brian Anthony"], "canonical_source_url": "https://arxiv.org/abs/2511.11672", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-11672/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "755d41575ae8ee5d07be0022c8f078c742da4628f12ecd1cad984269441021cf", "submitted_at": "2025-11-11T20:03:38Z", "updated_at": "2026-04-02T02:42:36Z", "version": 5, "versioned_url": "https://arxiv.org/abs/2511.11672v5"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.11672"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.11672"}], "name": "OSGym: Scalable OS Infra for Computer Use Agents", "publication": null, "published_at": "2025-11-11", "record_sha256": "e2a70e1137053235c7cd3c0574f25da59b66885e0f1f501d5f3280b492484146", "relations": [], "resource_id": "arxiv:2511.11672", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.11672v5", "topics": ["sandbox", "efficiency", "reinforcement-learning", "data-generation"], "updated_at": "2026-04-02"}, "text": "OSGym: Scalable OS Infra for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOSGym supplies full operating-system sandboxes with GUIs for parallel computer-use data collection and training. It combines per-replica state management, hardware-aware orchestration, copy-on-write virtual-machine disks, and layered failure recovery.\nPlatforms: unspecified\nTopics: sandbox, efficiency, reinforcement-learning, data-generation\nPublished: 2025-11-11\nAuthors: Zengyi Qin, Jinyuan Chen, Yunze Man, Shengcao Cao, Ziqi Pang, Zhuoyuan Wang, Han Fang, Ling Zhu, Zixin Xie, Zibu Wei, Tianshu Ran, Haoran Geng, Ray Pan, Qizhen Sun, Zachary Bright, Yuyang Cai, Chongye Yang, Jiace Zhao, Tianrui Liu, Han Cao, Yeyang Zhou, Rui Wang, Song Wang, Xiang Ren, Bo Zhang, Yutong Ban, Pieter Abbeel, Brian Anthony\nRecorded arXiv version: v5\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.11672v5\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-11672/"}
{"id": "arxiv:2511.12254", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ede08df1c10d8eeede43af75d26608b4917dd83becde23f28cd3adf335f79736", "submitted_at": "2025-11-15T15:22:42Z", "updated_at": "2026-03-12T08:45:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.12254v3"}, "authors": ["Yuxiang Zhou", "Jichang Li", "Yanhao Zhang", "Haonan Lu", "Guanbin Li"], "canonical_source_url": "https://arxiv.org/abs/2511.12254", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-12254/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ede08df1c10d8eeede43af75d26608b4917dd83becde23f28cd3adf335f79736", "submitted_at": "2025-11-15T15:22:42Z", "updated_at": "2026-03-12T08:45:07Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2511.12254v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2511.12254"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.12254"}], "name": "Mobile-Agent-RAG: Driving Smart Multi-Agent Coordination with Contextual Knowledge Empowerment for Long-Horizon Mobile Automation", "publication": null, "published_at": "2025-11-15", "record_sha256": "edbf6269ccf490335257f8378902eafdced39fbb53cd673f551b7848ca3e9509", "relations": [], "resource_id": "arxiv:2511.12254", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.12254v3", "topics": ["mobile-navigation", "memory", "long-horizon"], "updated_at": "2026-03-12"}, "text": "Mobile-Agent-RAG: Driving Smart Multi-Agent Coordination with Contextual Knowledge Empowerment for Long-Horizon Mobile Automation\nResource type: paper\nReview status: curated\nScope: core\n\nUses separate retrieval systems for high-level mobile-task planning and precise interface operations, with a benchmark for extended cross-app tasks.\nPlatforms: unspecified\nTopics: mobile-navigation, memory, long-horizon\nPublished: 2025-11-15\nAuthors: Yuxiang Zhou, Jichang Li, Yanhao Zhang, Haonan Lu, Guanbin Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.12254v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-12254/"}
{"id": "arxiv:2511.12997", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a105701d9265bacd3971f71f34a0e7ed315b6e54e9a7e675588c50713fcc4de", "submitted_at": "2025-11-17T05:38:50Z", "updated_at": "2026-07-22T21:19:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.12997v2"}, "authors": ["Genglin Liu", "Shijie Geng", "Sha Li", "Hejie Cui", "Sarah Zhang", "Xin Liu", "Tianyi Liu"], "canonical_source_url": "https://arxiv.org/abs/2511.12997", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-12997/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a105701d9265bacd3971f71f34a0e7ed315b6e54e9a7e675588c50713fcc4de", "submitted_at": "2025-11-17T05:38:50Z", "updated_at": "2026-07-22T21:19:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.12997v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.12997"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.12997"}], "name": "WebCoach: Self-Evolving Web Agents with Cross-Session Memory Guidance", "publication": null, "published_at": "2025-11-17", "record_sha256": "592e05607e633338120da7bd3312638da0d8802371e886316510df57202b6b40", "relations": [], "resource_id": "arxiv:2511.12997", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.12997v2", "topics": ["memory", "interaction-history", "web-navigation"], "updated_at": "2026-07-22"}, "text": "WebCoach: Self-Evolving Web Agents with Cross-Session Memory Guidance\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebCoach, which condenses navigation logs into persistent episodic memory and retrieves experiences by similarity and recency. Runtime hooks selectively supply advice to browser agents, evaluated on WebVoyager without retraining the underlying models.\nPlatforms: web\nTopics: memory, interaction-history, web-navigation\nPublished: 2025-11-17\nAuthors: Genglin Liu, Shijie Geng, Sha Li, Hejie Cui, Sarah Zhang, Xin Liu, Tianyi Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.12997v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-12997/"}
{"id": "arxiv:2511.13087", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cdabdd4fc56b95c2b0c801fefd76d4c724aa0a1d7f762623d4314ca9a38e85c", "submitted_at": "2025-11-17T07:38:05Z", "updated_at": "2025-11-17T07:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13087v1"}, "authors": ["SeokJoo Kwak", "Jihoon Kim", "Boyoun Kim", "Jung Jae Yoon", "Wooseok Jang", "Jeonghoon Hong", "Jaeho Yang", "Yeong-Dae Kwon"], "canonical_source_url": "https://arxiv.org/abs/2511.13087", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-13087/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2cdabdd4fc56b95c2b0c801fefd76d4c724aa0a1d7f762623d4314ca9a38e85c", "submitted_at": "2025-11-17T07:38:05Z", "updated_at": "2025-11-17T07:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13087v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2511.13087"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.13087"}], "name": "MEGA-GUI: Multi-stage Enhanced Grounding Agents for GUI Elements", "publication": null, "published_at": "2025-11-17", "record_sha256": "93b3acc43dfa9fbe99959460294e106ca16462fbe8856e287a2c024d82c23c44", "relations": [], "resource_id": "arxiv:2511.13087", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.13087v1", "topics": ["grounding", "screen-understanding"], "updated_at": "2025-11-17"}, "text": "MEGA-GUI: Multi-stage Enhanced Grounding Agents for GUI Elements\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MEGA-GUI, which separates coarse region selection from fine-grained interface-element grounding. Bidirectional zooming and context-aware instruction rewriting coordinate specialized vision-language agents across visual scales.\nPlatforms: unspecified\nTopics: grounding, screen-understanding\nPublished: 2025-11-17\nAuthors: SeokJoo Kwak, Jihoon Kim, Boyoun Kim, Jung Jae Yoon, Wooseok Jang, Jeonghoon Hong, Jaeho Yang, Yeong-Dae Kwon\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.13087v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-13087/"}
{"id": "arxiv:2511.13248", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d918a5754fe907fc273639f3277a3566d649c08da244714d0d7737456bede230", "submitted_at": "2025-11-17T11:08:45Z", "updated_at": "2025-11-17T11:08:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13248v1"}, "authors": ["Fuyao Zhang", "Jiaming Zhang", "Che Wang", "Xiongtao Sun", "Yurong Hao", "Guowei Guan", "Wenjie Li", "Longtao Huang", "Wei Yang Bryan Lim"], "canonical_source_url": "https://arxiv.org/abs/2511.13248", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-13248/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d918a5754fe907fc273639f3277a3566d649c08da244714d0d7737456bede230", "submitted_at": "2025-11-17T11:08:45Z", "updated_at": "2025-11-17T11:08:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.13248v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2511.13248"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.13248"}], "name": "DualTAP: A Dual-Task Adversarial Protector for Mobile MLLM Agents", "publication": null, "published_at": "2025-11-17", "record_sha256": "98acb9a218348d371a0c961ff6c5db2f489fb706b406c362211a29e56cc757c6", "relations": [], "resource_id": "arxiv:2511.13248", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.13248v1", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2025-11-17"}, "text": "DualTAP: A Dual-Task Adversarial Protector for Mobile MLLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nApplies dual-task adversarial perturbations to mobile screenshots to conceal personally identifiable information from untrusted MLLM routers while preserving agent task utility.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation\nPublished: 2025-11-17\nAuthors: Fuyao Zhang, Jiaming Zhang, Che Wang, Xiongtao Sun, Yurong Hao, Guowei Guan, Wenjie Li, Longtao Huang, Wei Yang Bryan Lim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.13248v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-13248/"}
{"id": "arxiv:2511.15567", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d63831bc0ff00979615415ce94abc140b646a4375cb33b950c4a7ccf00bf166", "submitted_at": "2025-11-19T16:00:02Z", "updated_at": "2025-11-19T16:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.15567v1"}, "authors": ["Kevin Qinghong Lin", "Siyuan Hu", "Linjie Li", "Zhengyuan Yang", "Lijuan Wang", "Philip Torr", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2511.15567", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-15567/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d63831bc0ff00979615415ce94abc140b646a4375cb33b950c4a7ccf00bf166", "submitted_at": "2025-11-19T16:00:02Z", "updated_at": "2025-11-19T16:00:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.15567v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.15567"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.15567"}], "name": "Computer-Use Agents as Judges for Generative User Interface", "publication": null, "published_at": "2025-11-19", "record_sha256": "4e0a55e69ed839f21ed1b9540e10c7a8e2172549e091f793c8d22000f8820811", "relations": [], "resource_id": "arxiv:2511.15567", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.15567v1", "topics": ["evaluation", "ui-design", "observability"], "updated_at": "2025-11-19"}, "text": "Computer-Use Agents as Judges for Generative User Interface\nResource type: paper\nReview status: curated\nScope: core\n\nAUI-Gym evaluates generated websites by having computer-use agents attempt tasks whose executability is checked programmatically. A feedback dashboard condenses navigation histories into visual summaries that a coding agent can use to revise an interface.\nPlatforms: web\nTopics: evaluation, ui-design, observability\nPublished: 2025-11-19\nAuthors: Kevin Qinghong Lin, Siyuan Hu, Linjie Li, Zhengyuan Yang, Lijuan Wang, Philip Torr, Mike Zheng Shou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.15567v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-15567/"}
{"id": "arxiv:2511.16590", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1679f7b56c7b60f92195e488c887114bcdcce7975a80fafbf2572c2b1f01ce0b", "submitted_at": "2025-11-20T17:43:46Z", "updated_at": "2025-11-20T17:43:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.16590v1"}, "authors": ["Sen Chen", "Tong Zhao", "Yi Bin", "Fei Ma", "Wenqi Shao", "Zheng Wang"], "canonical_source_url": "https://arxiv.org/abs/2511.16590", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-16590/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1679f7b56c7b60f92195e488c887114bcdcce7975a80fafbf2572c2b1f01ce0b", "submitted_at": "2025-11-20T17:43:46Z", "updated_at": "2025-11-20T17:43:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.16590v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.16590"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.16590"}], "name": "D-GARA: A Dynamic Benchmarking Framework for GUI Agent Robustness in Real-World Anomalies", "publication": null, "published_at": "2025-11-20", "record_sha256": "710c008749c73be4b0cd150736067d971f3a050e9b86f90fc46b05655e8f264b", "relations": [], "resource_id": "arxiv:2511.16590", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.16590v1", "topics": ["evaluation", "robustness", "mobile-navigation"], "updated_at": "2025-11-20"}, "text": "D-GARA: A Dynamic Benchmarking Framework for GUI Agent Robustness in Real-World Anomalies\nResource type: paper\nReview status: curated\nScope: core\n\nD-GARA evaluates Android GUI agents under runtime anomalies such as permission dialogs, battery warnings, and update prompts. Its modular environment embeds these interruptions in Android applications to study recovery and robustness.\nPlatforms: android\nTopics: evaluation, robustness, mobile-navigation\nPublished: 2025-11-20\nAuthors: Sen Chen, Tong Zhao, Yi Bin, Fei Ma, Wenqi Shao, Zheng Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.16590v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-16590/"}
{"id": "arxiv:2511.17131", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c746b329229d90dbe3d04a138f54c8b205b16ec6173cf681d47a319c8f24ddf", "submitted_at": "2025-11-21T10:47:22Z", "updated_at": "2025-11-21T10:47:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.17131v1"}, "authors": ["Horia Cristescu", "Charles Park", "Trong Canh Nguyen", "Sergiu Talmacel", "Alexandru-Gabriel Ilie", "Stefan Adam"], "canonical_source_url": "https://arxiv.org/abs/2511.17131", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-17131/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c746b329229d90dbe3d04a138f54c8b205b16ec6173cf681d47a319c8f24ddf", "submitted_at": "2025-11-21T10:47:22Z", "updated_at": "2025-11-21T10:47:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.17131v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.17131"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.17131"}], "name": "UI-CUBE: Enterprise-Grade Computer Use Agent Benchmarking Beyond Task Accuracy to Operational Reliability", "publication": null, "published_at": "2025-11-21", "record_sha256": "cc1867f285cd688d7a6968cf2c366470ea824ef962d4c5e8618b23a484c6f4ed", "relations": [], "resource_id": "arxiv:2511.17131", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.17131v1", "topics": ["evaluation", "industry", "workflow", "robustness"], "updated_at": "2025-11-21"}, "text": "UI-CUBE: Enterprise-Grade Computer Use Agent Benchmarking Beyond Task Accuracy to Operational Reliability\nResource type: paper\nReview status: curated\nScope: core\n\nUI-CUBE evaluates 226 tasks ranging from individual controls to copy-paste and enterprise workflows. Interface variations, multiple resolutions, and application-state validation support diagnosis of operational reliability beyond task accuracy alone.\nPlatforms: unspecified\nTopics: evaluation, industry, workflow, robustness\nPublished: 2025-11-21\nAuthors: Horia Cristescu, Charles Park, Trong Canh Nguyen, Sergiu Talmacel, Alexandru-Gabriel Ilie, Stefan Adam\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.17131v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-17131/"}
{"id": "arxiv:2511.19477", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17debfa3b59e5a2fcae950c2659e5851ae101cc58181e0cc4e0320befe9b4cb1", "submitted_at": "2025-11-22T12:18:35Z", "updated_at": "2025-11-22T12:18:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19477v1"}, "authors": ["Aram Vardanyan"], "canonical_source_url": "https://arxiv.org/abs/2511.19477", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-19477/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "17debfa3b59e5a2fcae950c2659e5851ae101cc58181e0cc4e0320befe9b4cb1", "submitted_at": "2025-11-22T12:18:35Z", "updated_at": "2025-11-22T12:18:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19477v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.19477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.19477"}], "name": "Building Browser Agents: Architecture, Security, and Practical Solutions", "publication": null, "published_at": "2025-11-22", "record_sha256": "1999ba9b3876ba3bbc69260e1e634885ce83e9d3453318c4950432f80f3c60cd", "relations": [], "resource_id": "arxiv:2511.19477", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.19477v1", "topics": ["web-navigation", "safety-security", "tool-integration"], "updated_at": "2025-11-22"}, "text": "Building Browser Agents: Architecture, Security, and Practical Solutions\nResource type: paper\nReview status: curated\nScope: core\n\nReports architectural and security lessons from building a browser agent, including hybrid accessibility-tree and visual context, browser tooling and programmatic constraints. A WebGames evaluation accompanies the discussion of reliability and specialization.\nPlatforms: web\nTopics: web-navigation, safety-security, tool-integration\nPublished: 2025-11-22\nAuthors: Aram Vardanyan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. Broader claims about the safety of general-purpose browsing are the paper's position, not a catalog conclusion.\n\nPrimary source: https://arxiv.org/abs/2511.19477v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-19477/"}
{"id": "arxiv:2511.19663", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0e23ba9b31a01956836e67329652882f84fe8ebef1075385a6af5e68cdae936", "submitted_at": "2025-11-24T19:56:28Z", "updated_at": "2025-11-24T19:56:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19663v1"}, "authors": ["Ahmed Awadallah", "Yash Lara", "Raghav Magazine", "Hussein Mozannar", "Akshay Nambi", "Yash Pandya", "Aravind Rajeswaran", "Corby Rosset", "Alexey Taymanov", "Vibhav Vineet", "Spencer Whitehead", "Andrew Zhao"], "canonical_source_url": "https://arxiv.org/abs/2511.19663", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-19663/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0e23ba9b31a01956836e67329652882f84fe8ebef1075385a6af5e68cdae936", "submitted_at": "2025-11-24T19:56:28Z", "updated_at": "2025-11-24T19:56:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.19663v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.19663"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.19663"}], "name": "Fara-7B: An Efficient Agentic Model for Computer Use", "publication": null, "published_at": "2025-11-24", "record_sha256": "f6531cb44bf5e6863623abe3f7846f732a2c9b4bb2f115270d1f9b7c17461e0e", "relations": [], "resource_id": "arxiv:2511.19663", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.19663v1", "topics": ["data-generation", "efficiency", "web-navigation"], "updated_at": "2025-11-24"}, "text": "Fara-7B: An Efficient Agentic Model for Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces FaraGen synthetic web trajectories, a compact screenshot-based action model, and WebTailBench for underrepresented browser tasks.\nPlatforms: web\nTopics: data-generation, efficiency, web-navigation\nPublished: 2025-11-24\nAuthors: Ahmed Awadallah, Yash Lara, Raghav Magazine, Hussein Mozannar, Akshay Nambi, Yash Pandya, Aravind Rajeswaran, Corby Rosset, Alexey Taymanov, Vibhav Vineet, Spencer Whitehead, Andrew Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.19663v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-19663/"}
{"id": "arxiv:2511.20067", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff9343bb706687caf2ea813220ee7221819cd5892ba950f3168b91aebe0a74e0", "submitted_at": "2025-11-25T08:40:33Z", "updated_at": "2025-11-25T08:40:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20067v1"}, "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "canonical_source_url": "https://arxiv.org/abs/2511.20067", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-20067/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff9343bb706687caf2ea813220ee7221819cd5892ba950f3168b91aebe0a74e0", "submitted_at": "2025-11-25T08:40:33Z", "updated_at": "2025-11-25T08:40:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20067v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.20067"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.20067"}], "name": "\"Are We Done Yet?\": A Vision-Based Judge for Autonomous Task Completion of Computer Use Agents", "publication": null, "published_at": "2025-11-25", "record_sha256": "e979470fd6dfdfcd1dc833c143ebe3f2134410293be3f67b609e365abb5b975c", "relations": [], "resource_id": "arxiv:2511.20067", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.20067v1", "topics": ["evaluation", "self-reflection"], "updated_at": "2025-11-25"}, "text": "\"Are We Done Yet?\": A Vision-Based Judge for Autonomous Task Completion of Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study tests vision-language models as task-completion judges using screenshots and instructions from 42 built-in macOS applications. It evaluates judgment accuracy on human-labeled tasks and separately studies how judge feedback affects agent task success.\nPlatforms: desktop, macos\nTopics: evaluation, self-reflection\nPublished: 2025-11-25\nAuthors: Marta Sumyk, Oleksandr Kosovan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.20067v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-20067/"}
{"id": "arxiv:2511.20766", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c288e7c42b5eedeea3f27a5622e87981f101f472db05ead6d6bede1371500f0", "submitted_at": "2025-11-25T19:00:22Z", "updated_at": "2025-11-25T19:00:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20766v1"}, "authors": ["Karen Ullrich", "Jingtong Su", "Claudia Shi", "Arjun Subramonian", "Amir Bar", "Ivan Evtimov", "Nikolaos Tsilivis", "Randall Balestriero", "Julia Kempe", "Mark Ibrahim"], "canonical_source_url": "https://arxiv.org/abs/2511.20766", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-20766/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c288e7c42b5eedeea3f27a5622e87981f101f472db05ead6d6bede1371500f0", "submitted_at": "2025-11-25T19:00:22Z", "updated_at": "2025-11-25T19:00:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.20766v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2511.20766"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.20766"}], "name": "OpenApps: Simulating Environment Variations to Measure UI-Agent Reliability", "publication": null, "published_at": "2025-11-25", "record_sha256": "b968fcc5519597c4287931fd0c1ccc56d64cfd4e09d742be52fb0938c3c42d5d", "relations": [], "resource_id": "arxiv:2511.20766", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.20766v1", "topics": ["evaluation", "robustness", "data-generation"], "updated_at": "2025-11-25"}, "text": "OpenApps: Simulating Environment Variations to Measure UI-Agent Reliability\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OpenApps, a configurable app ecosystem for measuring UI-agent reliability under changes in appearance and content. Its evaluation compares task success and failure behavior across many versions of the same apps.\nPlatforms: unspecified\nTopics: evaluation, robustness, data-generation\nPublished: 2025-11-25\nAuthors: Karen Ullrich, Jingtong Su, Claudia Shi, Arjun Subramonian, Amir Bar, Ivan Evtimov, Nikolaos Tsilivis, Randall Balestriero, Julia Kempe, Mark Ibrahim\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.20766v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-20766/"}
{"id": "arxiv:2511.21398", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af9da8a618674b5a40f95029c4797b48a850e0a99de827d3e526b21797f8eef4", "submitted_at": "2025-11-26T13:49:39Z", "updated_at": "2025-11-26T13:49:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.21398v1"}, "authors": ["Jiayuan Zhang", "Kaiquan Chen", "Zhihao Lu", "Enshen Zhou", "Qian Yu", "Jing Zhang"], "canonical_source_url": "https://arxiv.org/abs/2511.21398", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-21398/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af9da8a618674b5a40f95029c4797b48a850e0a99de827d3e526b21797f8eef4", "submitted_at": "2025-11-26T13:49:39Z", "updated_at": "2025-11-26T13:49:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2511.21398v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2511.21398"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.21398"}], "name": "Prune4Web: DOM Tree Pruning Programming for Web Agent", "publication": null, "published_at": "2025-11-26", "record_sha256": "1e5a0e4903c9711b7961fa4eba8b49a0f8380c367fa7eb819754002a0abc1e89", "relations": [], "resource_id": "arxiv:2511.21398", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.21398v1", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2025-11-26"}, "text": "Prune4Web: DOM Tree Pruning Programming for Web Agent\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Prune4Web, where generated scoring programs filter DOM elements according to decomposed subtasks before grounding. An annotation pipeline and two-turn training procedure coordinate the planner, programmatic filter and grounder.\nPlatforms: web\nTopics: grounding, screen-understanding, efficiency\nPublished: 2025-11-26\nAuthors: Jiayuan Zhang, Kaiquan Chen, Zhihao Lu, Enshen Zhou, Qian Yu, Jing Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.21398v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-21398/"}
{"id": "arxiv:2511.22235", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb395dfaad18b84ea4b52d97be194a42eb42d6ac922c55f2179e38cc258a31", "submitted_at": "2025-11-27T09:01:38Z", "updated_at": "2026-03-04T12:15:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.22235v2"}, "authors": ["Zehao Deng", "Tianjie Ju", "Zheng Wu", "Zhuosheng Zhang", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2511.22235", "citation_url": "https://awesomecua.github.io/resources/arxiv-2511-22235/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9acb395dfaad18b84ea4b52d97be194a42eb42d6ac922c55f2179e38cc258a31", "submitted_at": "2025-11-27T09:01:38Z", "updated_at": "2026-03-04T12:15:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2511.22235v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2511.22235"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2511.22235"}], "name": "Training High-Level Schedulers with Execution-Feedback Reinforcement Learning for Long-Horizon GUI Automation", "publication": null, "published_at": "2025-11-27", "record_sha256": "093caa4c10d280aeac443a48f84a00c9464fd13b778a17912ed6f2b111ecd8c4", "relations": [], "resource_id": "arxiv:2511.22235", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2511.22235v2", "topics": ["planning", "state-tracking", "reinforcement-learning", "long-horizon"], "updated_at": "2026-03-04"}, "text": "Training High-Level Schedulers with Execution-Feedback Reinforcement Learning for Long-Horizon GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nThis work trains high-level schedulers with staged reinforcement learning driven by execution feedback. Its CES framework separates a Coordinator for planning and decomposition, a StateTracker for compressed context, and a low-level Executor.\nPlatforms: unspecified\nTopics: planning, state-tracking, reinforcement-learning, long-horizon\nPublished: 2025-11-27\nAuthors: Zehao Deng, Tianjie Ju, Zheng Wu, Zhuosheng Zhang, Gongshen Liu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2511.22235v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2511-22235/"}
{"id": "arxiv:2512.00756", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf16987178fc24fb483bfa82efecb1924c7528fd16ac94f1144fcc345748850f", "submitted_at": "2025-11-30T06:47:33Z", "updated_at": "2026-04-28T03:32:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00756v2"}, "authors": ["Ruihan Chen", "Qiming Li", "Xiaocheng Feng", "Weihong Zhong", "Xiaoliang Yang", "Yuxuan Gu", "Zekun Zhou", "Yunfei Lu", "Haoyu Ren", "Kun Chen", "Dandan Tu", "Bing Qin"], "canonical_source_url": "https://arxiv.org/abs/2512.00756", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-00756/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf16987178fc24fb483bfa82efecb1924c7528fd16ac94f1144fcc345748850f", "submitted_at": "2025-11-30T06:47:33Z", "updated_at": "2026-04-28T03:32:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00756v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.00756"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.00756"}], "name": "MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents", "publication": null, "published_at": "2025-11-30", "record_sha256": "2edb28b141f24ac0aee0f6a0d899c6613da1294fbf893500c079bf98af191fb6", "relations": [], "resource_id": "arxiv:2512.00756", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.00756v2", "topics": ["evaluation", "screen-understanding", "robustness"], "updated_at": "2026-04-28"}, "text": "MPR-GUI: Benchmarking and Enhancing Multilingual Perception and Reasoning in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMPR-GUI evaluates GUI perception and reasoning through strictly aligned tasks in six languages. GUI-XLI adjusts non-English hidden representations toward English representations at language-sensitive layers during inference to address the measured cross-language gaps.\nPlatforms: unspecified\nTopics: evaluation, screen-understanding, robustness\nPublished: 2025-11-30\nAuthors: Ruihan Chen, Qiming Li, Xiaocheng Feng, Weihong Zhong, Xiaoliang Yang, Yuxuan Gu, Zekun Zhou, Yunfei Lu, Haoyu Ren, Kun Chen, Dandan Tu, Bing Qin\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.00756v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-00756/"}
{"id": "arxiv:2512.00846", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "faad266b89aed9b210a813b69184d7236f9098c0e095cc426402e154cbdd067b", "submitted_at": "2025-11-30T11:32:54Z", "updated_at": "2025-12-11T18:16:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00846v2"}, "authors": ["Neeraj Anand", "Rishabh Jain", "Sohan Patnaik", "Balaji Krishnamurthy", "Mausoom Sarkar"], "canonical_source_url": "https://arxiv.org/abs/2512.00846", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-00846/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "faad266b89aed9b210a813b69184d7236f9098c0e095cc426402e154cbdd067b", "submitted_at": "2025-11-30T11:32:54Z", "updated_at": "2025-12-11T18:16:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.00846v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.00846"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.00846"}], "name": "AFRAgent : An Adaptive Feature Renormalization Based High Resolution Aware GUI agent", "publication": null, "published_at": "2025-11-30", "record_sha256": "79af123a28d2ddb31c436e450d113ce5ee29f5cf01e1caa12019c58f8bb1746d", "relations": [], "resource_id": "arxiv:2512.00846", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.00846v2", "topics": ["grounding", "mobile-navigation", "efficiency"], "updated_at": "2025-12-11"}, "text": "AFRAgent : An Adaptive Feature Renormalization Based High Resolution Aware GUI agent\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AFRAgent, which uses adaptive feature renormalization to combine low-resolution visual embeddings with high-resolution details. The architecture is evaluated for smartphone GUI automation on Meta-GUI and AITW.\nPlatforms: unspecified\nTopics: grounding, mobile-navigation, efficiency\nPublished: 2025-11-30\nAuthors: Neeraj Anand, Rishabh Jain, Sohan Patnaik, Balaji Krishnamurthy, Mausoom Sarkar\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.00846v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-00846/"}
{"id": "arxiv:2512.01174", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e445d0ca79f727a6c1d51e83381c3fd1d8e1fddeff7de99fe70ca06bc1bff49", "submitted_at": "2025-12-01T01:18:21Z", "updated_at": "2025-12-01T01:18:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01174v1"}, "authors": ["Hyunjun Kim", "Sooyoung Ryu"], "canonical_source_url": "https://arxiv.org/abs/2512.01174", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-01174/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e445d0ca79f727a6c1d51e83381c3fd1d8e1fddeff7de99fe70ca06bc1bff49", "submitted_at": "2025-12-01T01:18:21Z", "updated_at": "2025-12-01T01:18:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01174v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2512.01174"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01174"}], "name": "DrawingBench: Evaluating Spatial Reasoning and UI Interaction Capabilities of Large Language Models through Mouse-Based Drawing Tasks", "publication": null, "published_at": "2025-12-01", "record_sha256": "84340fe72ab3bb370a5aaa3687eb76f461a6b7235b9cfc6b709d51dca96e2272", "relations": [], "resource_id": "arxiv:2512.01174", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.01174v1", "topics": ["evaluation", "grounding", "human-in-the-loop", "state-tracking"], "updated_at": "2025-12-01"}, "text": "DrawingBench: Evaluating Spatial Reasoning and UI Interaction Capabilities of Large Language Models through Mouse-Based Drawing Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nDrawingBench evaluates sequences of low-level mouse-based GUI actions through 250 drawing prompts and rule-based criteria. Action inspection and multi-turn external feedback support analysis of spatial reasoning, tool-state management, and refinement.\nPlatforms: unspecified\nTopics: evaluation, grounding, human-in-the-loop, state-tracking\nPublished: 2025-12-01\nAuthors: Hyunjun Kim, Sooyoung Ryu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.01174v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-01174/"}
{"id": "arxiv:2512.01763", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a6eb58f9b6349ab9cc33444d8ea82c8f17653cc4a7f4b19ffc94e3b4fb894b3", "submitted_at": "2025-12-01T15:06:45Z", "updated_at": "2026-03-08T10:18:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.01763v2"}, "authors": ["Xurui Zhou", "Gongwei Chen", "Yuquan Xie", "Zaijing Li", "Kaiwen Zhou", "Shuai Wang", "Shuo Yang", "Zhuotao Tian", "Rui Shao"], "canonical_source_url": "https://arxiv.org/abs/2512.01763", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-01763/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a6eb58f9b6349ab9cc33444d8ea82c8f17653cc4a7f4b19ffc94e3b4fb894b3", "submitted_at": "2025-12-01T15:06:45Z", "updated_at": "2026-03-08T10:18:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.01763v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.01763"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01763"}], "name": "HiconAgent: History Context-aware Policy Optimization for GUI Agents", "publication": null, "published_at": "2025-12-01", "record_sha256": "46fd9e7821f04bd076c9f67670757bb996a40ce901d690e92c0ca89b78f0a0f9", "relations": [], "resource_id": "arxiv:2512.01763", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.01763v2", "topics": ["interaction-history", "memory", "efficiency"], "updated_at": "2026-03-08"}, "text": "HiconAgent: History Context-aware Policy Optimization for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nHiconAgent trains GUI policies to use history through variable-length context sampling and paired policy updates with compressed and complete histories. Its compression retains past actions while removing historical observations, and an alignment loss couples the two branches.\nPlatforms: android\nTopics: interaction-history, memory, efficiency\nPublished: 2025-12-01\nAuthors: Xurui Zhou, Gongwei Chen, Yuquan Xie, Zaijing Li, Kaiwen Zhou, Shuai Wang, Shuo Yang, Zhuotao Tian, Rui Shao\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.01763v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-01763/"}
{"id": "arxiv:2512.01979", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4868c3c85545532e9c208623596d3810c85e32e3cf11429a68847ac11c2ba1cb", "submitted_at": "2025-12-01T18:37:19Z", "updated_at": "2025-12-01T18:37:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01979v1"}, "authors": ["Aiden Yiliu Li", "Bizhi Yu", "Daoan Lei", "Tianhe Ren", "Shilong Liu"], "canonical_source_url": "https://arxiv.org/abs/2512.01979", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-01979/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4868c3c85545532e9c208623596d3810c85e32e3cf11429a68847ac11c2ba1cb", "submitted_at": "2025-12-01T18:37:19Z", "updated_at": "2025-12-01T18:37:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.01979v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.01979"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.01979"}], "name": "Chain-of-Ground: Improving GUI Grounding via Iterative Reasoning and Reference Feedback", "publication": null, "published_at": "2025-12-01", "record_sha256": "7fe296eb15289bd11a39d87ff8008909b79746263881483b874dc6cb1c359a62", "relations": [], "resource_id": "arxiv:2512.01979", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.01979v1", "topics": ["grounding", "training-free", "robustness", "evaluation"], "updated_at": "2025-12-01"}, "text": "Chain-of-Ground: Improving GUI Grounding via Iterative Reasoning and Reference Feedback\nResource type: paper\nReview status: curated\nScope: core\n\nChain-of-Ground uses iterative visual reasoning and reference feedback to refine GUI localization without additional training. It evaluates on ScreenSpot-Pro and introduces TPanel-UI to probe grounding on industrial control panels with visual distortions.\nPlatforms: unspecified\nTopics: grounding, training-free, robustness, evaluation\nPublished: 2025-12-01\nAuthors: Aiden Yiliu Li, Bizhi Yu, Daoan Lei, Tianhe Ren, Shilong Liu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.01979v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-01979/"}
{"id": "arxiv:2512.02423", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fda844ac6767e6d9dfd16919042469e2070ec7f092e0b01fb5a243ed6eb4c282", "submitted_at": "2025-12-02T05:11:23Z", "updated_at": "2025-12-02T05:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.02423v1"}, "authors": ["Haolong Yan", "Yeqing Shen", "Xin Huang", "Jia Wang", "Kaijun Tan", "Zhixuan Liang", "Hongxin Li", "Zheng Ge", "Osamu Yoshie", "Si Li", "Xiangyu Zhang", "Daxin Jiang"], "canonical_source_url": "https://arxiv.org/abs/2512.02423", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-02423/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fda844ac6767e6d9dfd16919042469e2070ec7f092e0b01fb5a243ed6eb4c282", "submitted_at": "2025-12-02T05:11:23Z", "updated_at": "2025-12-02T05:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.02423v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2512.02423"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.02423"}], "name": "GUI Exploration Lab: Enhancing Screen Navigation in Agents via Multi-Turn Reinforcement Learning", "publication": null, "published_at": "2025-12-02", "record_sha256": "360556634de52cd132285d8006b812bac0492d6db15a00d880e3323319ad024c", "relations": [], "resource_id": "arxiv:2512.02423", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.02423v1", "topics": ["reinforcement-learning", "evaluation", "data-generation"], "updated_at": "2025-12-02"}, "text": "GUI Exploration Lab: Enhancing Screen Navigation in Agents via Multi-Turn Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI Exploration Lab, a simulation engine for composing screens, interface elements, and navigation graphs. It studies how supervised learning, single-turn RL, and multiturn RL affect GUI navigation and exploration.\nPlatforms: unspecified\nTopics: reinforcement-learning, evaluation, data-generation\nPublished: 2025-12-02\nAuthors: Haolong Yan, Yeqing Shen, Xin Huang, Jia Wang, Kaijun Tan, Zhixuan Liang, Hongxin Li, Zheng Ge, Osamu Yoshie, Si Li, Xiangyu Zhang, Daxin Jiang\nRecorded arXiv version: v1\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.02423v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-02423/"}
{"id": "arxiv:2512.04105", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d362a839fd1ee917ebdde6fcd794640523e930cd711769154532200903fe7de5", "submitted_at": "2025-11-28T07:26:29Z", "updated_at": "2025-11-28T07:26:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.04105v1"}, "authors": ["Jinzhe Tan", "Karim Benyekhlef"], "canonical_source_url": "https://arxiv.org/abs/2512.04105", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-04105/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d362a839fd1ee917ebdde6fcd794640523e930cd711769154532200903fe7de5", "submitted_at": "2025-11-28T07:26:29Z", "updated_at": "2025-11-28T07:26:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.04105v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.04105"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.04105"}], "name": "LegalWebAgent: Empowering Access to Justice via LLM-Based Web Agents", "publication": null, "published_at": "2025-11-28", "record_sha256": "a57213947e71ce402071da682a7579e50010dcf508247d5fec5099a2d88747d9", "relations": [], "resource_id": "arxiv:2512.04105", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.04105v1", "topics": ["industry", "web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2025-11-28"}, "text": "LegalWebAgent: Empowering Access to Justice via LLM-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nLegalWebAgent separates user-needs clarification, webpage browsing, and actions such as form completion or scheduling. A 15-task evaluation studies legal-service workflows relevant to Québec civil-law users, using HTML and screenshot observations.\nPlatforms: web\nTopics: industry, web-navigation, human-in-the-loop, evaluation\nPublished: 2025-11-28\nAuthors: Jinzhe Tan, Karim Benyekhlef\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.04105v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-04105/"}
{"id": "arxiv:2512.05941", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac971bd0e7cd946f81e28d5976eb933cbb95b97ca846ee624a0c36b48760de7", "submitted_at": "2025-12-05T18:39:12Z", "updated_at": "2025-12-05T18:39:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.05941v1"}, "authors": ["Zhiyuan Jiang", "Shenghao Xie", "Wenyi Li", "Wenqiang Zu", "Peihang Li", "Jiahao Qiu", "Siqi Pei", "Lei Ma", "Tiejun Huang", "Mengdi Wang", "Shilong Liu"], "canonical_source_url": "https://arxiv.org/abs/2512.05941", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-05941/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac971bd0e7cd946f81e28d5976eb933cbb95b97ca846ee624a0c36b48760de7", "submitted_at": "2025-12-05T18:39:12Z", "updated_at": "2025-12-05T18:39:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.05941v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.05941"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.05941"}], "name": "Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI Grounding", "publication": null, "published_at": "2025-12-05", "record_sha256": "7a6aa63d0beed2a3acab732a7bfd44f66f708659f7f1ef51d133bef3765e7941", "relations": [], "resource_id": "arxiv:2512.05941", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.05941v1", "topics": ["grounding", "training-free", "evaluation", "screen-understanding"], "updated_at": "2025-12-05"}, "text": "Zoom in, Click out: Unlocking and Evaluating the Potential of Zooming for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nZoomClick uses repeated zooming and context adjustment to improve GUI grounding without training. The paper studies zoom depth, crop sizes, and related choices and introduces GUIZoom-Bench to evaluate how grounding models respond to zoomed views.\nPlatforms: unspecified\nTopics: grounding, training-free, evaluation, screen-understanding\nPublished: 2025-12-05\nAuthors: Zhiyuan Jiang, Shenghao Xie, Wenyi Li, Wenqiang Zu, Peihang Li, Jiahao Qiu, Siqi Pei, Lei Ma, Tiejun Huang, Mengdi Wang, Shilong Liu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.05941v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-05941/"}
{"id": "arxiv:2512.07725", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2576701bbdc4fd8f1f571ed691b06185d1e16bf4955a5c99b10200d796ba6525", "submitted_at": "2025-12-08T17:16:12Z", "updated_at": "2025-12-08T17:16:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.07725v1"}, "authors": ["Alisha Ukani", "Hamed Haddadi", "Ali Shahin Shamsabadi", "Peter Snyder"], "canonical_source_url": "https://arxiv.org/abs/2512.07725", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-07725/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2576701bbdc4fd8f1f571ed691b06185d1e16bf4955a5c99b10200d796ba6525", "submitted_at": "2025-12-08T17:16:12Z", "updated_at": "2025-12-08T17:16:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.07725v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.07725"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.07725"}], "name": "Privacy Practices of Browser Agents", "publication": null, "published_at": "2025-12-08", "record_sha256": "6500e8cfb739adaf8fe4ae7886ac5bb8ee2f99fd49ee42928f125175ef4a80d0", "relations": [], "resource_id": "arxiv:2512.07725", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.07725v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2025-12-08"}, "text": "Privacy Practices of Browser Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a privacy evaluation framework for browser agents covering component weaknesses, website defenses, tracking, privacy-related prompts and personal-data leakage. The framework is applied to deployed browser-agent tools.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2025-12-08\nAuthors: Alisha Ukani, Hamed Haddadi, Ali Shahin Shamsabadi, Peter Snyder\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced. The abstract describes artifact release as planned; the annotation does not independently verify availability.\n\nPrimary source: https://arxiv.org/abs/2512.07725v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-07725/"}
{"id": "arxiv:2512.08529", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95f6e22cda1084d32bbd3f102ea780bfbf392ee578b7c20542314ef76618a56b", "submitted_at": "2025-12-09T12:19:00Z", "updated_at": "2025-12-09T12:19:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.08529v1"}, "authors": ["Yunzhu Zhang", "Zeyu Pan", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "canonical_source_url": "https://arxiv.org/abs/2512.08529", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-08529/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "95f6e22cda1084d32bbd3f102ea780bfbf392ee578b7c20542314ef76618a56b", "submitted_at": "2025-12-09T12:19:00Z", "updated_at": "2025-12-09T12:19:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.08529v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.08529"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.08529"}], "name": "MVP: Multiple View Prediction Improves GUI Grounding", "publication": null, "published_at": "2025-12-09", "record_sha256": "7bd7ae61f9e25ce6dcd0e0c1795554209507c9c858dd2d8684be01a840deb032", "relations": [], "resource_id": "arxiv:2512.08529", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.08529v1", "topics": ["grounding", "training-free", "robustness"], "updated_at": "2025-12-09"}, "text": "MVP: Multiple View Prediction Improves GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nMVP addresses unstable GUI coordinates by predicting across several attention-guided cropped views. It clusters the resulting coordinates and selects the densest cluster centroid, evaluating this training-free inference method across grounding models.\nPlatforms: unspecified\nTopics: grounding, training-free, robustness\nPublished: 2025-12-09\nAuthors: Yunzhu Zhang, Zeyu Pan, Zhengwen Zeng, Shuheng Shen, Changhua Meng, Linchao Zhu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.08529v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-08529/"}
{"id": "arxiv:2512.09396", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49e76d19d0477bcc56f19e8170fed7e9cc80f9c4bafb5f03af3a2d172cb113bc", "submitted_at": "2025-12-10T07:40:23Z", "updated_at": "2025-12-10T07:40:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.09396v1"}, "authors": ["Zishu Wei", "Qixiang Ma", "Xavier Hu", "Yuhang Liu", "Hui Zang", "Yudong Zhao", "Tao Wang", "Shengyu Zhang", "Fei Wu"], "canonical_source_url": "https://arxiv.org/abs/2512.09396", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-09396/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49e76d19d0477bcc56f19e8170fed7e9cc80f9c4bafb5f03af3a2d172cb113bc", "submitted_at": "2025-12-10T07:40:23Z", "updated_at": "2025-12-10T07:40:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.09396v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2512.09396"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.09396"}], "name": "GAIR: GUI Automation via Information-Joint Reasoning and Group Reflection", "publication": null, "published_at": "2025-12-10", "record_sha256": "d3f8903e633c5f970018a3beafe9769c7df6e0fc4dc92f5b57ae834c2accf3d5", "relations": [], "resource_id": "arxiv:2512.09396", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.09396v1", "topics": ["planning", "self-reflection", "automation"], "updated_at": "2025-12-10"}, "text": "GAIR: GUI Automation via Information-Joint Reasoning and Group Reflection\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GAIR, which combines information from specialized GUI models through a general-purpose coordinating model. When the available information is insufficient, group reflection sends targeted requests back to the specialists before selecting an operation.\nPlatforms: unspecified\nTopics: planning, self-reflection, automation\nPublished: 2025-12-10\nAuthors: Zishu Wei, Qixiang Ma, Xavier Hu, Yuhang Liu, Hui Zang, Yudong Zhao, Tao Wang, Shengyu Zhang, Fei Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.09396v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-09396/"}
{"id": "arxiv:2512.10371", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72b2db237a52ae31eaba58183a2777e79379aa8e849d18de5d8633daf27ccbf9", "submitted_at": "2025-12-11T07:37:38Z", "updated_at": "2026-05-08T09:40:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.10371v2"}, "authors": ["Shizuo Tian", "Hao Wen", "Yuxuan Chen", "Jiacheng Liu", "Shanhui Zhao", "Guohong Liu", "Ju Ren", "Yunxin Liu", "Yuanchun Li"], "canonical_source_url": "https://arxiv.org/abs/2512.10371", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-10371/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "72b2db237a52ae31eaba58183a2777e79379aa8e849d18de5d8633daf27ccbf9", "submitted_at": "2025-12-11T07:37:38Z", "updated_at": "2026-05-08T09:40:02Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.10371v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.10371"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.10371"}], "name": "AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management", "publication": null, "published_at": "2025-12-11", "record_sha256": "cd50553af816ed4625aca1962a56ad784139eb6920d8c7275c9cc13cbd915547", "relations": [], "resource_id": "arxiv:2512.10371", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.10371v2", "topics": ["interaction-history", "memory", "long-horizon"], "updated_at": "2026-05-08"}, "text": "AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AgentProg, organizing mobile-agent interaction history as a program with variables and control flow to decide what context to retain. A global belief state handles partial observability, with evaluation on AndroidWorld and extended long-horizon tasks.\nPlatforms: android\nTopics: interaction-history, memory, long-horizon\nPublished: 2025-12-11\nAuthors: Shizuo Tian, Hao Wen, Yuxuan Chen, Jiacheng Liu, Shanhui Zhao, Guohong Liu, Ju Ren, Yunxin Liu, Yuanchun Li\nRecorded arXiv version: v2\n\nReviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.10371v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-10371/"}
{"id": "arxiv:2512.10962", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7ccc7923f4e6f2ea296b4a6bfc28ba4dc7a52ed256008f153f13fccac3f018f", "submitted_at": "2025-11-22T23:12:56Z", "updated_at": "2026-02-04T19:36:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.10962v3"}, "authors": ["Yifei He", "Pranit Chawla", "Yaser Souri", "Subhojit Som", "Xia Song"], "canonical_source_url": "https://arxiv.org/abs/2512.10962", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-10962/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7ccc7923f4e6f2ea296b4a6bfc28ba4dc7a52ed256008f153f13fccac3f018f", "submitted_at": "2025-11-22T23:12:56Z", "updated_at": "2026-02-04T19:36:03Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.10962v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.10962"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.10962"}], "name": "WebSTAR: Scalable Data Synthesis for Computer Use Agents with Step-Level Filtering", "publication": null, "published_at": "2025-11-22", "record_sha256": "edfa1a84dadd2883c30ab0ed643f5d1c45cbc9ab6c9dca18885a1bbfaf657e5c", "relations": [], "resource_id": "arxiv:2512.10962", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.10962v3", "topics": ["data-generation", "reward-models", "imitation-learning"], "updated_at": "2026-02-04"}, "text": "WebSTAR: Scalable Data Synthesis for Computer Use Agents with Step-Level Filtering\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a pipeline that filters individual steps from noisy computer-agent rollouts and augments them with reasoning for training. WebSTAR and WebSCORE support GUI-policy supervision and a distilled process reward model, StepRM.\nPlatforms: web\nTopics: data-generation, reward-models, imitation-learning\nPublished: 2025-11-22\nAuthors: Yifei He, Pranit Chawla, Yaser Souri, Subhojit Som, Xia Song\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.10962v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-10962/"}
{"id": "arxiv:2512.11611", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "484a51656db4466d5e8630d64fa78d3cf10043568488f5a2e9bae679fa800cbe", "submitted_at": "2025-12-12T14:49:32Z", "updated_at": "2025-12-12T14:49:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.11611v1"}, "authors": ["Chunyi Li", "Longfei Li", "Zicheng Zhang", "Xiaohong Liu", "Min Tang", "Weisi Lin", "Guangtao Zhai"], "canonical_source_url": "https://arxiv.org/abs/2512.11611", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-11611/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "484a51656db4466d5e8630d64fa78d3cf10043568488f5a2e9bae679fa800cbe", "submitted_at": "2025-12-12T14:49:32Z", "updated_at": "2025-12-12T14:49:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.11611v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.11611"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.11611"}], "name": "Using GUI Agent for Electronic Design Automation", "publication": null, "published_at": "2025-12-12", "record_sha256": "d28a831b53f569f03f8f5066063dbb1adef1fe9a766af376859a8d75c041b0a9", "relations": [], "resource_id": "arxiv:2512.11611", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.11611v1", "topics": ["industry", "data-generation", "evaluation", "workflow"], "updated_at": "2025-12-12"}, "text": "Using GUI Agent for Electronic Design Automation\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-EDA collects more than 2,000 screenshot-answer-action pairs from engineers working across five CAD tools and five physical domains. The accompanying benchmark evaluates more than 30 GUI agents on specialized electronic-design workflows.\nPlatforms: unspecified\nTopics: industry, data-generation, evaluation, workflow\nPublished: 2025-12-12\nAuthors: Chunyi Li, Longfei Li, Zicheng Zhang, Xiaohong Liu, Min Tang, Weisi Lin, Guangtao Zhai\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.11611v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-11611/"}
{"id": "arxiv:2512.12594", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa6b4d9fa34610ec6fa8c0ac8b075127365aad57a20d814a2e838a223cf8bd56", "submitted_at": "2025-12-14T08:25:31Z", "updated_at": "2026-03-04T01:49:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.12594v2"}, "authors": ["Luoxi Meng", "Henry Feng", "Ilia Shumailov", "Earlence Fernandes"], "canonical_source_url": "https://arxiv.org/abs/2512.12594", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-12594/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa6b4d9fa34610ec6fa8c0ac8b075127365aad57a20d814a2e838a223cf8bd56", "submitted_at": "2025-12-14T08:25:31Z", "updated_at": "2026-03-04T01:49:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.12594v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "identifiers": {"arxiv": "2512.12594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.12594"}], "name": "ceLLMate: Sandboxing Browser AI Agents", "publication": null, "published_at": "2025-12-14", "record_sha256": "0392f3d1f89019ef7f684066fa64b2ecdd421e23cce3513c82f45150d34a2196", "relations": [], "resource_id": "arxiv:2512.12594", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.12594v2", "topics": ["safety-security", "sandbox", "web-navigation", "robustness", "evaluation"], "updated_at": "2026-03-04"}, "text": "ceLLMate: Sandboxing Browser AI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nceLLMate constrains browser-agent authority through an extension that applies policies at the HTTP layer. It evaluates this sandbox against prompt-injection attacks in WASP, avoiding dependence on brittle low-level click and keystroke policies.\nPlatforms: web\nTopics: safety-security, sandbox, web-navigation, robustness, evaluation\nPublished: 2025-12-14\nAuthors: Luoxi Meng, Henry Feng, Ilia Shumailov, Earlence Fernandes\nRecorded arXiv version: v2\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2512.12594v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-12594/"}
{"id": "arxiv:2512.12634", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5982901598e43dcadf63b4a950b2fff68e85be71d0f483648b535d82d7745bd7", "submitted_at": "2025-12-14T10:41:39Z", "updated_at": "2026-05-31T23:48:10Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2512.12634v4"}, "authors": ["Youngmin Im", "Byeongung Jo", "Jaeyoung Wi", "Seungwoo Baek", "Tae Hoon Min", "Joo Hyung Lee", "Sangeun Oh", "Insik Shin", "Sunjae Lee"], "canonical_source_url": "https://arxiv.org/abs/2512.12634", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-12634/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5982901598e43dcadf63b4a950b2fff68e85be71d0f483648b535d82d7745bd7", "submitted_at": "2025-12-14T10:41:39Z", "updated_at": "2026-05-31T23:48:10Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2512.12634v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.12634"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.12634"}], "name": "MobiBench: Multi-Branch, Modular Benchmark for Mobile GUI Agents", "publication": null, "published_at": "2025-12-14", "record_sha256": "21778e9fb102491ea87fcdbcd9b5d3edff3ef153fb27a5622f651ddfd5770d9c", "relations": [], "resource_id": "arxiv:2512.12634", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.12634v4", "topics": ["evaluation", "mobile-navigation", "observability"], "updated_at": "2026-05-31"}, "text": "MobiBench: Multi-Branch, Modular Benchmark for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMobiBench evaluates mobile GUI agents offline while allowing multiple valid action paths. It decomposes agent pipelines into modules to study component contributions, configurations, and performance bottlenecks.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, observability\nPublished: 2025-12-14\nAuthors: Youngmin Im, Byeongung Jo, Jaeyoung Wi, Seungwoo Baek, Tae Hoon Min, Joo Hyung Lee, Sangeun Oh, Insik Shin, Sunjae Lee\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.12634v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-12634/"}
{"id": "arxiv:2512.14014", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6228822cf64c6aeb1836bee52a6515123bd0d44d77cceb23e785bb4ff9aef47", "submitted_at": "2025-12-16T02:16:42Z", "updated_at": "2026-10-07T23:06:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.14014v2"}, "authors": ["Shufan Li", "Konstantinos Kallidromitis", "Akash Gokul", "Yusuke Kato", "Kazuki Kozuka", "Aditya Grover"], "canonical_source_url": "https://arxiv.org/abs/2512.14014", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-14014/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6228822cf64c6aeb1836bee52a6515123bd0d44d77cceb23e785bb4ff9aef47", "submitted_at": "2025-12-16T02:16:42Z", "updated_at": "2026-10-07T23:06:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.14014v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.14014"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.14014"}], "name": "MobileWorldBench: Towards Semantic World Modeling For Mobile Agents", "publication": null, "published_at": "2025-12-16", "record_sha256": "5b39891720ed55119c48f196b83cd9f1ddc3e24cf60d7616a9ffc9ebf1b29e99", "relations": [], "resource_id": "arxiv:2512.14014", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.14014v2", "topics": ["world-models", "planning", "evaluation"], "updated_at": "2026-10-07"}, "text": "MobileWorldBench: Towards Semantic World Modeling For Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates semantic world models for mobile interfaces and studies natural-language state transitions as a tool for agent planning.\nPlatforms: android\nTopics: world-models, planning, evaluation\nPublished: 2025-12-16\nAuthors: Shufan Li, Konstantinos Kallidromitis, Akash Gokul, Yusuke Kato, Kazuki Kozuka, Aditya Grover\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.14014v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-14014/"}
{"id": "arxiv:2512.15431", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fc824bdeabca3451fd895447a0321c507da1ac7bc7cc24184c7768992549daf", "submitted_at": "2025-12-17T13:26:30Z", "updated_at": "2025-12-19T17:36:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.15431v2"}, "authors": ["Haolong Yan", "Jia Wang", "Xin Huang", "Yeqing Shen", "Ziyang Meng", "Zhimin Fan", "Kaijun Tan", "Jin Gao", "Lieyu Shi", "Mi Yang", "Shiliang Yang", "Zhirui Wang", "Brian Li", "Kang An", "Chenyang Li", "Lei Lei", "Mengmeng Duan", "Danxun Liang", "Guodong Liu", "Hang Cheng", "Hao Wu", "Jie Dong", "Junhao Huang", "Mei Chen", "Renjie Yu", "Shunshan Li", "Xu Zhou", "Yiting Dai", "Yineng Deng", "Yingdan Liang", "Zelin Chen", "Wen Sun", "Chengxu Yan", "Chunqin Xu", "Dong Li", "Fengqiong Xiao", "Guanghao Fan", "Guopeng Li", "Guozhen Peng", "Hongbing Li", "Hang Li", "Hongming Chen", "Jingjing Xie", "Jianyong Li", "Jingyang Zhang", "Jiaju Ren", "Jiayu Yuan", "Jianpeng Yin", "Kai Cao", "Liang Zhao", "Liguo Tan", "Liying Shi", "Mengqiang Ren", "Min Xu", "Manjiao Liu", "Mao Luo", "Mingxin Wan", "Na Wang", "Nan Wu", "Ning Wang", "Peiyao Ma", "Qingzhou Zhang", "Qiao Wang", "Qinlin Zeng", "Qiong Gao", "Qiongyao Li", "Shangwu Zhong", "Shuli Gao", "Shaofan Liu", "Shisi Gao", "Shuang Luo", "Xingbin Liu", "Xiaojia Liu", "Xiaojie Hou", "Xin Liu", "Xuanti Feng", "Xuedan Cai", "Xuan Wen", "Xianwei Zhu", "Xin Liang", "Xin Liu", "Xin Zhou", "Yifan Sui", "Yingxiu Zhao", "Yukang Shi", "Yunfang Xu", "Yuqing Zeng", "Yixun Zhang", "Zejia Weng", "Zhonghao Yan", "Zhiguo Huang", "Zhuoyu Wang", "Zihan Yan", "Zheng Ge", "Jing Li", "Yibo Zhu", "Binxing Jiao", "Xiangyu Zhang", "Daxin Jiang"], "canonical_source_url": "https://arxiv.org/abs/2512.15431", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-15431/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fc824bdeabca3451fd895447a0321c507da1ac7bc7cc24184c7768992549daf", "submitted_at": "2025-12-17T13:26:30Z", "updated_at": "2025-12-19T17:36:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2512.15431v2"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.15431"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.15431"}], "name": "Step-GUI Technical Report", "publication": null, "published_at": "2025-12-17", "record_sha256": "6860fa3a736a6a414956e74aae77d6721b36a4c67c8bb99649d31b81c9687d34", "relations": [], "resource_id": "arxiv:2512.15431", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.15431v2", "topics": ["data-generation", "reward-models", "tool-integration", "evaluation", "mobile-navigation"], "updated_at": "2025-12-19"}, "text": "Step-GUI Technical Report\nResource type: paper\nReview status: curated\nScope: core\n\nStep-GUI uses calibrated trajectory rewards to turn generated GUI interaction data into training signals for compact agent models. The report also introduces GUI-MCP for hierarchical GUI operations and local task delegation, and AndroidDaily for static-action and end-to-end mobile evaluation.\nPlatforms: android, desktop\nTopics: data-generation, reward-models, tool-integration, evaluation, mobile-navigation\nPublished: 2025-12-17\nAuthors: Haolong Yan, Jia Wang, Xin Huang, Yeqing Shen, Ziyang Meng, Zhimin Fan, Kaijun Tan, Jin Gao, Lieyu Shi, Mi Yang, Shiliang Yang, Zhirui Wang, Brian Li, Kang An, Chenyang Li, Lei Lei, Mengmeng Duan, Danxun Liang, Guodong Liu, Hang Cheng, Hao Wu, Jie Dong, Junhao Huang, Mei Chen, Renjie Yu, Shunshan Li, Xu Zhou, Yiting Dai, Yineng Deng, Yingdan Liang, Zelin Chen, Wen Sun, Chengxu Yan, Chunqin Xu, Dong Li, Fengqiong Xiao, Guanghao Fan, Guopeng Li, Guozhen Peng, Hongbing Li, Hang Li, Hongming Chen, Jingjing Xie, Jianyong Li, Jingyang Zhang, Jiaju Ren, Jiayu Yuan, Jianpeng Yin, Kai Cao, Liang Zhao, Liguo Tan, Liying Shi, Mengqiang Ren, Min Xu, Manjiao Liu, Mao Luo, Mingxin Wan, Na Wang, Nan Wu, Ning Wang, Peiyao Ma, Qingzhou Zhang, Qiao Wang, Qinlin Zeng, Qiong Gao, Qiongyao Li, Shangwu Zhong, Shuli Gao, Shaofan Liu, Shisi Gao, Shuang Luo, Xingbin Liu, Xiaojia Liu, Xiaojie Hou, Xin Liu, Xuanti Feng, Xuedan Cai, Xuan Wen, Xianwei Zhu, Xin Liang, Xin Liu, Xin Zhou, Yifan Sui, Yingxiu Zhao, Yukang Shi, Yunfang Xu, Yuqing Zeng, Yixun Zhang, Zejia Weng, Zhonghao Yan, Zhiguo Huang, Zhuoyu Wang, Zihan Yan, Zheng Ge, Jing Li, Yibo Zhu, Binxing Jiao, Xiangyu Zhang, Daxin Jiang\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.15431v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-15431/"}
{"id": "arxiv:2512.16295", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67a1bd40b117b34e5fb994b1c2c93cd285714097d1bc9730baeeaba6c31a7b8a", "submitted_at": "2025-12-18T08:29:50Z", "updated_at": "2025-12-18T08:29:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16295v1"}, "authors": ["Zhenyu Wu", "Jingjing Xie", "Zehao Li", "Bowen Yang", "Qiushi Sun", "Zhaoyang Liu", "Zhoumianze Liu", "Yu Qiao", "Xiangyu Yue", "Zun Wang", "Zichen Ding"], "canonical_source_url": "https://arxiv.org/abs/2512.16295", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-16295/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67a1bd40b117b34e5fb994b1c2c93cd285714097d1bc9730baeeaba6c31a7b8a", "submitted_at": "2025-12-18T08:29:50Z", "updated_at": "2025-12-18T08:29:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16295v1"}}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.16295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.16295"}], "name": "OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models", "publication": null, "published_at": "2025-12-18", "record_sha256": "7405146cbf2a0f246629d1012c5f5e357f88bbfe0b6548316dda02a249d92d57", "relations": [], "resource_id": "arxiv:2512.16295", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.16295v1", "topics": ["reward-models", "reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2025-12-18"}, "text": "OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models\nResource type: paper\nReview status: curated\nScope: core\n\nOS-Oracle synthesizes GUI critic examples and trains a pre-action critic with supervised learning followed by consistency-preserving GRPO. It introduces OS-Critic Bench across mobile, web, and desktop settings and tests critic-assisted agents in OSWorld and AndroidWorld.\nPlatforms: web, desktop, android\nTopics: reward-models, reinforcement-learning, data-generation, evaluation\nPublished: 2025-12-18\nAuthors: Zhenyu Wu, Jingjing Xie, Zehao Li, Bowen Yang, Qiushi Sun, Zhaoyang Liu, Zhoumianze Liu, Yu Qiao, Xiangyu Yue, Zun Wang, Zichen Ding\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.16295v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-16295/"}
{"id": "arxiv:2512.16501", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bc8a1da3993ead9b36af47a9332fe9a1a2427101ad0d91da654b4a4b25141df", "submitted_at": "2025-12-18T13:09:09Z", "updated_at": "2025-12-18T13:09:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16501v1"}, "authors": ["Beitong Zhou", "Zhexiao Huang", "Yuan Guo", "Zhangxuan Gu", "Tianyu Xia", "Zichen Luo", "Fei Tang", "Dehan Kong", "Yanyi Shang", "Suling Ou", "Zhenlin Guo", "Changhua Meng", "Shuheng Shen"], "canonical_source_url": "https://arxiv.org/abs/2512.16501", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-16501/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4bc8a1da3993ead9b36af47a9332fe9a1a2427101ad0d91da654b4a4b25141df", "submitted_at": "2025-12-18T13:09:09Z", "updated_at": "2025-12-18T13:09:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.16501v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.16501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.16501"}], "name": "VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks", "publication": null, "published_at": "2025-12-18", "record_sha256": "f02589d966e469763e23fe7097e0fab03f3393316aaff174c505e51250fcbd3a", "relations": [], "resource_id": "arxiv:2512.16501", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.16501v1", "topics": ["grounding", "evaluation", "data-generation", "robustness"], "updated_at": "2025-12-18"}, "text": "VenusBench-GD: A Comprehensive Multi-Platform GUI Benchmark for Diverse Grounding Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nVenusBench-GD is a bilingual, cross-platform GUI grounding benchmark organized into basic and advanced tasks. Its annotation pipeline and six-subtask taxonomy support comparisons of general multimodal models and GUI-specialized models across different grounding demands.\nPlatforms: cross-platform\nTopics: grounding, evaluation, data-generation, robustness\nPublished: 2025-12-18\nAuthors: Beitong Zhou, Zhexiao Huang, Yuan Guo, Zhangxuan Gu, Tianyu Xia, Zichen Luo, Fei Tang, Dehan Kong, Yanyi Shang, Suling Ou, Zhenlin Guo, Changhua Meng, Shuheng Shen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.16501v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-16501/"}
{"id": "arxiv:2512.17221", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cd0aa99ed8bad167e25f6e2b48e7cd1cf01249508555f657215213ee94a301", "submitted_at": "2025-12-19T04:09:24Z", "updated_at": "2025-12-31T17:30:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.17221v3"}, "authors": ["Brandon Huang", "Hang Hua", "Zhuoran Yu", "Trevor Darrell", "Rogerio Feris", "Roei Herzig"], "canonical_source_url": "https://arxiv.org/abs/2512.17221", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-17221/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cd0aa99ed8bad167e25f6e2b48e7cd1cf01249508555f657215213ee94a301", "submitted_at": "2025-12-19T04:09:24Z", "updated_at": "2025-12-31T17:30:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.17221v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.17221"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.17221"}], "name": "DAVE: A VLM Vision Encoder for Document Understanding and Web Agents", "publication": null, "published_at": "2025-12-19", "record_sha256": "2c63f7eb0436de428c3456742f612a70bcee569f4a1f007c39dcd186d9964661", "relations": [], "resource_id": "arxiv:2512.17221", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.17221v3", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2025-12-31"}, "text": "DAVE: A VLM Vision Encoder for Document Understanding and Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDAVE trains a vision encoder for document images and web agents using self-supervised pretraining followed by supervised parsing and localization. It combines decoder-specific encoder merging with feature fusion and evaluates web localization and agent benchmarks alongside document tasks.\nPlatforms: web\nTopics: screen-understanding, grounding, evaluation\nPublished: 2025-12-19\nAuthors: Brandon Huang, Hang Hua, Zhuoran Yu, Trevor Darrell, Rogerio Feris, Roei Herzig\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.17221v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-17221/"}
{"id": "arxiv:2512.19396", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03cdf8514fa9f036b4614a880d842d49bb9cb0a821858b7215784fba209faa62", "submitted_at": "2025-12-22T13:42:18Z", "updated_at": "2026-04-10T06:20:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19396v3"}, "authors": ["Runze Li", "Yuwen Zhai", "Bo Xu", "LiWu Xu", "Nian Shi", "Wei Zhang", "Ran Lin", "Liang Wang"], "canonical_source_url": "https://arxiv.org/abs/2512.19396", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-19396/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03cdf8514fa9f036b4614a880d842d49bb9cb0a821858b7215784fba209faa62", "submitted_at": "2025-12-22T13:42:18Z", "updated_at": "2026-04-10T06:20:00Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19396v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.19396"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.19396"}], "name": "EchoTrail-GUI: Building Actionable Memory for GUI Agents via Critic-Guided Self-Exploration", "publication": null, "published_at": "2025-12-22", "record_sha256": "51bd80c3784117288eeec5542e965226fd7d4b7238634a087e912dbf92768427", "relations": [], "resource_id": "arxiv:2512.19396", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.19396v3", "topics": ["memory", "data-generation", "reward-models", "mobile-navigation"], "updated_at": "2026-04-10"}, "text": "EchoTrail-GUI: Building Actionable Memory for GUI Agents via Critic-Guided Self-Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nEchoTrail-GUI builds a memory of successful GUI trajectories through autonomous exploration filtered by a reward model. It retrieves relevant trajectories as in-context guidance for new tasks and evaluates the approach on AndroidWorld and AndroidLab.\nPlatforms: android\nTopics: memory, data-generation, reward-models, mobile-navigation\nPublished: 2025-12-22\nAuthors: Runze Li, Yuwen Zhai, Bo Xu, LiWu Xu, Nian Shi, Wei Zhang, Ran Lin, Liang Wang\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.19396v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-19396/"}
{"id": "arxiv:2512.19432", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cf3258fd2286b0ee6babb69b9bfc459ecf40536eec498a3af3403d2971791f3d", "submitted_at": "2025-12-22T14:31:28Z", "updated_at": "2025-12-30T07:15:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19432v3"}, "authors": ["Quyu Kong", "Xu Zhang", "Zhenyu Yang", "Nolan Gao", "Chen Liu", "Panrong Tong", "Chenglin Cai", "Hanzhang Zhou", "Jianan Zhang", "Liangyu Chen", "Zhidan Liu", "Steven Hoi", "Yue Wang"], "canonical_source_url": "https://arxiv.org/abs/2512.19432", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-19432/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cf3258fd2286b0ee6babb69b9bfc459ecf40536eec498a3af3403d2971791f3d", "submitted_at": "2025-12-22T14:31:28Z", "updated_at": "2025-12-30T07:15:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.19432v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.19432"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.19432"}], "name": "MobileWorld: Benchmarking Autonomous Mobile Agents in Agent-User Interactive and MCP-Augmented Environments", "publication": null, "published_at": "2025-12-22", "record_sha256": "19322b7e1d31a6166613a25b39bc8ae6752bf5b9ac5ad99e60e0e0b0ca9b4cc8", "relations": [], "resource_id": "arxiv:2512.19432", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.19432v3", "topics": ["evaluation", "tool-integration", "long-horizon"], "updated_at": "2025-12-30"}, "text": "MobileWorld: Benchmarking Autonomous Mobile Agents in Agent-User Interactive and MCP-Augmented Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces long mobile workflows that combine agent-user interaction, cross-application tasks, and MCP-based tool access.\nPlatforms: android\nTopics: evaluation, tool-integration, long-horizon\nPublished: 2025-12-22\nAuthors: Quyu Kong, Xu Zhang, Zhenyu Yang, Nolan Gao, Chen Liu, Panrong Tong, Chenglin Cai, Hanzhang Zhou, Jianan Zhang, Liangyu Chen, Zhidan Liu, Steven Hoi, Yue Wang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.19432v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-19432/"}
{"id": "arxiv:2512.21302", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "597935923ce9e59720771f3d92e921f2842f908ad81b3ab31a87ed9d9f5c5774", "submitted_at": "2025-12-24T17:40:42Z", "updated_at": "2025-12-24T17:40:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.21302v1"}, "authors": ["Yue Cao", "Yingyao Wang", "Pi Bu", "Jingxuan Xing", "Wei Jiang", "Zekun Zhu", "Junpeng Ma", "Sashuai Zhou", "Tong Lu", "Jun Song", "Yu Cheng", "Yuning Jiang", "Bo Zheng"], "canonical_source_url": "https://arxiv.org/abs/2512.21302", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-21302/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "597935923ce9e59720771f3d92e921f2842f908ad81b3ab31a87ed9d9f5c5774", "submitted_at": "2025-12-24T17:40:42Z", "updated_at": "2025-12-24T17:40:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.21302v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.21302"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.21302"}], "name": "AndroidLens: Long-latency Evaluation with Nested Sub-targets for Android GUI Agents", "publication": null, "published_at": "2025-12-24", "record_sha256": "94a8b5bf55e2b9ee0f97204e7d249b620e32005e503df707ce84e361639dbce7", "relations": [], "resource_id": "arxiv:2512.21302", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.21302v1", "topics": ["evaluation", "long-horizon", "mobile-navigation", "robustness"], "updated_at": "2025-12-24"}, "text": "AndroidLens: Long-latency Evaluation with Nested Sub-targets for Android GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAndroidLens provides 571 extended mobile tasks in Chinese and English, with multiple valid paths and preserved environmental anomalies. Its dynamic evaluation uses nested milestones and Average Task Progress to measure partial completion of long workflows.\nPlatforms: android\nTopics: evaluation, long-horizon, mobile-navigation, robustness\nPublished: 2025-12-24\nAuthors: Yue Cao, Yingyao Wang, Pi Bu, Jingxuan Xing, Wei Jiang, Zekun Zhu, Junpeng Ma, Sashuai Zhou, Tong Lu, Jun Song, Yu Cheng, Yuning Jiang, Bo Zheng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.21302v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-21302/"}
{"id": "arxiv:2512.22009", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "041466768673c1ea49bceda2ad08784c0fa0e896bcee5452a1e41d51ac3b81d8", "submitted_at": "2025-12-26T12:09:15Z", "updated_at": "2025-12-26T12:09:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22009v1"}, "authors": ["Sarthak Mehrotra", "Sairam V C Rebbapragada", "Mani Hemanth Reddy Bonthu", "Vineeth N Balasubramanian"], "canonical_source_url": "https://arxiv.org/abs/2512.22009", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-22009/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "041466768673c1ea49bceda2ad08784c0fa0e896bcee5452a1e41d51ac3b81d8", "submitted_at": "2025-12-26T12:09:15Z", "updated_at": "2025-12-26T12:09:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22009v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.22009"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.22009"}], "name": "iSHIFT: Lightweight Slow-Fast GUI Agent with Adaptive Perception", "publication": null, "published_at": "2025-12-26", "record_sha256": "6244c81183e06fa225ed40f65e0f89babf9fae9e13687b95382b912580f77fa6", "relations": [], "resource_id": "arxiv:2512.22009", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.22009v1", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2025-12-26"}, "text": "iSHIFT: Lightweight Slow-Fast GUI Agent with Adaptive Perception\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces iSHIFT, which combines latent reasoning with perception-control tokens to adapt GUI inference. The agent switches between detailed regional grounding and faster use of global visual cues.\nPlatforms: unspecified\nTopics: efficiency, grounding, screen-understanding\nPublished: 2025-12-26\nAuthors: Sarthak Mehrotra, Sairam V C Rebbapragada, Mani Hemanth Reddy Bonthu, Vineeth N Balasubramanian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.22009v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-22009/"}
{"id": "arxiv:2512.22047", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a89eb3fd173680c127245933d252d4e245b7179d77bca4b463d50800c78f7238", "submitted_at": "2025-12-26T14:51:52Z", "updated_at": "2025-12-26T14:51:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22047v1"}, "authors": ["Hanzhang Zhou", "Xu Zhang", "Panrong Tong", "Jianan Zhang", "Liangyu Chen", "Quyu Kong", "Chenglin Cai", "Chen Liu", "Yue Wang", "Jingren Zhou", "Steven Hoi"], "canonical_source_url": "https://arxiv.org/abs/2512.22047", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-22047/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a89eb3fd173680c127245933d252d4e245b7179d77bca4b463d50800c78f7238", "submitted_at": "2025-12-26T14:51:52Z", "updated_at": "2025-12-26T14:51:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.22047v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.22047"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.22047"}], "name": "MAI-UI Technical Report: Real-World Centric Foundation GUI Agents", "publication": null, "published_at": "2025-12-26", "record_sha256": "da27acd392d824eb741fb7dbbd68bc4fc4daf36d53e7525eba1e5d87a6fc3c42", "relations": [], "resource_id": "arxiv:2512.22047", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.22047v1", "topics": ["human-in-the-loop", "tool-integration", "reinforcement-learning", "data-generation", "efficiency"], "updated_at": "2025-12-26"}, "text": "MAI-UI Technical Report: Real-World Centric Foundation GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMAI-UI combines GUI navigation, native user interaction, and MCP tool calls in a family of policies trained through generated data and online reinforcement learning. The report also studies device-cloud routing and scaling of parallel environments and interaction budgets, with grounding and Android navigation evaluation.\nPlatforms: android\nTopics: human-in-the-loop, tool-integration, reinforcement-learning, data-generation, efficiency\nPublished: 2025-12-26\nAuthors: Hanzhang Zhou, Xu Zhang, Panrong Tong, Jianan Zhang, Liangyu Chen, Quyu Kong, Chenglin Cai, Chen Liu, Yue Wang, Jingren Zhou, Steven Hoi\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained primary abstract. Experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.22047v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-22047/"}
{"id": "arxiv:2512.23128", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a255cccbe3f16f498918126724c391911e1d745c093189790cbab715fd28a2d2", "submitted_at": "2025-12-29T01:09:10Z", "updated_at": "2026-07-02T16:12:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.23128v3"}, "authors": ["Karolina Korgul", "Yushi Yang", "Arkadiusz Drohomirecki", "Piotr Błaszczyk", "Will Howard", "Lukas Aichberger", "Chris Russell", "Philip H. S. Torr", "Adam Mahdi", "Adel Bibi"], "canonical_source_url": "https://arxiv.org/abs/2512.23128", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-23128/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a255cccbe3f16f498918126724c391911e1d745c093189790cbab715fd28a2d2", "submitted_at": "2025-12-29T01:09:10Z", "updated_at": "2026-07-02T16:12:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2512.23128v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2512.23128"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.23128"}], "name": "It's a TRAP! Task-Redirecting Agent Persuasion Benchmark for Web Agents", "publication": null, "published_at": "2025-12-29", "record_sha256": "e977dfaed27ef771ecef41e1a857cc5e5df2f76636d3ce653e850ec391efcdde", "relations": [], "resource_id": "arxiv:2512.23128", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.23128v3", "topics": ["safety-security", "evaluation", "ui-design"], "updated_at": "2026-07-02"}, "text": "It's a TRAP! Task-Redirecting Agent Persuasion Benchmark for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces TRAP, which evaluates task-redirection attacks embedded in realistic web interfaces. Controlled website clones and a modular injection framework support comparisons of persuasion techniques and contextual changes across agents.\nPlatforms: web\nTopics: safety-security, evaluation, ui-design\nPublished: 2025-12-29\nAuthors: Karolina Korgul, Yushi Yang, Arkadiusz Drohomirecki, Piotr Błaszczyk, Will Howard, Lukas Aichberger, Chris Russell, Philip H. S. Torr, Adam Mahdi, Adel Bibi\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.23128v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-23128/"}
{"id": "arxiv:2512.23647", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fde71dd935acb953bc5d01bcd48f0d6c3351db6a66126761baf134c114e7893c", "submitted_at": "2025-12-29T17:59:14Z", "updated_at": "2025-12-29T17:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.23647v1"}, "authors": ["Baixuan Li", "Jialong Wu", "Wenbiao Yin", "Kuan Li", "Zhongwang Zhang", "Huifeng Yin", "Zhengwei Tao", "Liwen Zhang", "Pengjun Xie", "Jingren Zhou", "Yong Jiang"], "canonical_source_url": "https://arxiv.org/abs/2512.23647", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-23647/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fde71dd935acb953bc5d01bcd48f0d6c3351db6a66126761baf134c114e7893c", "submitted_at": "2025-12-29T17:59:14Z", "updated_at": "2025-12-29T17:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.23647v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "identifiers": {"arxiv": "2512.23647"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.23647"}], "name": "Nested Browser-Use Learning for Agentic Information Seeking", "publication": null, "published_at": "2025-12-29", "record_sha256": "c3b8e4c4dad81667e06bba242928202ff480f699e3cc643af1438b4d02382e74", "relations": [], "resource_id": "arxiv:2512.23647", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.23647v1", "topics": ["web-navigation", "planning", "efficiency", "tool-integration"], "updated_at": "2025-12-29"}, "text": "Nested Browser-Use Learning for Agentic Information Seeking\nResource type: paper\nReview status: curated\nScope: core\n\nNestBrowse organizes browser actions in a nested framework that separates interaction control from page exploration. It evaluates this design for deep information seeking beyond snippet retrieval and direct URL fetching.\nPlatforms: web\nTopics: web-navigation, planning, efficiency, tool-integration\nPublished: 2025-12-29\nAuthors: Baixuan Li, Jialong Wu, Wenbiao Yin, Kuan Li, Zhongwang Zhang, Huifeng Yin, Zhengwei Tao, Liwen Zhang, Pengjun Xie, Jingren Zhou, Yong Jiang\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2512.23647v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-23647/"}
{"id": "arxiv:2512.24965", "metadata": {"aliases": ["ShowUI-pi", "ShowUI Pi"], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7d1f358b450f2a6c3deb00f1b16c54d443f48fda3a7dc70c77456bebf35ad7c", "submitted_at": "2025-12-31T16:51:14Z", "updated_at": "2025-12-31T16:51:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.24965v1"}, "authors": ["Siyuan Hu", "Kevin Qinghong Lin", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2512.24965", "citation_url": "https://awesomecua.github.io/resources/arxiv-2512-24965/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b7d1f358b450f2a6c3deb00f1b16c54d443f48fda3a7dc70c77456bebf35ad7c", "submitted_at": "2025-12-31T16:51:14Z", "updated_at": "2025-12-31T16:51:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2512.24965v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}], "identifiers": {"arxiv": "2512.24965"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2512.24965"}], "name": "ShowUI-$π$: Flow-based Generative Models as GUI Dexterous Hands", "publication": null, "published_at": "2025-12-31", "record_sha256": "ac81594d05c5e29f4f1a8b84dcdc2c551cc59ad858da7a67378c8abd18637859", "relations": [], "resource_id": "arxiv:2512.24965", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2512.24965v1", "topics": ["grounding", "evaluation", "data-generation"], "updated_at": "2025-12-31"}, "text": "ShowUI-$π$: Flow-based Generative Models as GUI Dexterous Hands\nResource type: paper\nReview status: curated\nScope: core\n\nModels discrete clicks and continuous drags with flow-based action generation, and introduces drag trajectories and the ScreenDrag benchmark.\nRecorded aliases: ShowUI-pi, ShowUI Pi\nPlatforms: desktop\nTopics: grounding, evaluation, data-generation\nPublished: 2025-12-31\nAuthors: Siyuan Hu, Kevin Qinghong Lin, Mike Zheng Shou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2512.24965v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2512-24965/"}
{"id": "arxiv:2601.02399", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "afab3dcd2444bf8333f2a2fea887819f6fca55c41209d8c449f6211d1c62b819", "submitted_at": "2025-12-30T01:49:46Z", "updated_at": "2025-12-30T01:49:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.02399v1"}, "authors": ["Jiaxin Ai", "Yukang Feng", "Fanrui Zhang", "Jianwen Sun", "Zizhen Li", "Chuanhao Li", "Yifan Chang", "Wenxiao Wu", "Ruoxi Wang", "Mingliang Zhai", "Kaipeng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2601.02399", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-02399/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "afab3dcd2444bf8333f2a2fea887819f6fca55c41209d8c449f6211d1c62b819", "submitted_at": "2025-12-30T01:49:46Z", "updated_at": "2025-12-30T01:49:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.02399v1"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:52Z", "role": "primary", "sha256": "8623b0ff7ef9d2e8ba484cd67359e7b8817e78e9ca22b8f27432c2533f5b90a0", "supports": ["name", "kind", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2510.01670%2C2510.04673%2C2510.10073%2C2510.12200%2C2510.14621%2C2510.16051%2C2510.17790%2C2510.18488%2C2510.18596%2C2510.19488%2C2510.19949%2C2510.20286%2C2510.22009%2C2510.24168%2C2510.24411%2C2510.24563%2C2511.00810%2C2511.04307%2C2511.07332%2C2511.09157%2C2511.12254%2C2511.13087%2C2511.15567%2C2511.16590%2C2511.19663%2C2511.20597%2C2511.20766%2C2512.08529%2C2512.10371%2C2512.10962%2C2512.12634%2C2512.14014%2C2512.15431%2C2512.16295%2C2512.16501%2C2512.19432%2C2512.22894%2C2512.24965%2C2601.02399%2C2601.06031%2C2601.07181%2C2601.07262%2C2601.07779%2C2601.08406%2C2601.09923%2C2601.11044%2C2601.11868%2C2601.12822%2C2601.13060%2C2601.15876&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:36Z", "role": "primary", "sha256": "f9a44f8090369a296a584a9b94c4491662ee25bb8f7239ef75dfd8f775be8887", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2601.02399v1"}], "identifiers": {"arxiv": "2601.02399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.02399"}], "name": "ProSoftArena: Benchmarking Hierarchical Capabilities of Multimodal Agents in Professional Software Environments", "publication": null, "published_at": "2025-12-30", "record_sha256": "0c349ba8e3dca737a59aefd1cafe114fcd526608bd9c658ca2313d7a67687c55", "relations": [], "resource_id": "arxiv:2601.02399", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.02399v1", "topics": ["evaluation", "industry", "science", "cross-app"], "updated_at": "2025-12-30"}, "text": "ProSoftArena: Benchmarking Hierarchical Capabilities of Multimodal Agents in Professional Software Environments\nResource type: paper\nReview status: curated\nScope: core\n\nProSoftArena evaluates 436 professional-software tasks across six disciplines and 13 applications in a Windows 11 virtual-machine environment. Its capability hierarchy, execution-based checks, and human-in-the-loop evaluation cover precise GUI operations and multi-software workflows.\nPlatforms: windows, desktop\nTopics: evaluation, industry, science, cross-app\nPublished: 2025-12-30\nAuthors: Jiaxin Ai, Yukang Feng, Fanrui Zhang, Jianwen Sun, Zizhen Li, Chuanhao Li, Yifan Chang, Wenxiao Wu, Ruoxi Wang, Mingliang Zhai, Kaipeng Zhang\nRecorded arXiv version: v1\n\nThe paper abstract and official full-text sections were reviewed. The full text supports Windows 11 virtual-machine execution, screenshot/state observations, and mouse/keyboard actions. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.02399v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-02399/"}
{"id": "arxiv:2601.02439", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8751c4fae3b4f838eba59ffa8be37a1a5f4fa497dfd69fa3075f8581547f51", "submitted_at": "2026-01-05T09:35:11Z", "updated_at": "2026-05-02T18:45:09Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2601.02439v6"}, "authors": ["Hao Bai", "Alexey Taymanov", "Tong Zhang", "Aviral Kumar", "Spencer Whitehead"], "canonical_source_url": "https://arxiv.org/abs/2601.02439", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-02439/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae8751c4fae3b4f838eba59ffa8be37a1a5f4fa497dfd69fa3075f8581547f51", "submitted_at": "2026-01-05T09:35:11Z", "updated_at": "2026-05-02T18:45:09Z", "version": 6, "versioned_url": "https://arxiv.org/abs/2601.02439v6"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.02439"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.02439"}], "name": "WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks", "publication": null, "published_at": "2026-01-05", "record_sha256": "d368e207c1045f3bda559a905cb2e1315e8dfcd3fee810ad57e84501e78acf42", "relations": [], "resource_id": "arxiv:2601.02439", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.02439v6", "topics": ["web-navigation", "reinforcement-learning", "data-generation", "evaluation", "efficiency"], "updated_at": "2026-05-02"}, "text": "WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nWebGym supplies realistic web tasks with rubric-based evaluation and an asynchronous rollout system for reinforcement learning. It studies how training on an agent’s own interaction traces scales with task coverage and transfers to unseen websites.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, data-generation, evaluation, efficiency\nPublished: 2026-01-05\nAuthors: Hao Bai, Alexey Taymanov, Tong Zhang, Aviral Kumar, Spencer Whitehead\nRecorded arXiv version: v6\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.02439v6\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-02439/"}
{"id": "arxiv:2601.03928", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "518a9699a6c11e6f61b655f4016170c50bf5d949e477134521df1c0214320046", "submitted_at": "2026-01-07T13:48:12Z", "updated_at": "2026-01-07T13:48:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.03928v1"}, "authors": ["Mingyu Ouyang", "Kevin Qinghong Lin", "Mike Zheng Shou", "Hwee Tou Ng"], "canonical_source_url": "https://arxiv.org/abs/2601.03928", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-03928/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "518a9699a6c11e6f61b655f4016170c50bf5d949e477134521df1c0214320046", "submitted_at": "2026-01-07T13:48:12Z", "updated_at": "2026-01-07T13:48:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.03928v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:09Z", "role": "primary", "sha256": "f8cee8b281f54ecb3112bb26b59d166617d33f08f098f44818968a9c9b777ce6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2512.06866%2C2512.07287%2C2512.08769%2C2512.12594%2C2512.12730%2C2512.13159%2C2512.14012%2C2512.16925%2C2512.16962%2C2512.16975%2C2512.21373%2C2512.23647%2C2512.23738%2C2601.01743%2C2601.02046%2C2601.02163%2C2601.02454%2C2601.02702%2C2601.03468%2C2601.03928%2C2601.04795%2C2601.05107%2C2601.05432%2C2601.06411%2C2601.06487%2C2601.06490%2C2601.06565%2C2601.06663%2C2601.07143%2C2601.07375%2C2601.08012%2C2601.08079%2C2601.08225%2C2601.08237%2C2601.08731%2C2601.09113%2C2601.09749%2C2601.09883%2C2601.10378%2C2601.10440%2C2601.11653%2C2601.11895%2C2601.11903%2C2601.11974%2C2601.12042%2C2601.12307%2C2601.12560%2C2601.12641%2C2601.13186%2C2601.13383&max_results=50"}], "identifiers": {"arxiv": "2601.03928"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.03928"}], "name": "FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection", "publication": null, "published_at": "2026-01-07", "record_sha256": "7a615e326e6ef76baa2de296107d78ed40a349ab5808e2d6acc6f3ec81cd038a", "relations": [], "resource_id": "arxiv:2601.03928", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.03928v1", "topics": ["grounding", "efficiency", "screen-understanding"], "updated_at": "2026-01-07"}, "text": "FocusUI: Efficient UI Grounding via Position-Preserving Visual Token Selection\nResource type: paper\nReview status: curated\nScope: core\n\nFocusUI selects instruction-relevant screenshot patches while preserving their positional structure for GUI grounding. Its PosPad scheme replaces runs of dropped tokens with positional markers and is evaluated under reduced token budgets.\nPlatforms: unspecified\nTopics: grounding, efficiency, screen-understanding\nPublished: 2026-01-07\nAuthors: Mingyu Ouyang, Kevin Qinghong Lin, Mike Zheng Shou, Hwee Tou Ng\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2601.03928v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-03928/"}
{"id": "arxiv:2601.04035", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c40e3db3f538f089c19d7c8233ea01174d0591c3211a7acc412ee4df28e902a8", "submitted_at": "2026-01-07T15:51:44Z", "updated_at": "2026-08-31T13:27:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.04035v2"}, "authors": ["Yilin Cao", "Yufeng Zhong", "Zhixiong Zeng", "Siran Dai", "Liming Zheng", "Jing Huang", "Haibo Qiu", "Peng Shi", "Wenji Mao"], "canonical_source_url": "https://arxiv.org/abs/2601.04035", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-04035/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c40e3db3f538f089c19d7c8233ea01174d0591c3211a7acc412ee4df28e902a8", "submitted_at": "2026-01-07T15:51:44Z", "updated_at": "2026-08-31T13:27:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.04035v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.04035"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04035"}], "name": "MobileDreamer: Generative Sketch World Model for GUI Agent", "publication": null, "published_at": "2026-01-07", "record_sha256": "3eedb2e3a316afe451f58805b4cdd4b6312ff559a721c145a622bcb76bb0e169", "relations": [], "resource_id": "arxiv:2601.04035", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.04035v2", "topics": ["world-models", "planning", "mobile-navigation"], "updated_at": "2026-08-31"}, "text": "MobileDreamer: Generative Sketch World Model for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nMobileDreamer predicts post-action GUI states as task-relevant textual sketches with an order-invariant learning objective. Agents use these predictions for lookahead action selection, evaluated on AndroidWorld.\nPlatforms: android\nTopics: world-models, planning, mobile-navigation\nPublished: 2026-01-07\nAuthors: Yilin Cao, Yufeng Zhong, Zhixiong Zeng, Siran Dai, Liming Zheng, Jing Huang, Haibo Qiu, Peng Shi, Wenji Mao\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.04035v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-04035/"}
{"id": "arxiv:2601.04126", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4d0a59f1eb8197b4d389f5f40957fb8aeb1e22d0739f9877b26d57ed70fe59b", "submitted_at": "2026-01-07T17:40:08Z", "updated_at": "2026-06-30T13:54:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.04126v3"}, "authors": ["Ziyun Zhang", "Zezhou Wang", "Xiaoyi Zhang", "Zongyu Guo", "Jiahao Li", "Bin Li", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2601.04126", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-04126/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4d0a59f1eb8197b4d389f5f40957fb8aeb1e22d0739f9877b26d57ed70fe59b", "submitted_at": "2026-01-07T17:40:08Z", "updated_at": "2026-06-30T13:54:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.04126v3"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.04126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04126"}], "name": "InfiniteWeb: Scalable Web Environment Synthesis for GUI Agent Training", "publication": null, "published_at": "2026-01-07", "record_sha256": "0a729d31ac1bb13a8ec0944e844454c5a85b533bef25d8bda3e5f10185ca0c53", "relations": [], "resource_id": "arxiv:2601.04126", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.04126v3", "topics": ["data-generation", "reinforcement-learning", "evaluation", "sandbox"], "updated_at": "2026-06-30"}, "text": "InfiniteWeb: Scalable Web Environment Synthesis for GUI Agent Training\nResource type: paper\nReview status: curated\nScope: core\n\nInfiniteWeb generates functional, interconnected websites and task evaluators using shared specifications and task-driven development. It evaluates website construction and the transfer of GUI policies trained in these synthetic web environments to OSWorld and Online-Mind2Web.\nPlatforms: web, desktop\nTopics: data-generation, reinforcement-learning, evaluation, sandbox\nPublished: 2026-01-07\nAuthors: Ziyun Zhang, Zezhou Wang, Xiaoyi Zhang, Zongyu Guo, Jiahao Li, Bin Li, Yan Lu\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.04126v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-04126/"}
{"id": "arxiv:2601.04500", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47a340b19cf399f2ae7851482030d85e572b0f6ad8415c67657beec737a80e06", "submitted_at": "2026-01-08T02:07:53Z", "updated_at": "2026-01-08T02:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.04500v1"}, "authors": ["Yifei Gao", "Jiang Wu", "Xiaoyi Chen", "Yifan Yang", "Zhe Cui", "Tianyi Ma", "Jiaming Zhang", "Jitao Sang"], "canonical_source_url": "https://arxiv.org/abs/2601.04500", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-04500/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47a340b19cf399f2ae7851482030d85e572b0f6ad8415c67657beec737a80e06", "submitted_at": "2026-01-08T02:07:53Z", "updated_at": "2026-01-08T02:07:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.04500v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.04500"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.04500"}], "name": "GUITester: Enabling GUI Agents for Exploratory Defect Discovery", "publication": null, "published_at": "2026-01-08", "record_sha256": "ecf0656a3e9b974ff1f60312920039eadba757a9bdd93d2f781acf96febfeebe", "relations": [], "resource_id": "arxiv:2601.04500", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.04500v1", "topics": ["automation", "evaluation", "self-reflection"], "updated_at": "2026-01-08"}, "text": "GUITester: Enabling GUI Agents for Exploratory Defect Discovery\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUITestBench for interactive defect discovery and GUITester, which separates task navigation from anomaly verification. Testing-oriented plans and history-based reflection help distinguish software defects from agent execution mistakes.\nPlatforms: unspecified\nTopics: automation, evaluation, self-reflection\nPublished: 2026-01-08\nAuthors: Yifei Gao, Jiang Wu, Xiaoyi Chen, Yifan Yang, Zhe Cui, Tianyi Ma, Jiaming Zhang, Jitao Sang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.04500v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-04500/"}
{"id": "arxiv:2601.05787", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e342723078b50e612df749df80eadfd16e0fe29f5875777c1b3189d8b5981512", "submitted_at": "2026-01-09T13:26:38Z", "updated_at": "2026-02-10T15:12:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.05787v2"}, "authors": ["Zezhou Wang", "Ziyun Zhang", "Xiaoyi Zhang", "Zhuzhong Qian", "Yan Lu"], "canonical_source_url": "https://arxiv.org/abs/2601.05787", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-05787/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e342723078b50e612df749df80eadfd16e0fe29f5875777c1b3189d8b5981512", "submitted_at": "2026-01-09T13:26:38Z", "updated_at": "2026-02-10T15:12:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.05787v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:47Z", "role": "primary", "sha256": "ae98e77b8bb125e45277aacec768a047e7bb0e13f33ca9ff6ca7c19e29e56cca", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.05787"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:47Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.05787"}], "name": "From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation", "publication": null, "published_at": "2026-01-09", "record_sha256": "cb7fe1b2e346dd9748675a72aea4fc989dc4d4361761b875163171dbe6ef72f8", "relations": [], "resource_id": "arxiv:2601.05787", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.05787v2", "topics": ["reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2026-02-10"}, "text": "From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation\nResource type: paper\nReview status: curated\nScope: core\n\nBEPA adapts a small set of expert GUI traces into policy-aligned guidance through trajectories reachable by the learner and a dynamically updated task cache. It evaluates this two-level approach to reinforcement learning on desktop and web benchmarks.\nPlatforms: desktop, web\nTopics: reinforcement-learning, data-generation, evaluation\nPublished: 2026-01-09\nAuthors: Zezhou Wang, Ziyun Zhang, Xiaoyi Zhang, Zhuzhong Qian, Yan Lu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.05787v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-05787/"}
{"id": "arxiv:2601.07181", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c6a513f53c16ea6d989a7b2be56ad26b83e531bcea46b00f2e08476b686c97", "submitted_at": "2026-01-12T04:04:20Z", "updated_at": "2026-01-12T04:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07181v1"}, "authors": ["Yichun Zhang", "Xiangwu Guo", "Yauhong Goh", "Jessica Hu", "Zhiheng Chen", "Xin Wang", "Difei Gao", "Mike Zheng Shou"], "canonical_source_url": "https://arxiv.org/abs/2601.07181", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-07181/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c6a513f53c16ea6d989a7b2be56ad26b83e531bcea46b00f2e08476b686c97", "submitted_at": "2026-01-12T04:04:20Z", "updated_at": "2026-01-12T04:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07181v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.07181"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07181"}], "name": "ShowUI-Aloha: Human-Taught GUI Agent", "publication": null, "published_at": "2026-01-12", "record_sha256": "555a26777c42b684e8c543f235653a7280f06e4c38feacf4aa52435892f3296c", "relations": [], "resource_id": "arxiv:2601.07181", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.07181v1", "topics": ["recording", "workflow", "data-generation"], "updated_at": "2026-01-12"}, "text": "ShowUI-Aloha: Human-Taught GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nTransforms human screen recordings and input events into annotated demonstrations used by planning and execution components for GUI tasks.\nPlatforms: desktop\nTopics: recording, workflow, data-generation\nPublished: 2026-01-12\nAuthors: Yichun Zhang, Xiangwu Guo, Yauhong Goh, Jessica Hu, Zhiheng Chen, Xin Wang, Difei Gao, Mike Zheng Shou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.07181v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-07181/"}
{"id": "arxiv:2601.07262", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2afe8d0d8d83f91003b7885635e098e19f69024306c43c3edb23c110e553f63", "submitted_at": "2026-01-12T07:08:42Z", "updated_at": "2026-04-23T09:34:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.07262v3"}, "authors": ["Jihong Wang", "Jiamu Zhou", "Weiming Zhang", "Teng Wang", "Weiwen Liu", "Zhuosheng Zhang", "Xingyu Lou", "Weinan Zhang", "Huarong Deng", "Jun Wang"], "canonical_source_url": "https://arxiv.org/abs/2601.07262", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-07262/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2afe8d0d8d83f91003b7885635e098e19f69024306c43c3edb23c110e553f63", "submitted_at": "2026-01-12T07:08:42Z", "updated_at": "2026-04-23T09:34:51Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.07262v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.07262"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07262"}], "name": "ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution", "publication": null, "published_at": "2026-01-12", "record_sha256": "2c1dba468372412d808cf6308037efd41cf4063f59da4a269ae5d67b40bee1f0", "relations": [], "resource_id": "arxiv:2601.07262", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.07262v3", "topics": ["memory", "human-in-the-loop", "long-horizon", "web-navigation"], "updated_at": "2026-04-23"}, "text": "ColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nColorBrowserAgent converts human feedback into reusable website knowledge and uses that knowledge to guide progressive context summaries. It evaluates the resulting memory and adaptation mechanisms on long browser workflows.\nPlatforms: web\nTopics: memory, human-in-the-loop, long-horizon, web-navigation\nPublished: 2026-01-12\nAuthors: Jihong Wang, Jiamu Zhou, Weiming Zhang, Teng Wang, Weiwen Liu, Zhuosheng Zhang, Xingyu Lou, Weinan Zhang, Huarong Deng, Jun Wang\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.07262v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-07262/"}
{"id": "arxiv:2601.07263", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4af00e0632ee21f70409998fc22f67e9a4ed1de415b488eda122c6ef94cab7aa", "submitted_at": "2026-01-12T07:10:08Z", "updated_at": "2026-01-12T07:10:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07263v1"}, "authors": ["Xinyi Wu", "Geng Hong", "Yueyue Chen", "MingXuan Liu", "Feier Jin", "Xudong Pan", "Jiarun Dai", "Baojun Liu"], "canonical_source_url": "https://arxiv.org/abs/2601.07263", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-07263/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4af00e0632ee21f70409998fc22f67e9a4ed1de415b488eda122c6ef94cab7aa", "submitted_at": "2026-01-12T07:10:08Z", "updated_at": "2026-01-12T07:10:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07263v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2601.07263"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07263"}], "name": "When Bots Take the Bait: Exposing and Mitigating the Emerging Social Engineering Attack in Web Automation Agent", "publication": null, "published_at": "2026-01-12", "record_sha256": "1e5b22125c921046718a7ef4aa338732676a772fa083994929d794091fb028eb", "relations": [], "resource_id": "arxiv:2601.07263", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.07263v1", "topics": ["safety-security", "web-navigation", "robustness"], "updated_at": "2026-01-12"}, "text": "When Bots Take the Bait: Exposing and Mitigating the Emerging Social Engineering Attack in Web Automation Agent\nResource type: paper\nReview status: curated\nScope: core\n\nStudies deceptive social-engineering web patterns targeting browser automation agents and proposes defenses against bait-based manipulation.\nPlatforms: web\nTopics: safety-security, web-navigation, robustness\nPublished: 2026-01-12\nAuthors: Xinyi Wu, Geng Hong, Yueyue Chen, MingXuan Liu, Feier Jin, Xudong Pan, Jiarun Dai, Baojun Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.07263v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-07263/"}
{"id": "arxiv:2601.07779", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfdaabaf3c6df80eb718e51dc666c54fb2ecf38ec520f6a71ccf5e5ce54248c8", "submitted_at": "2026-01-12T17:55:51Z", "updated_at": "2026-01-12T17:55:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07779v1"}, "authors": ["Bowen Yang", "Kaiming Jin", "Zhenyu Wu", "Zhaoyang Liu", "Qiushi Sun", "Zehao Li", "JingJing Xie", "Zhoumianze Liu", "Fangzhi Xu", "Kanzhi Cheng", "Qingyun Li", "Yian Wang", "Yu Qiao", "Zun Wang", "Zichen Ding"], "canonical_source_url": "https://arxiv.org/abs/2601.07779", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-07779/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfdaabaf3c6df80eb718e51dc666c54fb2ecf38ec520f6a71ccf5e5ce54248c8", "submitted_at": "2026-01-12T17:55:51Z", "updated_at": "2026-01-12T17:55:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.07779v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.07779"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.07779"}], "name": "OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent", "publication": null, "published_at": "2026-01-12", "record_sha256": "2cd5223cace25d15a66c2c823bfccf73e7743310d75c9c10ebc31bf63f555c80", "relations": [], "resource_id": "arxiv:2601.07779", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.07779v1", "topics": ["memory", "self-reflection", "planning", "tool-integration"], "updated_at": "2026-01-12"}, "text": "OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent\nResource type: paper\nReview status: curated\nScope: core\n\nOS-Symphony coordinates milestone-based reflection and memory with tools that build visually aligned tutorials through browser interaction. The framework evaluates trajectory correction and adaptation on online computer-use benchmarks including OSWorld.\nPlatforms: desktop, web\nTopics: memory, self-reflection, planning, tool-integration\nPublished: 2026-01-12\nAuthors: Bowen Yang, Kaiming Jin, Zhenyu Wu, Zhaoyang Liu, Qiushi Sun, Zehao Li, JingJing Xie, Zhoumianze Liu, Fangzhi Xu, Kanzhi Cheng, Qingyun Li, Yian Wang, Yu Qiao, Zun Wang, Zichen Ding\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.07779v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-07779/"}
{"id": "arxiv:2601.08406", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "626702d22786336ff65b6e89d97e2034ff7aa144fab4ecc286d169fca6993410", "submitted_at": "2026-01-13T10:21:28Z", "updated_at": "2026-01-13T10:21:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08406v1"}, "authors": ["Xinyi Wu", "Jiagui Chen", "Geng Hong", "Jiayi Dong", "Xudong Pan", "Jiarun Dai", "Min Yang"], "canonical_source_url": "https://arxiv.org/abs/2601.08406", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-08406/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "626702d22786336ff65b6e89d97e2034ff7aa144fab4ecc286d169fca6993410", "submitted_at": "2026-01-13T10:21:28Z", "updated_at": "2026-01-13T10:21:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.08406v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.08406"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.08406"}], "name": "WebTrap Park: An Automated Platform for Systematic Security Evaluation of Web Agents", "publication": null, "published_at": "2026-01-13", "record_sha256": "ba31ba2b2a333532896f5c474abafadc37a8dcd64cc956ac737b80f2d063c451", "relations": [], "resource_id": "arxiv:2601.08406", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.08406v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-01-13"}, "text": "WebTrap Park: An Automated Platform for Systematic Security Evaluation of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebTrap Park, a platform that turns web-agent security risks into executable tasks and assesses concrete interactions with live pages. It compares agent frameworks without requiring modifications to their implementations.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2026-01-13\nAuthors: Xinyi Wu, Jiagui Chen, Geng Hong, Jiayi Dong, Xudong Pan, Jiarun Dai, Min Yang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.08406v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-08406/"}
{"id": "arxiv:2601.08605", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73df8c5b417c3387530b7730be3aed8b69374c06fdd3573f56147b19359c6912", "submitted_at": "2026-01-13T14:48:34Z", "updated_at": "2026-04-15T14:59:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.08605v2"}, "authors": ["Wenyuan Zhang", "Xinghua Zhang", "Haiyang Yu", "Shuaiyi Nie", "Bingli Wu", "Juwei Yue", "Tingwen Liu", "Yongbin Li"], "canonical_source_url": "https://arxiv.org/abs/2601.08605", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-08605/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73df8c5b417c3387530b7730be3aed8b69374c06fdd3573f56147b19359c6912", "submitted_at": "2026-01-13T14:48:34Z", "updated_at": "2026-04-15T14:59:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.08605v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.08605"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.08605"}], "name": "ExpSeek: Self-Triggered Experience Seeking for Web Agents", "publication": null, "published_at": "2026-01-13", "record_sha256": "82543779714aa39902a8ef5408ec65a49d3d2ddc70b8903371c72a4ec9a70fdc", "relations": [], "resource_id": "arxiv:2601.08605", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.08605v2", "topics": ["memory", "planning", "web-navigation", "evaluation"], "updated_at": "2026-04-15"}, "text": "ExpSeek: Self-Triggered Experience Seeking for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nExpSeek uses step-level entropy signals to decide when a web agent should request experience-based guidance. It supplies guidance tailored to the current step and evaluates the intervention across web-agent benchmarks.\nPlatforms: web\nTopics: memory, planning, web-navigation, evaluation\nPublished: 2026-01-13\nAuthors: Wenyuan Zhang, Xinghua Zhang, Haiyang Yu, Shuaiyi Nie, Bingli Wu, Juwei Yue, Tingwen Liu, Yongbin Li\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.08605v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-08605/"}
{"id": "arxiv:2601.09636", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8b697b91942291403491581a61c135c6bfc8b0f3c2a9b9bab0f926af7208fce", "submitted_at": "2026-01-14T17:12:48Z", "updated_at": "2026-05-13T02:54:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.09636v2"}, "authors": ["Yibo Lyu", "Gongwei Chen", "Rui Shao", "Weili Guan", "Liqiang Nie"], "canonical_source_url": "https://arxiv.org/abs/2601.09636", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-09636/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d8b697b91942291403491581a61c135c6bfc8b0f3c2a9b9bab0f926af7208fce", "submitted_at": "2026-01-14T17:12:48Z", "updated_at": "2026-05-13T02:54:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.09636v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.09636"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09636"}], "name": "PersonalAlign: Hierarchical Implicit Intent Alignment for Personalized GUI Agent with Long-Term User-Centric Records", "publication": null, "published_at": "2026-01-14", "record_sha256": "fc0d1baa4845a50acb36fe398d1f17211ca9294860762f3be1d059465eeb72c5", "relations": [], "resource_id": "arxiv:2601.09636", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.09636v2", "topics": ["memory", "evaluation", "mobile-navigation"], "updated_at": "2026-05-13"}, "text": "PersonalAlign: Hierarchical Implicit Intent Alignment for Personalized GUI Agent with Long-Term User-Centric Records\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AndroidIntent to evaluate how GUI agents infer omitted preferences and recurring routines from long-term user records. Its HIM-Agent maintains hierarchical memory of preferences and routines for personalized execution and proactive suggestions.\nPlatforms: android\nTopics: memory, evaluation, mobile-navigation\nPublished: 2026-01-14\nAuthors: Yibo Lyu, Gongwei Chen, Rui Shao, Weili Guan, Liqiang Nie\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.09636v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-09636/"}
{"id": "arxiv:2601.09770", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d5ed7e72534fc55d028d4f25cb98b2373c88b622cf76396933d7a49d8ecf2e8", "submitted_at": "2026-01-14T14:27:28Z", "updated_at": "2026-01-14T14:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09770v1"}, "authors": ["Chen Chen", "Jiawei Shao", "Dakuan Lu", "Haoyi Hu", "Xiangcheng Liu", "Hantao Yao", "Wu Liu"], "canonical_source_url": "https://arxiv.org/abs/2601.09770", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-09770/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d5ed7e72534fc55d028d4f25cb98b2373c88b622cf76396933d7a49d8ecf2e8", "submitted_at": "2026-01-14T14:27:28Z", "updated_at": "2026-01-14T14:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.09770v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.09770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09770"}], "name": "GUI-Eyes: Tool-Augmented Perception for Visual Grounding in GUI Agents", "publication": null, "published_at": "2026-01-14", "record_sha256": "dd0691aaee49ed52e4a4bd6fa5ee66e4d44ae2b44db3468b63d3d5d139c5e538", "relations": [], "resource_id": "arxiv:2601.09770", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.09770v1", "topics": ["grounding", "reinforcement-learning", "tool-integration", "screen-understanding"], "updated_at": "2026-01-14"}, "text": "GUI-Eyes: Tool-Augmented Perception for Visual Grounding in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Eyes trains GUI grounders to decide when and how to use visual tools such as crops and zooms. Its coarse-to-fine policy uses dense spatial rewards based on location proximity and region overlap, with evaluation on ScreenSpot-Pro.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, tool-integration, screen-understanding\nPublished: 2026-01-14\nAuthors: Chen Chen, Jiawei Shao, Dakuan Lu, Haoyi Hu, Xiangcheng Liu, Hantao Yao, Wu Liu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.09770v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-09770/"}
{"id": "arxiv:2601.09923", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e32917bc92b65701eea06c09a2c681c7504418f8ce888af2b72012a1da50c9f6", "submitted_at": "2026-01-14T23:06:35Z", "updated_at": "2026-06-04T15:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.09923v3"}, "authors": ["Hanna Foerster", "Tom Blanchard", "Kristina Nikolić", "Ilia Shumailov", "Cheng Zhang", "Robert Mullins", "Nicolas Papernot", "Florian Tramèr", "Yiren Zhao"], "canonical_source_url": "https://arxiv.org/abs/2601.09923", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-09923/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e32917bc92b65701eea06c09a2c681c7504418f8ce888af2b72012a1da50c9f6", "submitted_at": "2026-01-14T23:06:35Z", "updated_at": "2026-06-04T15:38:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.09923v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.09923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.09923"}], "name": "CaMeLs Can Use Computers Too: System-level Security for Computer Use Agents", "publication": null, "published_at": "2026-01-14", "record_sha256": "30bc5e8ac19ca93e64f008dfdcd7d0413f6d0235348e1934e0bcae85b21ddff7", "relations": [], "resource_id": "arxiv:2601.09923", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.09923v3", "topics": ["safety-security", "planning", "robustness"], "updated_at": "2026-06-04"}, "text": "CaMeLs Can Use Computers Too: System-level Security for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nNOVA separates trusted task planning from screen-derived observations by producing a branching plan before execution and resolving runtime values through perception. Its OSWorld study examines security and utility while identifying remaining risks from perception-driven branch steering.\nPlatforms: desktop\nTopics: safety-security, planning, robustness\nPublished: 2026-01-14\nAuthors: Hanna Foerster, Tom Blanchard, Kristina Nikolić, Ilia Shumailov, Cheng Zhang, Robert Mullins, Nicolas Papernot, Florian Tramèr, Yiren Zhao\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.09923v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-09923/"}
{"id": "arxiv:2601.11631", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "600f2a44695886fc527b04e1a8b1ad07cec90cf5c58dee834605d5133a760b1e", "submitted_at": "2026-01-14T03:52:57Z", "updated_at": "2026-01-14T03:52:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11631v1"}, "authors": ["Yurun Song", "Jiong Yin", "Rongjunchen Zhang", "Ian G. Harris"], "canonical_source_url": "https://arxiv.org/abs/2601.11631", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-11631/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "600f2a44695886fc527b04e1a8b1ad07cec90cf5c58dee834605d5133a760b1e", "submitted_at": "2026-01-14T03:52:57Z", "updated_at": "2026-01-14T03:52:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.11631v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.11631"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.11631"}], "name": "Compress to Focus: Efficient Coordinate Compression for Policy Optimization in Multi-Turn GUI Agents", "publication": null, "published_at": "2026-01-14", "record_sha256": "b892848be6e2b33b630474ee9d650db9ed984459432cae51b1f236e02b5b3160", "relations": [], "resource_id": "arxiv:2601.11631", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.11631v1", "topics": ["grounding", "reinforcement-learning", "efficiency", "interaction-history"], "updated_at": "2026-01-14"}, "text": "Compress to Focus: Efficient Coordinate Compression for Policy Optimization in Multi-Turn GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCCPO couples GUI policy optimization with coordinate-guided compression of past visual observations. It aggregates targets across rollouts to focus retained context and uses distance-based advantages to train grounding and compression together.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, efficiency, interaction-history\nPublished: 2026-01-14\nAuthors: Yurun Song, Jiong Yin, Rongjunchen Zhang, Ian G. Harris\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.11631v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-11631/"}
{"id": "arxiv:2601.12349", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21848dea9e000125550fff7fe3844a57bf100215bf9eb800cbdbb612675f7074", "submitted_at": "2026-01-18T10:54:54Z", "updated_at": "2026-07-15T15:32:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.12349v3"}, "authors": ["Yi Qian", "Kunwei Qian", "Xingbang He", "Ligeng Chen", "Jikang Zhang", "Tiantai Zhang", "Haiyang Wei", "Linzhang Wang", "Hao Wu", "Bing Mao"], "canonical_source_url": "https://arxiv.org/abs/2601.12349", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-12349/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "21848dea9e000125550fff7fe3844a57bf100215bf9eb800cbdbb612675f7074", "submitted_at": "2026-01-18T10:54:54Z", "updated_at": "2026-07-15T15:32:11Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.12349v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.12349"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.12349"}], "name": "Mind the Gap: Action Rebinding Attacks against Android GUI Agents", "publication": null, "published_at": "2026-01-18", "record_sha256": "fbb8235e5f96432e9737aaf617ffd1c5e3cda81246f6f9586d786a044bf19d7d", "relations": [], "resource_id": "arxiv:2601.12349", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.12349v3", "topics": ["safety-security", "cross-app", "evaluation"], "updated_at": "2026-07-15"}, "text": "Mind the Gap: Action Rebinding Attacks against Android GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies action-rebinding attacks in which an Android GUI agent’s planned input is redirected to a different application during the observation–action delay. The evaluation examines cross-application execution and the consequences of treating an earlier screen as the current action target.\nPlatforms: android\nTopics: safety-security, cross-app, evaluation\nPublished: 2026-01-18\nAuthors: Yi Qian, Kunwei Qian, Xingbang He, Ligeng Chen, Jikang Zhang, Tiantai Zhang, Haiyang Wei, Linzhang Wang, Hao Wu, Bing Mao\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.12349v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-12349/"}
{"id": "arxiv:2601.12822", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57586682d4675387dea2687e7c15d70cc10aecac883e35fe1812c39abcbc1dc", "submitted_at": "2026-01-19T08:32:09Z", "updated_at": "2026-01-19T08:32:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12822v1"}, "authors": ["Wenqi Zhang", "Yulin Shen", "Changyue Jiang", "Jiarun Dai", "Geng Hong", "Xudong Pan"], "canonical_source_url": "https://arxiv.org/abs/2601.12822", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-12822/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57586682d4675387dea2687e7c15d70cc10aecac883e35fe1812c39abcbc1dc", "submitted_at": "2026-01-19T08:32:09Z", "updated_at": "2026-01-19T08:32:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.12822v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.12822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.12822"}], "name": "MirrorGuard: Toward Secure Computer-Use Agents via Simulation-to-Real Reasoning Correction", "publication": null, "published_at": "2026-01-19", "record_sha256": "f94f2d674e321f047dbae525039c721788131b38f2d9cae3915026d2b4c4ce6f", "relations": [], "resource_id": "arxiv:2601.12822", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.12822v1", "topics": ["safety-security", "data-generation", "robustness"], "updated_at": "2026-01-19"}, "text": "MirrorGuard: Toward Secure Computer-Use Agents via Simulation-to-Real Reasoning Correction\nResource type: paper\nReview status: curated\nScope: core\n\nMirrorGuard trains a reasoning-correction defense using GUI risk trajectories generated in a text-based neural-symbolic simulator. It evaluates whether corrections transfer to computer-use agents operating real interfaces, keeping simulated training separate from deployment tests.\nPlatforms: unspecified\nTopics: safety-security, data-generation, robustness\nPublished: 2026-01-19\nAuthors: Wenqi Zhang, Yulin Shen, Changyue Jiang, Jiarun Dai, Geng Hong, Xudong Pan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.12822v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-12822/"}
{"id": "arxiv:2601.13060", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa8cb4137e862bf8031a55bc8f81e1d6c903733f2cfac9cf758e5f2f92884481", "submitted_at": "2026-01-19T13:50:43Z", "updated_at": "2026-01-19T13:50:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13060v1"}, "authors": ["Zecheng Li", "Zhihui Cao", "Wenke Huang", "Yudong Zhang", "Keying Qi", "Rui Wang", "Zeyu Zheng", "Jian Zhao", "Hao Zhu", "Hengxin Wu", "Yuran Wang", "Guitao Fan", "Guokun Wu", "Yicong Liu", "Zhilin Gao", "Haikun Xu", "He Yang", "Minqi Xiang", "Xingyu Liu", "Zuojian Wang"], "canonical_source_url": "https://arxiv.org/abs/2601.13060", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-13060/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa8cb4137e862bf8031a55bc8f81e1d6c903733f2cfac9cf758e5f2f92884481", "submitted_at": "2026-01-19T13:50:43Z", "updated_at": "2026-01-19T13:50:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.13060v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.13060"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.13060"}], "name": "MagicGUI-RMS: A Multi-Agent Reward Model System for Self-Evolving GUI Agents via Automated Feedback Reflux", "publication": null, "published_at": "2026-01-19", "record_sha256": "d78f5f671a8fbd7a78d6da68a32fc9a65bb71a9ebcf6c35f29958b94e32b0ce2", "relations": [], "resource_id": "arxiv:2601.13060", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.13060v1", "topics": ["reward-models", "data-generation", "evaluation", "robustness"], "updated_at": "2026-01-19"}, "text": "MagicGUI-RMS: A Multi-Agent Reward Model System for Self-Evolving GUI Agents via Automated Feedback Reflux\nResource type: paper\nReview status: curated\nScope: core\n\nMagicGUI-RMS combines domain-specific and general reward models to assess GUI actions and propose corrections. A structured data pipeline and feedback-recycling process support repeated reward-model and agent improvement.\nPlatforms: unspecified\nTopics: reward-models, data-generation, evaluation, robustness\nPublished: 2026-01-19\nAuthors: Zecheng Li, Zhihui Cao, Wenke Huang, Yudong Zhang, Keying Qi, Rui Wang, Zeyu Zheng, Jian Zhao, Hao Zhu, Hengxin Wu, Yuran Wang, Guitao Fan, Guokun Wu, Yicong Liu, Zhilin Gao, Haikun Xu, He Yang, Minqi Xiang, Xingyu Liu, Zuojian Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.13060v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-13060/"}
{"id": "arxiv:2601.15876", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38b2798ce9ebd69a46202f0fa8ceca859394acda4f5173b5e30248914c8c35ac", "submitted_at": "2026-01-22T11:36:43Z", "updated_at": "2026-01-23T08:52:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.15876v2"}, "authors": ["Taofeng Xue", "Chong Peng", "Mianqiu Huang", "Linsen Guo", "Tiancheng Han", "Haozhe Wang", "Jianing Wang", "Xiaocheng Zhang", "Xin Yang", "Dengchang Zhao", "Jinrui Ding", "Xiandi Ma", "Yuchen Xie", "Peng Pei", "Xunliang Cai", "Xipeng Qiu"], "canonical_source_url": "https://arxiv.org/abs/2601.15876", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-15876/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "38b2798ce9ebd69a46202f0fa8ceca859394acda4f5173b5e30248914c8c35ac", "submitted_at": "2026-01-22T11:36:43Z", "updated_at": "2026-01-23T08:52:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.15876v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.15876"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.15876"}], "name": "EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience", "publication": null, "published_at": "2026-01-22", "record_sha256": "ab45de501057bb5886fc3f837b3f6bf74b0acdd7453b52b2969dc43ce96864dd", "relations": [], "resource_id": "arxiv:2601.15876", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.15876v2", "topics": ["data-generation", "sandbox", "long-horizon", "self-reflection"], "updated_at": "2026-01-23"}, "text": "EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience\nResource type: paper\nReview status: curated\nScope: core\n\nEvoCUA combines task synthesis with executable validators, asynchronous sandbox rollouts and iterative policy updates. It turns successful routines and analyzed failures into supervision and evaluates the evolving agent on OSWorld.\nPlatforms: desktop\nTopics: data-generation, sandbox, long-horizon, self-reflection\nPublished: 2026-01-22\nAuthors: Taofeng Xue, Chong Peng, Mianqiu Huang, Linsen Guo, Tiancheng Han, Haozhe Wang, Jianing Wang, Xiaocheng Zhang, Xin Yang, Dengchang Zhao, Jinrui Ding, Xiandi Ma, Yuchen Xie, Peng Pei, Xunliang Cai, Xipeng Qiu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.15876v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-15876/"}
{"id": "arxiv:2601.16356", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bdc29648446cd79e0c0b9f72814457b55258d501283ce2fa2dedcd8262f885b", "submitted_at": "2026-01-22T22:42:40Z", "updated_at": "2026-01-22T22:42:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.16356v1"}, "authors": ["Simret Araya Gebreegziabher", "Yukun Yang", "Charles Chiang", "Hojun Yoo", "Chaoran Chen", "Hyo Jin Do", "Zahra Ashktorab", "Werner Geyer", "Diego Gómez-Zará", "Toby Jia-Jun Li"], "canonical_source_url": "https://arxiv.org/abs/2601.16356", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-16356/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bdc29648446cd79e0c0b9f72814457b55258d501283ce2fa2dedcd8262f885b", "submitted_at": "2026-01-22T22:42:40Z", "updated_at": "2026-01-22T22:42:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.16356v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.16356"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.16356"}], "name": "The Behavioral Fabric of LLM-Powered GUI Agents: Human Values and Interaction Outcomes", "publication": null, "published_at": "2026-01-22", "record_sha256": "eef0a1b6750392692b731fab53e8f3bb8797cdd89715dd5bbf78120ebd71269b", "relations": [], "resource_id": "arxiv:2601.16356", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.16356v1", "topics": ["evaluation", "ui-design", "human-in-the-loop"], "updated_at": "2026-01-22"}, "text": "The Behavioral Fabric of LLM-Powered GUI Agents: Human Values and Interaction Outcomes\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a controlled web-task testbed for studying how stated preferences and human values influence GUI-agent decisions. It analyzes how personas alter navigation and how prominent interface cues can override the intended preference guidance.\nPlatforms: web\nTopics: evaluation, ui-design, human-in-the-loop\nPublished: 2026-01-22\nAuthors: Simret Araya Gebreegziabher, Yukun Yang, Charles Chiang, Hojun Yoo, Chaoran Chen, Hyo Jin Do, Zahra Ashktorab, Werner Geyer, Diego Gómez-Zará, Toby Jia-Jun Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.16356v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-16356/"}
{"id": "arxiv:2601.17418", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce0909d1c9081f099ae0dc0b0182e0442fd71fd58c62b13d486d154afc2187e", "submitted_at": "2026-01-24T11:35:37Z", "updated_at": "2026-01-24T11:35:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17418v1"}, "authors": ["Mingxian Yu", "Siqi Luo", "Xu Chen"], "canonical_source_url": "https://arxiv.org/abs/2601.17418", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-17418/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ce0909d1c9081f099ae0dc0b0182e0442fd71fd58c62b13d486d154afc2187e", "submitted_at": "2026-01-24T11:35:37Z", "updated_at": "2026-01-24T11:35:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17418v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2601.17418"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.17418"}], "name": "GraphPilot: GUI Task Automation with One-Step LLM Reasoning Powered by Knowledge Graph", "publication": null, "published_at": "2026-01-24", "record_sha256": "14bf61b2028b2ea32ad9d3dbe3a036221394a1bf57ddc8ef23437192c0f346a1", "relations": [], "resource_id": "arxiv:2601.17418", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.17418v1", "topics": ["mobile-navigation", "planning", "memory", "efficiency"], "updated_at": "2026-01-24"}, "text": "GraphPilot: GUI Task Automation with One-Step LLM Reasoning Powered by Knowledge Graph\nResource type: paper\nReview status: curated\nScope: core\n\nGraphPilot builds app-specific knowledge graphs from explored pages, elements and transitions, then uses them to plan action sequences. It requests additional interface observations when uncertain and checks plans against graph transition rules before correction and execution.\nPlatforms: unspecified\nTopics: mobile-navigation, planning, memory, efficiency\nPublished: 2026-01-24\nAuthors: Mingxian Yu, Siqi Luo, Xu Chen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.17418v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-17418/"}
{"id": "arxiv:2601.17722", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa7c8ce1b0e584c5279691ff06cd522d7c8ea1f4801d7074324c6845ce482efb", "submitted_at": "2026-01-25T06:58:15Z", "updated_at": "2026-01-25T06:58:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17722v1"}, "authors": ["Ying Mo", "Yu Bai", "Dapeng Sun", "Yuqian Shi", "Yukai Miao", "Li Chen", "Dan Li"], "canonical_source_url": "https://arxiv.org/abs/2601.17722", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-17722/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa7c8ce1b0e584c5279691ff06cd522d7c8ea1f4801d7074324c6845ce482efb", "submitted_at": "2026-01-25T06:58:15Z", "updated_at": "2026-01-25T06:58:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.17722v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.17722"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.17722"}], "name": "EntWorld: A Holistic Environment and Benchmark for Verifiable Enterprise GUI Agents", "publication": null, "published_at": "2026-01-25", "record_sha256": "487c3929075209022b5e77f5a83bf32b2805134b5fd298e26fc039223b92468d", "relations": [], "resource_id": "arxiv:2601.17722", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.17722v1", "topics": ["evaluation", "industry", "workflow"], "updated_at": "2026-01-25"}, "text": "EntWorld: A Holistic Environment and Benchmark for Verifiable Enterprise GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces EntWorld, an enterprise GUI benchmark built by deriving task logic from database schemas. Long workflows across enterprise domains are assessed with deterministic SQL-based state checks rather than screenshot similarity alone.\nPlatforms: unspecified\nTopics: evaluation, industry, workflow\nPublished: 2026-01-25\nAuthors: Ying Mo, Yu Bai, Dapeng Sun, Yuqian Shi, Yukai Miao, Li Chen, Dan Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.17722v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-17722/"}
{"id": "arxiv:2601.18092", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae1716f673e47cd0e61b5f2225ca4871dfd8522e35037fc8d226f9c7cd0e3e75", "submitted_at": "2026-01-26T03:02:26Z", "updated_at": "2026-01-26T03:02:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.18092v1"}, "authors": ["Nan Chen", "Jing Lu", "Zilong Wang", "Luna K. Qiu", "Siming Chen", "Yuqing Yang"], "canonical_source_url": "https://arxiv.org/abs/2601.18092", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-18092/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ae1716f673e47cd0e61b5f2225ca4871dfd8522e35037fc8d226f9c7cd0e3e75", "submitted_at": "2026-01-26T03:02:26Z", "updated_at": "2026-01-26T03:02:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.18092v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.18092"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18092"}], "name": "From Struggle to Success: Context-Aware Guidance for Screen Reader Users in Computer Use", "publication": null, "published_at": "2026-01-26", "record_sha256": "7c1bb8832689d0a6910c71259cc94bd5d17c4a6a6d63fe5a877f5b61b05ab98a", "relations": [], "resource_id": "arxiv:2601.18092", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2601.18092v1", "topics": ["human-in-the-loop", "evaluation", "workflow"], "updated_at": "2026-01-26"}, "text": "From Struggle to Success: Context-Aware Guidance for Screen Reader Users in Computer Use\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces AskEase, an on-demand assistant that combines contextual information to provide screen-reader-friendly computer-use guidance. A study with screen-reader users evaluates task completion and perceived workload while preserving human control of the interaction.\nPlatforms: unspecified\nTopics: human-in-the-loop, evaluation, workflow\nPublished: 2026-01-26\nAuthors: Nan Chen, Jing Lu, Zilong Wang, Luna K. Qiu, Siming Chen, Yuqing Yang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.18092v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-18092/"}
{"id": "arxiv:2601.18197", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccb46b86b0c0b854ed2f1e17f9e12bf031c5ed1349f3d3c44d3276b2a3a195ee", "submitted_at": "2026-01-26T06:29:41Z", "updated_at": "2026-06-26T07:19:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18197v2"}, "authors": ["Shaokang Wang", "Pei Fu", "Ruoceng Zhang", "Shaojie Zhang", "Xiuwen Xi", "Jiahui Yang", "Bin Qin", "Ying Huang", "Zhenbo Luo", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2601.18197", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-18197/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccb46b86b0c0b854ed2f1e17f9e12bf031c5ed1349f3d3c44d3276b2a3a195ee", "submitted_at": "2026-01-26T06:29:41Z", "updated_at": "2026-06-26T07:19:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18197v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:55Z", "role": "primary", "sha256": "e6061c8dd9c00a70fcf1d4325b671c50a6912063ecbe2b6359175ae4622983a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2511.13091%2C2511.21398%2C2511.22235%2C2512.00756%2C2512.00846%2C2512.01174%2C2512.01763%2C2512.02423%2C2512.04105%2C2512.05941%2C2512.07725%2C2512.09396%2C2512.11611%2C2512.12692%2C2512.17221%2C2512.19396%2C2512.21302%2C2512.22009%2C2512.22047%2C2512.22322%2C2512.23128%2C2601.02371%2C2601.02439%2C2601.03641%2C2601.04035%2C2601.04126%2C2601.04500%2C2601.05787%2C2601.06899%2C2601.08605%2C2601.09636%2C2601.09770%2C2601.11631%2C2601.12349%2C2601.16356%2C2601.17418%2C2601.18197%2C2601.18305%2C2601.19199%2C2601.20732%2C2601.21352%2C2601.21872%2C2601.21961%2C2601.22149%2C2601.22491%2C2601.22528%2C2602.02468%2C2602.05354%2C2602.06090%2C2602.06351&max_results=50"}], "identifiers": {"arxiv": "2601.18197"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18197"}], "name": "GAIA: A Data Flywheel System for Training GUI Test-Time Scaling Critic Models", "publication": null, "published_at": "2026-01-26", "record_sha256": "c75dda3f5615a625575c9ed137762421e79b015dfe28150dd1990f4c1b7222c8", "relations": [], "resource_id": "arxiv:2601.18197", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.18197v2", "topics": ["reward-models", "data-generation", "evaluation"], "updated_at": "2026-06-26"}, "text": "GAIA: A Data Flywheel System for Training GUI Test-Time Scaling Critic Models\nResource type: paper\nReview status: curated\nScope: core\n\nGAIA trains a GUI action critic on positive and negative examples, then uses that critic to collect improved examples for another training round. It studies how recycling these data affects the test-time action choices of different GUI agents.\nPlatforms: unspecified\nTopics: reward-models, data-generation, evaluation\nPublished: 2026-01-26\nAuthors: Shaokang Wang, Pei Fu, Ruoceng Zhang, Shaojie Zhang, Xiuwen Xi, Jiahui Yang, Bin Qin, Ying Huang, Zhenbo Luo, Jian Luan\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.18197v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-18197/"}
{"id": "arxiv:2601.18305", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb90eed33e989d38cd4058d54ba7d073f1b8e5c1453a9106dfe5e533da7b08a1", "submitted_at": "2026-01-26T09:35:10Z", "updated_at": "2026-08-20T07:50:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18305v2"}, "authors": ["Xuan Wang", "Siyuan Su", "Quantong Fu", "Yongxiang Hu", "Yangfan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2601.18305", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-18305/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb90eed33e989d38cd4058d54ba7d073f1b8e5c1453a9106dfe5e533da7b08a1", "submitted_at": "2026-01-26T09:35:10Z", "updated_at": "2026-08-20T07:50:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.18305v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.18305"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18305"}], "name": "SwipeGen: Bridging the Execution Gap in GUI Agents via Human-like Swipe Synthesis", "publication": null, "published_at": "2026-01-26", "record_sha256": "f4ace9d6f73334f1faaf21c183f0b09089517550217498d859da7ee1f3a2268f", "relations": [], "resource_id": "arxiv:2601.18305", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.18305v2", "topics": ["action-prediction", "data-generation", "evaluation"], "updated_at": "2026-08-20"}, "text": "SwipeGen: Bridging the Execution Gap in GUI Agents via Human-like Swipe Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SwipeGen for synthesizing varied swipe trajectories and SwipeBench for assessing their execution quality. The study examines how richer control of trajectory, speed, and timing changes GUI agents’ interaction performance.\nPlatforms: unspecified\nTopics: action-prediction, data-generation, evaluation\nPublished: 2026-01-26\nAuthors: Xuan Wang, Siyuan Su, Quantong Fu, Yongxiang Hu, Yangfan Zhou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.18305v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-18305/"}
{"id": "arxiv:2601.18842", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca912ad67a681e8e908a5a403e982f39b5080bee4df376eb0b598468f3bb504", "submitted_at": "2026-01-26T11:33:40Z", "updated_at": "2026-05-13T07:11:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.18842v3"}, "authors": ["Yanxi Wang", "Zhiling Zhang", "Wenbo Zhou", "Weiming Zhang", "Jie Zhang", "Qiannan Zhu", "Yu Shi", "Shuxin Zheng", "Jiyan He"], "canonical_source_url": "https://arxiv.org/abs/2601.18842", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-18842/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1ca912ad67a681e8e908a5a403e982f39b5080bee4df376eb0b598468f3bb504", "submitted_at": "2026-01-26T11:33:40Z", "updated_at": "2026-05-13T07:11:36Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.18842v3"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.18842"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.18842"}], "name": "GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents", "publication": null, "published_at": "2026-01-26", "record_sha256": "febf16ff885b9dbd27961911a12a141c6d18939ebf52c41464c8d43fe1809b4a", "relations": [], "resource_id": "arxiv:2601.18842", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.18842v3", "topics": ["safety-security", "evaluation", "grounding"], "updated_at": "2026-05-13"}, "text": "GUIGuard-Bench: Toward a General Evaluation for Privacy-Preserving GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUIGuard-Bench, which annotates private regions in GUI trajectories with categories, risk levels, and relevance to the user’s task. It evaluates privacy recognition and the effect of screenshot protection on offline planning and task utility.\nPlatforms: android, desktop\nTopics: safety-security, evaluation, grounding\nPublished: 2026-01-26\nAuthors: Yanxi Wang, Zhiling Zhang, Wenbo Zhou, Weiming Zhang, Jie Zhang, Qiannan Zhu, Yu Shi, Shuxin Zheng, Jiyan He\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.18842v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-18842/"}
{"id": "arxiv:2601.19199", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1930b4efca7ee38fc6cd1ceb283c14ed9b1388e7b758b951ba9469ed3067c3e0", "submitted_at": "2026-01-27T04:58:48Z", "updated_at": "2026-01-31T15:17:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.19199v2"}, "authors": ["Libo Sun", "Jiwen Zhang", "Siyuan Wang", "Zhongyu Wei"], "canonical_source_url": "https://arxiv.org/abs/2601.19199", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-19199/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1930b4efca7ee38fc6cd1ceb283c14ed9b1388e7b758b951ba9469ed3067c3e0", "submitted_at": "2026-01-27T04:58:48Z", "updated_at": "2026-01-31T15:17:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.19199v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.19199"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.19199"}], "name": "MAGNET: Towards Adaptive GUI Agents with Memory-Driven Knowledge Evolution", "publication": null, "published_at": "2026-01-27", "record_sha256": "d56948cc3712c46a738bfe24b07f8811d8ba87aa1e1036aeacabdfcfb0565bb5", "relations": [], "resource_id": "arxiv:2601.19199", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.19199v2", "topics": ["memory", "grounding", "robustness", "evaluation"], "updated_at": "2026-01-31"}, "text": "MAGNET: Towards Adaptive GUI Agents with Memory-Driven Knowledge Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nMAGNET maintains separate memories for stable interface functions and task procedures across changing layouts and workflows. It updates frequently used knowledge and evaluates adaptation in AndroidWorld and offline distribution-shift settings.\nPlatforms: android\nTopics: memory, grounding, robustness, evaluation\nPublished: 2026-01-27\nAuthors: Libo Sun, Jiwen Zhang, Siyuan Wang, Zhongyu Wei\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.19199v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-19199/"}
{"id": "arxiv:2601.19306", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "223b9a915e5ea9849644318e1e5426e64ad8655ab46d0b007c30e1967130dc50", "submitted_at": "2026-01-27T07:46:05Z", "updated_at": "2026-01-27T07:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.19306v1"}, "authors": ["Sijia Li", "Xiaoyu Tan", "Shahir Ali", "Niels Schmidt", "Gengchen Ma", "Xihe Qiu"], "canonical_source_url": "https://arxiv.org/abs/2601.19306", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-19306/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "223b9a915e5ea9849644318e1e5426e64ad8655ab46d0b007c30e1967130dc50", "submitted_at": "2026-01-27T07:46:05Z", "updated_at": "2026-01-27T07:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.19306v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:01:04.629294+00:00", "role": "primary", "sha256": "53521b0a80976320b10efcf7840629e4a2085119710324dd48db2cfd9c5d9cf6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:01:04.629294+00:00", "role": "primary", "sha256": "53521b0a80976320b10efcf7840629e4a2085119710324dd48db2cfd9c5d9cf6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=500&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:20.444793+00:00", "role": "primary", "sha256": "ef3bbc8adcc147239ea910e0c59904d7a26ee1610a7a4d1613be2cb845098536", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2601.19306v1"}], "identifiers": {"arxiv": "2601.19306"}, "kind": "paper", "last_verified_at": "2026-10-18T10:01:04.629294+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.19306"}], "name": "Curiosity Driven Knowledge Retrieval for Mobile Agents", "publication": null, "published_at": "2026-01-27", "record_sha256": "b5989d398d72b3439ec51094826ccd43d02a90d980c94e874df04ef2e851d39e", "relations": [], "resource_id": "arxiv:2601.19306", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.19306v1", "topics": ["memory", "planning", "mobile-navigation"], "updated_at": "2026-01-27"}, "text": "Curiosity Driven Knowledge Retrieval for Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses execution uncertainty to trigger retrieval from application documentation, code, and past trajectories, organizing the results into structured AppCards for Android agent planning.\nPlatforms: android\nTopics: memory, planning, mobile-navigation\nPublished: 2026-01-27\nAuthors: Sijia Li, Xiaoyu Tan, Shahir Ali, Niels Schmidt, Gengchen Ma, Xihe Qiu\nRecorded arXiv version: v1\n\nThe curiosity signal is a model-based proxy, not a calibrated probability of failure. Accurate external app information is a dependency, and AndroidWorld is the reported evaluation environment. No performance numbers were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Problem Formulation; 3.2 Curiosity-Driven Information Gain Estimation; 3.3 Retrieving and Embedding External Knowledge); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2601.19306v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-19306/"}
{"id": "arxiv:2601.20335", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1cbf16dd3c790996758e00c085a9c2bccabf5b24f945bbd8440cc1f953e38568", "submitted_at": "2026-01-28T07:49:48Z", "updated_at": "2026-01-29T02:32:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.20335v2"}, "authors": ["Qinzhuo Wu", "Zhizhuo Yang", "Hanhao Li", "Pengzhi Gao", "Wei Liu", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2601.20335", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-20335/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1cbf16dd3c790996758e00c085a9c2bccabf5b24f945bbd8440cc1f953e38568", "submitted_at": "2026-01-28T07:49:48Z", "updated_at": "2026-01-29T02:32:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.20335v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.20335"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20335"}], "name": "MobileBench-OL: A Comprehensive Chinese Benchmark for Evaluating Mobile GUI Agents in Real-World Environment", "publication": null, "published_at": "2026-01-28", "record_sha256": "46520d6b02449b60f5e660f3c0e0147f7551a58f8f294d1c26e6e189f0d0e6d5", "relations": [], "resource_id": "arxiv:2601.20335", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.20335v2", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2026-01-29"}, "text": "MobileBench-OL: A Comprehensive Chinese Benchmark for Evaluating Mobile GUI Agents in Real-World Environment\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MobileBench-OL, an online benchmark of Chinese-app tasks spanning execution, reasoning, exploration, and noisy conditions. Its automatic evaluation and reset framework supports repeated testing of mobile GUI agents in real environments.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, robustness\nPublished: 2026-01-28\nAuthors: Qinzhuo Wu, Zhizhuo Yang, Hanhao Li, Pengzhi Gao, Wei Liu, Jian Luan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.20335v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-20335/"}
{"id": "arxiv:2601.20380", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad764cf86a2ecab81f3f730d6d26ea582544a355a90a8be55be220c98aabd7bf", "submitted_at": "2026-01-28T08:45:17Z", "updated_at": "2026-01-28T08:45:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20380v1"}, "authors": ["Le Zhang", "Yixiong Xiao", "Xinjiang Lu", "Jingjia Cao", "Yusai Zhao", "Jingbo Zhou", "Lang An", "Zikan Feng", "Wanxiang Sha", "Yu Shi", "Congxi Xiao", "Jian Xiong", "Yankai Zhang", "Hua Wu", "Haifeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2601.20380", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-20380/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ad764cf86a2ecab81f3f730d6d26ea582544a355a90a8be55be220c98aabd7bf", "submitted_at": "2026-01-28T08:45:17Z", "updated_at": "2026-01-28T08:45:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.20380v1"}}, "environments": ["android", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.20380"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20380"}], "name": "OmegaUse: Building a General-Purpose GUI Agent for Autonomous Task Execution", "publication": null, "published_at": "2026-01-28", "record_sha256": "d338a8186880c6980581850d1cb3cdf950624883ce98463b300d7f33daf5d54b", "relations": [], "resource_id": "arxiv:2601.20380", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.20380v1", "topics": ["data-generation", "reinforcement-learning", "grounding", "evaluation"], "updated_at": "2026-01-28"}, "text": "OmegaUse: Building a General-Purpose GUI Agent for Autonomous Task Execution\nResource type: paper\nReview status: curated\nScope: core\n\nOmegaUse combines exploratory and taxonomy-guided data synthesis with supervised fine-tuning followed by GRPO for GUI control. The report introduces OS-Nav for offline Android and Ubuntu evaluation alongside established grounding and action-prediction benchmarks.\nPlatforms: android, desktop, linux\nTopics: data-generation, reinforcement-learning, grounding, evaluation\nPublished: 2026-01-28\nAuthors: Le Zhang, Yixiong Xiao, Xinjiang Lu, Jingjia Cao, Yusai Zhao, Jingbo Zhou, Lang An, Zikan Feng, Wanxiang Sha, Yu Shi, Congxi Xiao, Jian Xiong, Yankai Zhang, Hua Wu, Haifeng Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.20380v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-20380/"}
{"id": "arxiv:2601.20650", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0139f7a1402d8dbfdecf65cc53c032495b2084f24cc38ff43841642c9102f29c", "submitted_at": "2026-01-28T14:35:23Z", "updated_at": "2026-08-26T22:31:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.20650v3"}, "authors": ["Jing Wu", "Wenjie Ai", "Daphne Barretto", "Yiye Chen", "Qingyu Chen", "Yuhang He", "Pranit Chawla", "Nicholas Gydé", "Yanan Jian", "Vibhav Vineet"], "canonical_source_url": "https://arxiv.org/abs/2601.20650", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-20650/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0139f7a1402d8dbfdecf65cc53c032495b2084f24cc38ff43841642c9102f29c", "submitted_at": "2026-01-28T14:35:23Z", "updated_at": "2026-08-26T22:31:44Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2601.20650v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.20650"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20650"}], "name": "OS-Marathon: Benchmarking Computer-Use Agents on Vast-Horizon, Repetitive Tasks", "publication": null, "published_at": "2026-01-28", "record_sha256": "8bee359fca20dbb0f017269cecd6f23a6897f6ae071c796045375d26b6ed7ca1", "relations": [], "resource_id": "arxiv:2601.20650", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.20650v3", "topics": ["evaluation", "long-horizon", "human-in-the-loop"], "updated_at": "2026-08-26"}, "text": "OS-Marathon: Benchmarking Computer-Use Agents on Vast-Horizon, Repetitive Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OS-Marathon to evaluate long repetitive computer-use workflows whose execution grows with the amount of input data. It studies failure accumulation under task decomposition and GraphDemo, which adapts agents using a demonstration of the recurring workflow.\nPlatforms: unspecified\nTopics: evaluation, long-horizon, human-in-the-loop\nPublished: 2026-01-28\nAuthors: Jing Wu, Wenjie Ai, Daphne Barretto, Yiye Chen, Qingyu Chen, Yuhang He, Pranit Chawla, Nicholas Gydé, Yanan Jian, Vibhav Vineet\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.20650v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-20650/"}
{"id": "arxiv:2601.20732", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1083cf063ccf1781f2cdfe5e266b75b52079fd12ad16d05755577b880c010aec", "submitted_at": "2026-01-28T16:06:31Z", "updated_at": "2026-03-25T10:25:50Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.20732v4"}, "authors": ["Ziwei Liu", "Borui Kang", "Hangjie Yuan", "Zixiang Zhao", "Wei Li", "Yifan Zhu", "Tao Feng"], "canonical_source_url": "https://arxiv.org/abs/2601.20732", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-20732/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1083cf063ccf1781f2cdfe5e266b75b52079fd12ad16d05755577b880c010aec", "submitted_at": "2026-01-28T16:06:31Z", "updated_at": "2026-03-25T10:25:50Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2601.20732v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.20732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.20732"}], "name": "Continual GUI Agents", "publication": null, "published_at": "2026-01-28", "record_sha256": "3332453dbe5164ad86840f3bae3549ead6801305d3fa281b40e4d7e1f3e32e0f", "relations": [], "resource_id": "arxiv:2601.20732", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.20732v4", "topics": ["grounding", "reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-03-25"}, "text": "Continual GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nContinual GUI Agents defines grounding under changing application domains and screen resolutions. Its GUI-AiF training framework uses separate point- and region-alignment rewards to adapt grounding as these distributions shift.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, robustness, evaluation\nPublished: 2026-01-28\nAuthors: Ziwei Liu, Borui Kang, Hangjie Yuan, Zixiang Zhao, Wei Li, Yifan Zhu, Tao Feng\nRecorded arXiv version: v4\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.20732v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-20732/"}
{"id": "arxiv:2601.21123", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3efa3cd0add437e327ea49896a3ee794daa178664dfb4096aba66fed27554b", "submitted_at": "2026-01-28T23:38:25Z", "updated_at": "2026-02-02T23:11:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21123v2"}, "authors": ["Tianyi Chen", "Yinheng Li", "Michael Solodko", "Sen Wang", "Nan Jiang", "Tingyuan Cui", "Junheng Hao", "Jongwoo Ko", "Sara Abdali", "Leon Xu", "Suzhen Zheng", "Hao Fan", "Pashmina Cameron", "Justin Wagle", "Kazuhito Koishida"], "canonical_source_url": "https://arxiv.org/abs/2601.21123", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-21123/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3efa3cd0add437e327ea49896a3ee794daa178664dfb4096aba66fed27554b", "submitted_at": "2026-01-28T23:38:25Z", "updated_at": "2026-02-02T23:11:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21123v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.21123"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21123"}], "name": "CUA-Skill: Develop Skills for Computer Using Agent", "publication": null, "published_at": "2026-01-28", "record_sha256": "b4c9d8d011627a62bd984c66c44fe47a814648cd43ab8636226b580cabc1291e", "relations": [], "resource_id": "arxiv:2601.21123", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.21123v2", "topics": ["tool-integration", "planning", "memory", "automation"], "updated_at": "2026-02-02"}, "text": "CUA-Skill: Develop Skills for Computer Using Agent\nResource type: paper\nReview status: curated\nScope: core\n\nCUA-Skill encodes Windows application knowledge in parameterized skills and execution/composition graphs. Its agent retrieves and instantiates these skills and uses memory for failure recovery, with end-to-end evaluation on WindowsAgentArena.\nPlatforms: desktop, windows\nTopics: tool-integration, planning, memory, automation\nPublished: 2026-01-28\nAuthors: Tianyi Chen, Yinheng Li, Michael Solodko, Sen Wang, Nan Jiang, Tingyuan Cui, Junheng Hao, Jongwoo Ko, Sara Abdali, Leon Xu, Suzhen Zheng, Hao Fan, Pashmina Cameron, Justin Wagle, Kazuhito Koishida\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.21123v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-21123/"}
{"id": "arxiv:2601.21253", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac3f148047f7e63a33780fed3fd1ca461bead2128914c46f353efab4675df444", "submitted_at": "2026-01-29T04:21:11Z", "updated_at": "2026-01-29T04:21:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21253v1"}, "authors": ["Wei Minn", "Biniam Fisseha Demissie", "Yan Naing Tun", "Jiakun Liu", "Mariano Ceccato", "Lwin Khin Shar", "David Lo"], "canonical_source_url": "https://arxiv.org/abs/2601.21253", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-21253/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ac3f148047f7e63a33780fed3fd1ca461bead2128914c46f353efab4675df444", "submitted_at": "2026-01-29T04:21:11Z", "updated_at": "2026-01-29T04:21:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21253v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "identifiers": {"arxiv": "2601.21253"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21253"}], "name": "CovAgent: Overcoming the 30% Curse of Mobile Application Coverage with Agentic AI and Dynamic Instrumentation", "publication": null, "published_at": "2026-01-29", "record_sha256": "aa05a979a6b997c67ffc06aaa6183837e5cf6fad3739786244c2d3021175839b", "relations": [], "resource_id": "arxiv:2601.21253", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.21253v1", "topics": ["automation", "mobile-navigation", "evaluation"], "updated_at": "2026-01-29"}, "text": "CovAgent: Overcoming the 30% Curse of Mobile Application Coverage with Agentic AI and Dynamic Instrumentation\nResource type: paper\nReview status: curated\nScope: core\n\nCovAgent augments Android GUI fuzzers by analyzing decompiled app code and component transitions to identify unsatisfied activation conditions. A second agent generates instrumentation that enables otherwise difficult-to-reach activities, with coverage-based evaluation.\nPlatforms: android\nTopics: automation, mobile-navigation, evaluation\nPublished: 2026-01-29\nAuthors: Wei Minn, Biniam Fisseha Demissie, Yan Naing Tun, Jiakun Liu, Mariano Ceccato, Lwin Khin Shar, David Lo\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2601.21253v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-21253/"}
{"id": "arxiv:2601.21352", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa36a9fbd0b189717b2a5311f06b6127d54f2eca09961cedd2a0b9d2a78b1b8b", "submitted_at": "2026-01-29T07:22:50Z", "updated_at": "2026-01-29T07:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21352v1"}, "authors": ["Ziyu Lu", "Tengjin Weng", "Yiying Yang", "Yuhang Zhao", "Xinxin Huang", "Wenhao Jiang"], "canonical_source_url": "https://arxiv.org/abs/2601.21352", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-21352/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa36a9fbd0b189717b2a5311f06b6127d54f2eca09961cedd2a0b9d2a78b1b8b", "submitted_at": "2026-01-29T07:22:50Z", "updated_at": "2026-01-29T07:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.21352v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.21352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21352"}], "name": "BEAP-Agent: Backtrackable Execution and Adaptive Planning for GUI Agents", "publication": null, "published_at": "2026-01-29", "record_sha256": "34953d40e08d30ac04793998267a49be65ba7254d2cb93cf05c4022a2ee38f4c", "relations": [], "resource_id": "arxiv:2601.21352", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.21352v1", "topics": ["planning", "state-tracking", "robustness", "long-horizon"], "updated_at": "2026-01-29"}, "text": "BEAP-Agent: Backtrackable Execution and Adaptive Planning for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nBEAP-Agent treats GUI execution as depth-first search with coordinated planning, execution and task-state tracking. It supports backtracking across multiple levels of earlier states and evaluates recovery on OSWorld.\nPlatforms: desktop\nTopics: planning, state-tracking, robustness, long-horizon\nPublished: 2026-01-29\nAuthors: Ziyu Lu, Tengjin Weng, Yiying Yang, Yuhang Zhao, Xinxin Huang, Wenhao Jiang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.21352v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-21352/"}
{"id": "arxiv:2601.21872", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68ae67856eb871d9a8f259dfb180bde02d29ee97fd7abf0e60719fc2c7abd7ab", "submitted_at": "2026-01-29T15:39:50Z", "updated_at": "2026-04-09T17:48:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21872v2"}, "authors": ["Yao Zhang", "Shijie Tang", "Zeyu Li", "Zhen Han", "Volker Tresp"], "canonical_source_url": "https://arxiv.org/abs/2601.21872", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-21872/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68ae67856eb871d9a8f259dfb180bde02d29ee97fd7abf0e60719fc2c7abd7ab", "submitted_at": "2026-01-29T15:39:50Z", "updated_at": "2026-04-09T17:48:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21872v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.21872"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21872"}], "name": "WebArbiter: A Principle-Guided Reasoning Process Reward Model for Web Agents", "publication": null, "published_at": "2026-01-29", "record_sha256": "4a24865168cda7ea87e0f6a61a6cfbd4446e7c17b9ac0f98666d409e1783f0b1", "relations": [], "resource_id": "arxiv:2601.21872", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.21872v2", "topics": ["reward-models", "reinforcement-learning", "evaluation", "web-navigation"], "updated_at": "2026-04-09"}, "text": "WebArbiter: A Principle-Guided Reasoning Process Reward Model for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebArbiter generates reasoned action-preference verdicts for web navigation, using reasoning distillation followed by reinforcement learning. It introduces WebPRMBench and evaluates both preference judgments and reward-guided trajectory search.\nPlatforms: web\nTopics: reward-models, reinforcement-learning, evaluation, web-navigation\nPublished: 2026-01-29\nAuthors: Yao Zhang, Shijie Tang, Zeyu Li, Zhen Han, Volker Tresp\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.21872v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-21872/"}
{"id": "arxiv:2601.21961", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1380697158d3a91334427cf067f3f91a751bd89d31103edecfa68171b28b8d66", "submitted_at": "2026-01-29T16:40:15Z", "updated_at": "2026-02-27T16:47:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21961v2"}, "authors": ["Kuai Yu", "Naicheng Yu", "Han Wang", "Rui Yang", "Huan Zhang"], "canonical_source_url": "https://arxiv.org/abs/2601.21961", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-21961/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1380697158d3a91334427cf067f3f91a751bd89d31103edecfa68171b28b8d66", "submitted_at": "2026-01-29T16:40:15Z", "updated_at": "2026-02-27T16:47:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.21961v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.21961"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.21961"}], "name": "How do Visual Attributes Influence Web Agents? A Comprehensive Evaluation of User Interface Design Factors", "publication": null, "published_at": "2026-01-29", "record_sha256": "b7f6a618a898afbcda74ad212c47da4286e14b77f2528484ce196a1b35881cb0", "relations": [], "resource_id": "arxiv:2601.21961", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.21961v2", "topics": ["ui-design", "evaluation", "web-navigation"], "updated_at": "2026-02-27"}, "text": "How do Visual Attributes Influence Web Agents? A Comprehensive Evaluation of User Interface Design Factors\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VAF, a controlled pipeline that varies webpage appearance while retaining item semantics. It measures both agent clicks and reasoning references to study which visual design factors change browsing decisions.\nPlatforms: web\nTopics: ui-design, evaluation, web-navigation\nPublished: 2026-01-29\nAuthors: Kuai Yu, Naicheng Yu, Han Wang, Rui Yang, Huan Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.21961v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-21961/"}
{"id": "arxiv:2601.22149", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aedb5ed90c91d72f62f5ab60be5374b6f431f887425bc760883ce3fb85a46d43", "submitted_at": "2026-01-29T18:59:07Z", "updated_at": "2026-04-17T18:27:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22149v2"}, "authors": ["Hang Ding", "Peidong Liu", "Junqiao Wang", "Ziwei Ji", "Meng Cao", "Rongzhao Zhang", "Lynn Ai", "Eric Yang", "Tianyu Shi", "Lei Yu"], "canonical_source_url": "https://arxiv.org/abs/2601.22149", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-22149/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aedb5ed90c91d72f62f5ab60be5374b6f431f887425bc760883ce3fb85a46d43", "submitted_at": "2026-01-29T18:59:07Z", "updated_at": "2026-04-17T18:27:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2601.22149v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.22149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22149"}], "name": "DynaWeb: Model-Based Reinforcement Learning of Web Agents", "publication": null, "published_at": "2026-01-29", "record_sha256": "0b888260ad8f0f5ca0713c763ca4c8955c9641d1584fd9293ecb35ada3ff8e29", "relations": [], "resource_id": "arxiv:2601.22149", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.22149v2", "topics": ["world-models", "reinforcement-learning", "data-generation", "web-navigation"], "updated_at": "2026-04-17"}, "text": "DynaWeb: Model-Based Reinforcement Learning of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDynaWeb learns a world model that predicts web-page representations after actions and uses it to generate policy-training rollouts. It interleaves these rollouts with expert trajectories and evaluates the trained web agents on WebArena and WebVoyager.\nPlatforms: web\nTopics: world-models, reinforcement-learning, data-generation, web-navigation\nPublished: 2026-01-29\nAuthors: Hang Ding, Peidong Liu, Junqiao Wang, Ziwei Ji, Meng Cao, Rongzhao Zhang, Lynn Ai, Eric Yang, Tianyu Shi, Lei Yu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.22149v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-22149/"}
{"id": "arxiv:2601.22528", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0b3dbaacdf640a1314a40688133f7386c0eb8b901c43393ad21ec887814a606", "submitted_at": "2026-01-30T04:01:21Z", "updated_at": "2026-01-30T04:01:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22528v1"}, "authors": ["Hongze Mi", "Yibo Feng", "WenJie Lu", "Song Cao", "Jinyuan Li", "Yanming Li", "Xuelin Zhang", "Haotian Luo", "Songyang Peng", "He Cui", "Tengfei Tian", "Jun Fang", "Hua Chai", "Naiqiang Tan"], "canonical_source_url": "https://arxiv.org/abs/2601.22528", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-22528/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c0b3dbaacdf640a1314a40688133f7386c0eb8b901c43393ad21ec887814a606", "submitted_at": "2026-01-30T04:01:21Z", "updated_at": "2026-01-30T04:01:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22528v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.22528"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22528"}], "name": "Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent Evolution", "publication": null, "published_at": "2026-01-30", "record_sha256": "260deae849904a9fa1c0715d64b6011684b88a86b02b22b86a572a3bb622354e", "relations": [], "resource_id": "arxiv:2601.22528", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.22528v1", "topics": ["memory", "training-free", "cross-app", "long-horizon"], "updated_at": "2026-01-30"}, "text": "Darwinian Memory: A Training-Free Self-Regulating Memory System for GUI Agent Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nDarwinian Memory splits GUI trajectories into reusable units and updates the memory by retaining useful paths and pruning poor ones. It evaluates this training-free memory mechanism on long multi-application tasks.\nPlatforms: unspecified\nTopics: memory, training-free, cross-app, long-horizon\nPublished: 2026-01-30\nAuthors: Hongze Mi, Yibo Feng, WenJie Lu, Song Cao, Jinyuan Li, Yanming Li, Xuelin Zhang, Haotian Luo, Songyang Peng, He Cui, Tengfei Tian, Jun Fang, Hua Chai, Naiqiang Tan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.22528v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-22528/"}
{"id": "arxiv:2601.22781", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d66365682d0ea4cee744c6c59d3a585c9272a31da5f8ff7c27983d719f070a7", "submitted_at": "2026-01-30T10:03:20Z", "updated_at": "2026-01-30T10:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22781v1"}, "authors": ["Linjia Kang", "Zhimin Wang", "Yongkang Zhang", "Duo Wu", "Jinghe Wang", "Ming Ma", "Haopeng Yan", "Zhi Wang"], "canonical_source_url": "https://arxiv.org/abs/2601.22781", "citation_url": "https://awesomecua.github.io/resources/arxiv-2601-22781/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d66365682d0ea4cee744c6c59d3a585c9272a31da5f8ff7c27983d719f070a7", "submitted_at": "2026-01-30T10:03:20Z", "updated_at": "2026-01-30T10:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2601.22781v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2601.22781"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2601.22781"}], "name": "Learning with Challenges: Adaptive Difficulty-Aware Data Generation for Mobile GUI Agent Training", "publication": null, "published_at": "2026-01-30", "record_sha256": "ca9b32ff010f2543b11b0987fc655099ff50e0fb9dd5e02f0fa8b82dd62829ad", "relations": [], "resource_id": "arxiv:2601.22781", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2601.22781v1", "topics": ["data-generation", "mobile-navigation", "evaluation"], "updated_at": "2026-01-30"}, "text": "Learning with Challenges: Adaptive Difficulty-Aware Data Generation for Mobile GUI Agent Training\nResource type: paper\nReview status: curated\nScope: core\n\nMobileGen estimates a mobile agent’s strengths across structural and semantic task difficulty, then samples new training targets around that profile. A controllable multi-agent generator creates the corresponding instructions and GUI trajectories.\nPlatforms: unspecified\nTopics: data-generation, mobile-navigation, evaluation\nPublished: 2026-01-30\nAuthors: Linjia Kang, Zhimin Wang, Yongkang Zhang, Duo Wu, Jinghe Wang, Ming Ma, Haopeng Yan, Zhi Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2601.22781v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2601-22781/"}
{"id": "arxiv:2602.00575", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86225f2e52e1394e2a0a56b1e82327d756bff95739e01ce7cc5898af49944aba", "submitted_at": "2026-01-31T07:36:54Z", "updated_at": "2026-07-25T06:54:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.00575v2"}, "authors": ["Chaoqun Cui", "Jing Huang", "Shijing Wang", "Liming Zheng", "Qingchao Kong", "Zhixiong Zeng"], "canonical_source_url": "https://arxiv.org/abs/2602.00575", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-00575/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86225f2e52e1394e2a0a56b1e82327d756bff95739e01ce7cc5898af49944aba", "submitted_at": "2026-01-31T07:36:54Z", "updated_at": "2026-07-25T06:54:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.00575v2"}}, "environments": ["desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.00575"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.00575"}], "name": "Agentic Reward Modeling: Verifying GUI Agent via Progressive Trajectory-Grounded Interaction", "publication": null, "published_at": "2026-01-31", "record_sha256": "e788769cd2cbd7df2d2aff8fa63b08de113120b1ec9bde04862d07fd5823d24d", "relations": [], "resource_id": "arxiv:2602.00575", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.00575v2", "topics": ["reward-models", "tool-integration", "evaluation", "efficiency"], "updated_at": "2026-07-25"}, "text": "Agentic Reward Modeling: Verifying GUI Agent via Progressive Trajectory-Grounded Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nVAGEN verifies GUI trajectories by first extracting readily available evidence and escalating to tool-based environment probes when needed. Evaluation on OSWorld-Verified and AndroidWorld studies verification accuracy and its cost.\nPlatforms: desktop, android\nTopics: reward-models, tool-integration, evaluation, efficiency\nPublished: 2026-01-31\nAuthors: Chaoqun Cui, Jing Huang, Shijing Wang, Liming Zheng, Qingchao Kong, Zhixiong Zeng\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.00575v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-00575/"}
{"id": "arxiv:2602.01576", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c29d1404b02c4f5ef2cdaf78bdb0f9062b1b85ec1a154408bc8f49446c28d08", "submitted_at": "2026-02-02T03:12:16Z", "updated_at": "2026-05-23T02:41:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.01576v2"}, "authors": ["Woosung Koh", "Sungjun Han", "Segyu Lee", "Se-Young Yun", "Jamin Shin"], "canonical_source_url": "https://arxiv.org/abs/2602.01576", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-01576/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c29d1404b02c4f5ef2cdaf78bdb0f9062b1b85ec1a154408bc8f49446c28d08", "submitted_at": "2026-02-02T03:12:16Z", "updated_at": "2026-05-23T02:41:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.01576v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.01576"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.01576"}], "name": "Generative Visual Code Mobile World Models", "publication": null, "published_at": "2026-02-02", "record_sha256": "34f3bacadb274b022f66c6d0ed4627c03fea16ef61291ea577cb9c5f8fdc3621", "relations": [], "resource_id": "arxiv:2602.01576", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.01576v2", "topics": ["world-models", "data-generation", "screen-understanding", "mobile-navigation"], "updated_at": "2026-05-23"}, "text": "Generative Visual Code Mobile World Models\nResource type: paper\nReview status: curated\nScope: core\n\ngWorld predicts future mobile GUI states as renderable web code and supplies an automated pipeline for code-based training data. The study evaluates visual state prediction and how stronger world models affect downstream mobile policies.\nPlatforms: unspecified\nTopics: world-models, data-generation, screen-understanding, mobile-navigation\nPublished: 2026-02-02\nAuthors: Woosung Koh, Sungjun Han, Segyu Lee, Se-Young Yun, Jamin Shin\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.01576v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-01576/"}
{"id": "arxiv:2602.01725", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1a6f98c0ab0a221c169796d9680e4b2a7bf7a0775c0493254f47d5e0a9b3c4e", "submitted_at": "2026-02-02T07:04:06Z", "updated_at": "2026-02-02T07:04:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.01725v1"}, "authors": ["Yurun Chen", "Zeyi Liao", "Ping Yin", "Taotao Xie", "Keting Yin", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2602.01725", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-01725/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1a6f98c0ab0a221c169796d9680e4b2a7bf7a0775c0493254f47d5e0a9b3c4e", "submitted_at": "2026-02-02T07:04:06Z", "updated_at": "2026-02-02T07:04:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.01725v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:38+00:00", "role": "primary", "sha256": "6e1266c5a47cd667d681d3143a0331b74b6f64416b7e01f44a93a001c38ef74b", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.01725v1"}], "identifiers": {"arxiv": "2602.01725"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.01725"}], "name": "SafePred: A Predictive Guardrail for Computer-Using Agents via World Models", "publication": null, "published_at": "2026-02-02", "record_sha256": "5d8a3271c562cbc5d1bce7332d530733a7fa1c8913cffbbe1f5f93deb7add849", "relations": [], "resource_id": "arxiv:2602.01725", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.01725v1", "topics": ["safety-security", "world-models", "planning", "evaluation"], "updated_at": "2026-02-02"}, "text": "SafePred: A Predictive Guardrail for Computer-Using Agents via World Models\nResource type: paper\nReview status: curated\nScope: core\n\nSafePred uses world-model risk forecasts to guide action filtering and task replanning. Its desktop and web safety evaluations distinguish policy compliance, task success and success under the specified policy.\nPlatforms: desktop, web\nTopics: safety-security, world-models, planning, evaluation\nPublished: 2026-02-02\nAuthors: Yurun Chen, Zeyi Liao, Ping Yin, Taotao Xie, Keting Yin, Shengyu Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.01725v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-01725/"}
{"id": "arxiv:2602.02419", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9557fc608a7fd7667746e9b3e06bc699b1796cdd77fe5e975ba1bd5a64690c1", "submitted_at": "2026-02-02T18:22:45Z", "updated_at": "2026-02-03T08:31:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.02419v2"}, "authors": ["Qingni Wang", "Yue Fan", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2602.02419", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-02419/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c9557fc608a7fd7667746e9b3e06bc699b1796cdd77fe5e975ba1bd5a64690c1", "submitted_at": "2026-02-02T18:22:45Z", "updated_at": "2026-02-03T08:31:29Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.02419v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.02419"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02419"}], "name": "SafeGround: Know When to Trust GUI Grounding Models via Uncertainty Calibration", "publication": null, "published_at": "2026-02-02", "record_sha256": "155ccb3895fe337cb2a06e1bf87975e2392cbf545d8f6afe57cdd353030b6a60", "relations": [], "resource_id": "arxiv:2602.02419", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.02419v2", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-02-03"}, "text": "SafeGround: Know When to Trust GUI Grounding Models via Uncertainty Calibration\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SafeGround, which estimates GUI-grounding uncertainty from the spatial dispersion of sampled predictions and calibrates a decision threshold before testing. It studies risk-aware selection on ScreenSpot-Pro, with statistical control scoped to the threshold’s false-discovery rate.\nPlatforms: unspecified\nTopics: grounding, robustness, evaluation\nPublished: 2026-02-02\nAuthors: Qingni Wang, Yue Fan, Xin Eric Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.02419v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-02419/"}
{"id": "arxiv:2602.02468", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5b730fa81936d6a1b2b0a8b74a522a22de39b977929165c44a5050e490c301f", "submitted_at": "2026-02-02T18:50:07Z", "updated_at": "2026-02-02T18:50:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02468v1"}, "authors": ["Aiden Yiliu Li", "Xinyue Hao", "Shilong Liu", "Mengdi Wang"], "canonical_source_url": "https://arxiv.org/abs/2602.02468", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-02468/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5b730fa81936d6a1b2b0a8b74a522a22de39b977929165c44a5050e490c301f", "submitted_at": "2026-02-02T18:50:07Z", "updated_at": "2026-02-02T18:50:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02468v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.02468"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02468"}], "name": "Avenir-Web: Human-Experience-Imitating Multimodal Web Agents with Mixture of Grounding Experts", "publication": null, "published_at": "2026-02-02", "record_sha256": "5518f3ccc6233c1f99c9560a75496c9152754366d2db1e949c710fe16d0b747c", "relations": [], "resource_id": "arxiv:2602.02468", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.02468v1", "topics": ["grounding", "planning", "memory", "web-navigation"], "updated_at": "2026-02-02"}, "text": "Avenir-Web: Human-Experience-Imitating Multimodal Web Agents with Mixture of Grounding Experts\nResource type: paper\nReview status: curated\nScope: core\n\nAvenir-Web combines multiple grounding specialists with procedural experience, a task checklist and adaptive memory. It evaluates this planning and execution framework on live Online-Mind2Web tasks.\nPlatforms: web\nTopics: grounding, planning, memory, web-navigation\nPublished: 2026-02-02\nAuthors: Aiden Yiliu Li, Xinyue Hao, Shilong Liu, Mengdi Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.02468v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-02468/"}
{"id": "arxiv:2602.02548", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b0dbff31dcf8c966bebd367ae05984e46a7fd13b2cfb1ad5ef061332b43f00", "submitted_at": "2026-01-30T08:38:05Z", "updated_at": "2026-01-30T08:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02548v1"}, "authors": ["Xiaoce Wang", "Guibin Zhang", "Junzhe Li", "Jinzhe Tu", "Chun Li", "Ming Li"], "canonical_source_url": "https://arxiv.org/abs/2602.02548", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-02548/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b0dbff31dcf8c966bebd367ae05984e46a7fd13b2cfb1ad5ef061332b43f00", "submitted_at": "2026-01-30T08:38:05Z", "updated_at": "2026-01-30T08:38:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.02548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02548"}], "name": "ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents", "publication": null, "published_at": "2026-01-30", "record_sha256": "60ef5c36738f1e1869b3b6a40acecfd933707f895a8a8cc791e0213deb26cdd7", "relations": [], "resource_id": "arxiv:2602.02548", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.02548v1", "topics": ["tool-integration", "grounding", "data-generation"], "updated_at": "2026-01-30"}, "text": "ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nToolTok represents GUI operations with learned tool tokens and frames interaction as progressive tool selection. Semantic anchoring and an easy-to-hard curriculum teach tool meanings before visual pathfinding, with evaluation across grounding scenarios.\nPlatforms: unspecified\nTopics: tool-integration, grounding, data-generation\nPublished: 2026-01-30\nAuthors: Xiaoce Wang, Guibin Zhang, Junzhe Li, Jinzhe Tu, Chun Li, Ming Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.02548v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-02548/"}
{"id": "arxiv:2602.02995", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a01499ce704bb7916a3157ff2d3fc755a1115d9b27c7e529c153bce360a5404", "submitted_at": "2026-02-03T02:07:12Z", "updated_at": "2026-02-03T02:07:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02995v1"}, "authors": ["Sizhe Tang", "Rongqian Chen", "Tian Lan"], "canonical_source_url": "https://arxiv.org/abs/2602.02995", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-02995/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a01499ce704bb7916a3157ff2d3fc755a1115d9b27c7e529c153bce360a5404", "submitted_at": "2026-02-03T02:07:12Z", "updated_at": "2026-02-03T02:07:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.02995v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.02995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.02995"}], "name": "Agent Alpha: Tree Search Unifying Generation, Exploration and Evaluation for Computer-Use Agents", "publication": null, "published_at": "2026-02-03", "record_sha256": "cd6db06216c5e82728828bbe60bfa95f51f94c9f4aaaa4ef9acf32f36ca163d6", "relations": [], "resource_id": "arxiv:2602.02995", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.02995v1", "topics": ["planning", "efficiency", "evaluation"], "updated_at": "2026-02-03"}, "text": "Agent Alpha: Tree Search Unifying Generation, Exploration and Evaluation for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAgent Alpha uses step-level Monte Carlo tree search to reuse successful prefixes and recover from early GUI mistakes. Comparison-based evaluation and diversity-constrained expansion guide the search, with experiments on OSWorld.\nPlatforms: desktop\nTopics: planning, efficiency, evaluation\nPublished: 2026-02-03\nAuthors: Sizhe Tang, Rongqian Chen, Tian Lan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.02995v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-02995/"}
{"id": "arxiv:2602.03792", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86702018906e64294d7b8daea618c5f152e8b94b604d27d6b164e82aff462564", "submitted_at": "2026-02-03T17:55:04Z", "updated_at": "2026-02-03T17:55:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.03792v1"}, "authors": ["Xilong Wang", "Yinuo Liu", "Zhun Wang", "Dawn Song", "Neil Gong"], "canonical_source_url": "https://arxiv.org/abs/2602.03792", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-03792/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86702018906e64294d7b8daea618c5f152e8b94b604d27d6b164e82aff462564", "submitted_at": "2026-02-03T17:55:04Z", "updated_at": "2026-02-03T17:55:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.03792v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.03792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.03792"}], "name": "WebSentinel: Detecting and Localizing Prompt Injection Attacks for Web Agents", "publication": null, "published_at": "2026-02-03", "record_sha256": "35059dde12900957052ff95069eca97716ef05b0025f63080a390435b5c39d6e", "relations": [], "resource_id": "arxiv:2602.03792", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.03792v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-02-03"}, "text": "WebSentinel: Detecting and Localizing Prompt Injection Attacks for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebSentinel identifies suspicious webpage segments and checks each against its surrounding page context to detect and locate prompt injections. Its evaluation uses both contaminated and clean webpage datasets.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2026-02-03\nAuthors: Xilong Wang, Yinuo Liu, Zhun Wang, Dawn Song, Neil Gong\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.03792v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-03792/"}
{"id": "arxiv:2602.05354", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76fbeae086fb9ef21f820b02777a2ceb66ee37f8f1c978c7426589ea4e888e49", "submitted_at": "2026-02-05T06:24:23Z", "updated_at": "2026-09-08T14:40:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.05354v3"}, "authors": ["Syed Nazmus Sakib", "Nafiul Haque", "Tapodhir Karmakar Taton", "Shahrear Bin Amin", "Shifat E. Arman"], "canonical_source_url": "https://arxiv.org/abs/2602.05354", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-05354/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76fbeae086fb9ef21f820b02777a2ceb66ee37f8f1c978c7426589ea4e888e49", "submitted_at": "2026-02-05T06:24:23Z", "updated_at": "2026-09-08T14:40:01Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.05354v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.05354"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05354"}], "name": "Do Web Agents Investigate Before They Decide?", "publication": null, "published_at": "2026-02-05", "record_sha256": "83e0605f86be47c57eb418ae6ff733557dd831f74826c57e417cc62e00439524", "relations": [], "resource_id": "arxiv:2602.05354", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.05354v3", "topics": ["evaluation", "web-navigation", "planning"], "updated_at": "2026-09-08"}, "text": "Do Web Agents Investigate Before They Decide?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MIRAGE, a web decision benchmark where agents must actively uncover evidence that can contradict the visible context. Separate investigation, reasoning, decision, and hallucination measures diagnose failures beyond reaching the right page.\nPlatforms: web\nTopics: evaluation, web-navigation, planning\nPublished: 2026-02-05\nAuthors: Syed Nazmus Sakib, Nafiul Haque, Tapodhir Karmakar Taton, Shahrear Bin Amin, Shifat E. Arman\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.05354v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-05354/"}
{"id": "arxiv:2602.05429", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84d822655c810958cff6035584542d7e151d765734ec7030f841c15561ee5677", "submitted_at": "2026-02-05T08:19:39Z", "updated_at": "2026-02-05T08:19:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05429v1"}, "authors": ["Rui Lv", "Juncheng Mo", "Tianyi Chu", "Chen Rao", "Hongyi Jing", "Jiajie Teng", "Jiafu Chen", "Shiqi Zhang", "Liangzi Ding", "Shuo Fang", "Huaizhong Lin", "Ziqiang Dang", "Chenguang Ma", "Lei Zhao"], "canonical_source_url": "https://arxiv.org/abs/2602.05429", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-05429/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "84d822655c810958cff6035584542d7e151d765734ec7030f841c15561ee5677", "submitted_at": "2026-02-05T08:19:39Z", "updated_at": "2026-02-05T08:19:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05429v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.05429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05429"}], "name": "M$^2$-Miner: Multi-Agent Enhanced MCTS for Mobile GUI Agent Data Mining", "publication": null, "published_at": "2026-02-05", "record_sha256": "40255856e8bd77f68afc60357fa4b02d4776441bfcaaf02358347c62c8a6c7fe", "relations": [], "resource_id": "arxiv:2602.05429", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.05429v1", "topics": ["data-generation", "mobile-navigation", "planning"], "updated_at": "2026-02-05"}, "text": "M$^2$-Miner: Multi-Agent Enhanced MCTS for Mobile GUI Agent Data Mining\nResource type: paper\nReview status: curated\nScope: core\n\nM2-Miner uses multi-agent-guided Monte Carlo tree search to collect mobile GUI intent–trajectory pairs. Intent recycling and progressive collector training aim to increase the diversity and usefulness of the mined interaction data.\nPlatforms: unspecified\nTopics: data-generation, mobile-navigation, planning\nPublished: 2026-02-05\nAuthors: Rui Lv, Juncheng Mo, Tianyi Chu, Chen Rao, Hongyi Jing, Jiajie Teng, Jiafu Chen, Shiqi Zhang, Liangzi Ding, Shuo Fang, Huaizhong Lin, Ziqiang Dang, Chenguang Ma, Lei Zhao\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.05429v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-05429/"}
{"id": "arxiv:2602.05832", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f51eb0b7102a37fcc705c53640c5a1d30fcc115cedcdca4a00493ed45b9d5507", "submitted_at": "2026-02-05T16:21:43Z", "updated_at": "2026-02-05T16:21:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05832v1"}, "authors": ["Han Xiao", "Guozhi Wang", "Hao Wang", "Shilong Liu", "Yuxiang Chai", "Yue Pan", "Yufeng Zhou", "Xiaoxin Chen", "Yafei Wen", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2602.05832", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-05832/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f51eb0b7102a37fcc705c53640c5a1d30fcc115cedcdca4a00493ed45b9d5507", "submitted_at": "2026-02-05T16:21:43Z", "updated_at": "2026-02-05T16:21:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.05832v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.05832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.05832"}], "name": "UI-Mem: Self-Evolving Experience Memory for Online Reinforcement Learning in Mobile GUI Agents", "publication": null, "published_at": "2026-02-05", "record_sha256": "03de73ab8abf529ee5be24df1e24d2f76c229b6b010926235f79984c38391e4a", "relations": [], "resource_id": "arxiv:2602.05832", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.05832v1", "topics": ["memory", "reinforcement-learning", "cross-app"], "updated_at": "2026-02-05"}, "text": "UI-Mem: Self-Evolving Experience Memory for Online Reinforcement Learning in Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUI-Mem stores workflows, skills and failure patterns as reusable templates to guide online GUI reinforcement learning. It varies guidance within rollout groups and continually updates the memory so an unguided policy can learn from guided experiences.\nPlatforms: unspecified\nTopics: memory, reinforcement-learning, cross-app\nPublished: 2026-02-05\nAuthors: Han Xiao, Guozhi Wang, Hao Wang, Shilong Liu, Yuxiang Chai, Yue Pan, Yufeng Zhou, Xiaoxin Chen, Yafei Wen, Hongsheng Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.05832v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-05832/"}
{"id": "arxiv:2602.06075", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0692d6936a3e3c20a7ad61c69ae770f5669c3783ff837092254f8b3fb170cd0a", "submitted_at": "2026-02-03T17:01:59Z", "updated_at": "2026-08-27T06:48:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.06075v3"}, "authors": ["Guangyi Liu", "Pengxiang Zhao", "Yaozhen Liang", "Qinyi Luo", "Shunye Tang", "Yuxiang Chai", "Weifeng Lin", "Han Xiao", "WenHao Wang", "Siheng Chen", "Zhengxi Lu", "Gao Wu", "Hao Wang", "Liang Liu", "Yong Liu"], "canonical_source_url": "https://arxiv.org/abs/2602.06075", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-06075/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0692d6936a3e3c20a7ad61c69ae770f5669c3783ff837092254f8b3fb170cd0a", "submitted_at": "2026-02-03T17:01:59Z", "updated_at": "2026-08-27T06:48:18Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.06075v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.06075"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06075"}], "name": "MemGUI-Bench: Benchmarking Memory of Mobile GUI Agents in Dynamic Environments", "publication": null, "published_at": "2026-02-03", "record_sha256": "bd397e6f469dc7c6c2c31a2a5fc097c515e4092f164de642f1308d10aa106870", "relations": [], "resource_id": "arxiv:2602.06075", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.06075v3", "topics": ["memory", "evaluation", "mobile-navigation", "cross-app"], "updated_at": "2026-08-27"}, "text": "MemGUI-Bench: Benchmarking Memory of Mobile GUI Agents in Dynamic Environments\nResource type: paper\nReview status: curated\nScope: core\n\nMemGUI-Bench evaluates short-term retention and cross-session experience reuse in mobile GUI agents using paired tasks and repeated-attempt protocols. Its staged evaluator measures memory fidelity, learning and execution efficiency.\nPlatforms: unspecified\nTopics: memory, evaluation, mobile-navigation, cross-app\nPublished: 2026-02-03\nAuthors: Guangyi Liu, Pengxiang Zhao, Yaozhen Liang, Qinyi Luo, Shunye Tang, Yuxiang Chai, Weifeng Lin, Han Xiao, WenHao Wang, Siheng Chen, Zhengxi Lu, Gao Wu, Hao Wang, Liang Liu, Yong Liu\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.06075v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-06075/"}
{"id": "arxiv:2602.06351", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2591e0dab6a84b5db2631ddc94a2c0b65b13a97d06d411ab6ad4e243fe3d3e1e", "submitted_at": "2026-02-06T03:27:55Z", "updated_at": "2026-02-06T03:27:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06351v1"}, "authors": ["Longhui Ma", "Di Zhao", "Siwei Wang", "Zhao Lv", "Miao Wang"], "canonical_source_url": "https://arxiv.org/abs/2602.06351", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-06351/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2591e0dab6a84b5db2631ddc94a2c0b65b13a97d06d411ab6ad4e243fe3d3e1e", "submitted_at": "2026-02-06T03:27:55Z", "updated_at": "2026-02-06T03:27:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06351v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.06351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06351"}], "name": "Trifuse: Enhancing Attention-Based GUI Grounding via Multimodal Fusion", "publication": null, "published_at": "2026-02-06", "record_sha256": "890391feeb2b6eec8433b4db5666f70468375c5ad1241dbd45cad58794101b62", "relations": [], "resource_id": "arxiv:2602.06351", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.06351v1", "topics": ["grounding", "screen-understanding", "training-free"], "updated_at": "2026-02-06"}, "text": "Trifuse: Enhancing Attention-Based GUI Grounding via Multimodal Fusion\nResource type: paper\nReview status: curated\nScope: core\n\nTrifuse combines attention-based localization with OCR text cues and icon captions to ground GUI instructions. Its fusion procedure favors cross-modal agreement and sharp location peaks, and is evaluated without task-specific fine-tuning.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, training-free\nPublished: 2026-02-06\nAuthors: Longhui Ma, Di Zhao, Siwei Wang, Zhao Lv, Miao Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.06351v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-06351/"}
{"id": "arxiv:2602.06391", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "852813305299178baeda3540cbad539dc30b840681e7bbdaed34dac1341dc0de", "submitted_at": "2026-02-06T05:14:11Z", "updated_at": "2026-02-06T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06391v1"}, "authors": ["Zhongyin Zhao", "Yuan Liu", "Yikun Liu", "Haicheng Wang", "Le Tian", "Xiao Zhou", "Yangxiu You", "Zilin Yu", "Yang Yu", "Jie Zhou"], "canonical_source_url": "https://arxiv.org/abs/2602.06391", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-06391/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "852813305299178baeda3540cbad539dc30b840681e7bbdaed34dac1341dc0de", "submitted_at": "2026-02-06T05:14:11Z", "updated_at": "2026-02-06T05:14:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.06391v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.06391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.06391"}], "name": "POINTS-GUI-G: GUI-Grounding Journey", "publication": null, "published_at": "2026-02-06", "record_sha256": "406356f24b54091f7303965b5290ca56b4743278fa151933d279a01bf10382a5", "relations": [], "resource_id": "arxiv:2602.06391", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.06391v1", "topics": ["grounding", "reinforcement-learning"], "updated_at": "2026-02-06"}, "text": "POINTS-GUI-G: GUI-Grounding Journey\nResource type: paper\nReview status: curated\nScope: core\n\nStudies GUI-grounding development from a weakly grounded base model through data preparation, visual training, and verifiable rewards.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\nPublished: 2026-02-06\nAuthors: Zhongyin Zhao, Yuan Liu, Yikun Liu, Haicheng Wang, Le Tian, Xiao Zhou, Yangxiu You, Zilin Yu, Yang Yu, Jie Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.06391v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-06391/"}
{"id": "arxiv:2602.07153", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce896d4d846bf07fedc68b7f952532c4fefe24dcb49b51556e0d6c2bc648bb91", "submitted_at": "2026-02-06T19:55:26Z", "updated_at": "2026-04-12T06:09:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.07153v2"}, "authors": ["Jinbiao Wei", "Yilun Zhao", "Kangqi Ni", "Arman Cohan"], "canonical_source_url": "https://arxiv.org/abs/2602.07153", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-07153/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce896d4d846bf07fedc68b7f952532c4fefe24dcb49b51556e0d6c2bc648bb91", "submitted_at": "2026-02-06T19:55:26Z", "updated_at": "2026-04-12T06:09:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.07153v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.07153"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.07153"}], "name": "ANCHOR: Branch-Point Data Generation for GUI Agents", "publication": null, "published_at": "2026-02-06", "record_sha256": "276ded089468c3073eed03d8ab4c6ca41c5e786a282191d109f0e066313009ee", "relations": [], "resource_id": "arxiv:2602.07153", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.07153v2", "topics": ["data-generation", "planning", "evaluation"], "updated_at": "2026-04-12"}, "text": "ANCHOR: Branch-Point Data Generation for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nANCHOR expands verified desktop demonstrations by generating new tasks at meaningful branch points and executing their instructions. State-aware verification and action filtering refine the resulting trajectories, which are evaluated through fine-tuned agents on OSWorld and WindowsAgentArena.\nPlatforms: desktop, windows\nTopics: data-generation, planning, evaluation\nPublished: 2026-02-06\nAuthors: Jinbiao Wei, Yilun Zhao, Kangqi Ni, Arman Cohan\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.07153v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-07153/"}
{"id": "arxiv:2602.07283", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d556ebbfbf4fc5875395b449a92b8ff27a7ed59d41bc062ac60fcf175f9064f8", "submitted_at": "2026-02-07T00:19:08Z", "updated_at": "2026-02-07T00:19:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07283v1"}, "authors": ["Ruijia Cheng", "Jenny T. Liang", "Eldon Schoop", "Jeffrey Nichols"], "canonical_source_url": "https://arxiv.org/abs/2602.07283", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-07283/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d556ebbfbf4fc5875395b449a92b8ff27a7ed59d41bc062ac60fcf175f9064f8", "submitted_at": "2026-02-07T00:19:08Z", "updated_at": "2026-02-07T00:19:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.07283v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.07283"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.07283"}], "name": "Mapping the Design Space of User Experience for Computer Use Agents", "publication": null, "published_at": "2026-02-07", "record_sha256": "cd5ea59d40cba4362281abc62f34619fc8339f8219d3a00e7e807a7ab7c9202f", "relations": [], "resource_id": "arxiv:2602.07283", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.07283v1", "topics": ["human-in-the-loop", "ui-design", "evaluation"], "updated_at": "2026-02-07"}, "text": "Mapping the Design Space of User Experience for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDevelops a taxonomy of computer-use-agent interaction design through system review and practitioner interviews. A Wizard-of-Oz study then examines how users respond to normal, erroneous, and risky web-based execution, focusing on control, explanation, prompts, and mental models.\nPlatforms: web\nTopics: human-in-the-loop, ui-design, evaluation\nPublished: 2026-02-07\nAuthors: Ruijia Cheng, Jenny T. Liang, Eldon Schoop, Jeffrey Nichols\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.07283v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-07283/"}
{"id": "arxiv:2602.08995", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "853f64c020f47129be01778d5212ade8ef78251959ad634ee04ad0fb9fe2989c", "submitted_at": "2026-02-09T18:41:15Z", "updated_at": "2026-06-25T04:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.08995v2"}, "authors": ["Yuting Ning", "Jaylen Jones", "Zhehao Zhang", "Chentao Ye", "Weitong Ruan", "Junyi Li", "Rahul Gupta", "Huan Sun"], "canonical_source_url": "https://arxiv.org/abs/2602.08995", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-08995/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "853f64c020f47129be01778d5212ade8ef78251959ad634ee04ad0fb9fe2989c", "submitted_at": "2026-02-09T18:41:15Z", "updated_at": "2026-06-25T04:36:12Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.08995v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:26:41+00:00", "role": "primary", "sha256": "487a5d21d2385320315eb5a4a2c76e1017973d6027e219be6d31d9328fd5ff9a", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.08995v2"}], "identifiers": {"arxiv": "2602.08995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.08995"}], "name": "When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents", "publication": null, "published_at": "2026-02-09", "record_sha256": "9bba5bc25f4f35b1213308ed7e398de97ba8204ee13501e70498096f12345d91", "relations": [], "resource_id": "arxiv:2602.08995", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.08995v2", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-25"}, "text": "When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMisActBench provides action-level labels for task alignment, while DeAction checks proposed computer-use actions and requests revisions through structured feedback. The paper separately evaluates offline detection and interactive operation in OSWorld and RedTeamCUA.\nPlatforms: desktop, web\nTopics: safety-security, evaluation, robustness\nPublished: 2026-02-09\nAuthors: Yuting Ning, Jaylen Jones, Zhehao Zhang, Chentao Ye, Weitong Ruan, Junyi Li, Rahul Gupta, Huan Sun\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.08995v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-08995/"}
{"id": "arxiv:2602.09007", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a01e0eb848a57a84767836d2c3c982bbf6094d22a3e70652dd42af52760bc5", "submitted_at": "2026-02-09T18:52:02Z", "updated_at": "2026-09-26T18:28:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09007v3"}, "authors": ["Haodong Li", "Jingwei Wu", "Quan Sun", "Guopeng Li", "Juanxi Tian", "Ziyu Guo", "Donghao Zhou", "Tianfei Ren", "Huanyu Zhang", "Yanlin Lai", "Ruichuan An", "Hongbo Peng", "Yuhong Dai", "Jia Wang", "Ziyang Meng", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang", "Pheng-Ann Heng"], "canonical_source_url": "https://arxiv.org/abs/2602.09007", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09007/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43a01e0eb848a57a84767836d2c3c982bbf6094d22a3e70652dd42af52760bc5", "submitted_at": "2026-02-09T18:52:02Z", "updated_at": "2026-09-26T18:28:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.09007v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "identifiers": {"arxiv": "2602.09007"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09007"}], "name": "GUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments", "publication": null, "published_at": "2026-02-09", "record_sha256": "75589b85498188f8542000b5308f64d6e04328c4258097b54a15107a7ca48e81", "relations": [], "resource_id": "arxiv:2602.09007", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09007v3", "topics": ["evaluation", "world-models", "grounding", "screen-understanding"], "updated_at": "2026-09-26"}, "text": "GUI-GenBench: Evaluating Image Generation Models as Interactive GUI Environments\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-GenBench tests image-generation models as action-conditioned GUI environments across five suites, including transitions, trajectories and grounding. Its VLM-based GUI-Score separates goal completion, interaction logic, consistency and visual properties.\nPlatforms: unspecified\nTopics: evaluation, world-models, grounding, screen-understanding\nPublished: 2026-02-09\nAuthors: Haodong Li, Jingwei Wu, Quan Sun, Guopeng Li, Juanxi Tian, Ziyu Guo, Donghao Zhou, Tianfei Ren, Huanyu Zhang, Yanlin Lai, Ruichuan An, Hongbo Peng, Yuhong Dai, Jia Wang, Ziyang Meng, Zheng Ge, Xiangyu Zhang, Daxin Jiang, Pheng-Ann Heng\nRecorded arXiv version: v3\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2602.09007v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09007/"}
{"id": "arxiv:2602.09012", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3178684a322fcf07f847473139efdddf0b02213af468f1afbd285549d15bde56", "submitted_at": "2026-02-09T18:55:33Z", "updated_at": "2026-06-21T08:25:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09012v2"}, "authors": ["Jiacheng Liu", "Yaxin Luo", "Jiacheng Cui", "Xinyi Shang", "Xiaohan Zhao", "Zhiqiang Shen"], "canonical_source_url": "https://arxiv.org/abs/2602.09012", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09012/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3178684a322fcf07f847473139efdddf0b02213af468f1afbd285549d15bde56", "submitted_at": "2026-02-09T18:55:33Z", "updated_at": "2026-06-21T08:25:39Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09012v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.09012"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09012"}], "name": "Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense", "publication": null, "published_at": "2026-02-09", "record_sha256": "02fe393eabbb896c89ea6bded7f115ff49c6dd3132c0964935d2096ba384fff4", "relations": [], "resource_id": "arxiv:2602.09012", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09012v2", "topics": ["evaluation", "safety-security", "data-generation"], "updated_at": "2026-06-21"}, "text": "Next-Gen CAPTCHAs: Leveraging the Cognitive Gap for Scalable and Diverse GUI-Agent Defense\nResource type: paper\nReview status: curated\nScope: core\n\nNext-Gen CAPTCHAs introduces generated interactive challenges intended to distinguish human users from GUI agents. The benchmark targets perception, memory, decisions and actions, with backend-supported challenge types designed for scalable instance generation.\nPlatforms: web\nTopics: evaluation, safety-security, data-generation\nPublished: 2026-02-09\nAuthors: Jiacheng Liu, Yaxin Luo, Jiacheng Cui, Xinyi Shang, Xiaohan Zhao, Zhiqiang Shen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.09012v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09012/"}
{"id": "arxiv:2602.09082", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efc4ad8f13d55939abcc783bdd652f276c3ac0164ea35232a3fb9366642d4558", "submitted_at": "2026-02-09T18:43:40Z", "updated_at": "2026-02-24T15:17:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09082v2"}, "authors": ["Venus Team", "Changlong Gao", "Zhangxuan Gu", "Yulin Liu", "Xinyu Qiu", "Shuheng Shen", "Yue Wen", "Tianyu Xia", "Zhenyu Xu", "Zhengwen Zeng", "Beitong Zhou", "Xingran Zhou", "Weizhi Chen", "Sunhao Dai", "Jingya Dou", "Yichen Gong", "Yuan Guo", "Zhenlin Guo", "Feng Li", "Qian Li", "Jinzhen Lin", "Yuqi Zhou", "Linchao Zhu", "Liang Chen", "Zhenyu Guo", "Changhua Meng", "Weiqiang Wang"], "canonical_source_url": "https://arxiv.org/abs/2602.09082", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09082/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efc4ad8f13d55939abcc783bdd652f276c3ac0164ea35232a3fb9366642d4558", "submitted_at": "2026-02-09T18:43:40Z", "updated_at": "2026-02-24T15:17:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.09082v2"}}, "environments": ["cross-platform", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.09082"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09082"}], "name": "UI-Venus-1.5 Technical Report", "publication": null, "published_at": "2026-02-09", "record_sha256": "e28bb84b9219d213ba57be24273f5d41a9f10fe3bcb09e822126db7b5468d242", "relations": [], "resource_id": "arxiv:2602.09082", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09082v2", "topics": ["reinforcement-learning", "grounding", "mobile-navigation"], "updated_at": "2026-02-24"}, "text": "UI-Venus-1.5 Technical Report\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a GUI-agent family combining broad intermediate training, full-trajectory online reinforcement learning, and merging of specialized models.\nPlatforms: cross-platform, web, android\nTopics: reinforcement-learning, grounding, mobile-navigation\nPublished: 2026-02-09\nAuthors: Venus Team, Changlong Gao, Zhangxuan Gu, Yulin Liu, Xinyu Qiu, Shuheng Shen, Yue Wen, Tianyu Xia, Zhenyu Xu, Zhengwen Zeng, Beitong Zhou, Xingran Zhou, Weizhi Chen, Sunhao Dai, Jingya Dou, Yichen Gong, Yuan Guo, Zhenlin Guo, Feng Li, Qian Li, Jinzhen Lin, Yuqi Zhou, Linchao Zhu, Liang Chen, Zhenyu Guo, Changhua Meng, Weiqiang Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.09082v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09082/"}
{"id": "arxiv:2602.09310", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c3606d090749737d4223a9f826400c2b2aebde245c623391f2a8c6702b2468f8", "submitted_at": "2026-02-10T01:00:55Z", "updated_at": "2026-02-10T01:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09310v1"}, "authors": ["Ananya Gubbi Mohanbabu", "Rosiana Natalie", "Brandon Kim", "Anhong Guo", "Amy Pavel"], "canonical_source_url": "https://arxiv.org/abs/2602.09310", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09310/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c3606d090749737d4223a9f826400c2b2aebde245c623391f2a8c6702b2468f8", "submitted_at": "2026-02-10T01:00:55Z", "updated_at": "2026-02-10T01:00:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09310v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.09310"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09310"}], "name": "A11y-CUA Dataset: Characterizing the Accessibility Gap in Computer Use Agents", "publication": null, "published_at": "2026-02-10", "record_sha256": "5514c4e4107a4428042c75482b467478eb0e793cb2d264f6cde236375d3d092d", "relations": [], "resource_id": "arxiv:2602.09310", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09310v1", "topics": ["human-in-the-loop", "evaluation", "data-generation"], "updated_at": "2026-02-10"}, "text": "A11y-CUA Dataset: Characterizing the Accessibility Gap in Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces A11y-CUA, a dataset comparing how blind or low-vision users and sighted users perform everyday computer tasks. It contrasts those traces with agent behavior under default, keyboard-only, and magnifier conditions to identify barriers to shared control.\nPlatforms: unspecified\nTopics: human-in-the-loop, evaluation, data-generation\nPublished: 2026-02-10\nAuthors: Ananya Gubbi Mohanbabu, Rosiana Natalie, Brandon Kim, Anhong Guo, Amy Pavel\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.09310v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09310/"}
{"id": "arxiv:2602.09662", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b5080050b85e38c8c03078c96c206e92d4ed63d106d55983563d51fcbc9f66c5", "submitted_at": "2026-02-10T11:16:57Z", "updated_at": "2026-02-10T11:16:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09662v1"}, "authors": ["Deyang Jiang", "Jing Huang", "Xuanle Zhao", "Lei Chen", "Liming Zheng", "Fanfan Liu", "Haibo Qiu", "Peng Shi", "Zhixiong Zeng"], "canonical_source_url": "https://arxiv.org/abs/2602.09662", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09662/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b5080050b85e38c8c03078c96c206e92d4ed63d106d55983563d51fcbc9f66c5", "submitted_at": "2026-02-10T11:16:57Z", "updated_at": "2026-02-10T11:16:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09662v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.09662"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09662"}], "name": "TreeCUA: Efficiently Scaling GUI Automation with Tree-Structured Verifiable Evolution", "publication": null, "published_at": "2026-02-10", "record_sha256": "07ae5b3944ce37a93eefdfecec77d1605dd6707cde40bf8cb7d7f8d6ac6be843", "relations": [], "resource_id": "arxiv:2602.09662", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09662v1", "topics": ["data-generation", "planning", "memory", "efficiency"], "updated_at": "2026-02-10"}, "text": "TreeCUA: Efficiently Scaling GUI Automation with Tree-Structured Verifiable Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nTreeCUA organizes GUI exploration trajectories as reusable trees and balances deeper tasks against broader coverage. Its data pipeline combines action verification, memory-backed recovery and a DPO variant that uses information from adjacent branches.\nPlatforms: unspecified\nTopics: data-generation, planning, memory, efficiency\nPublished: 2026-02-10\nAuthors: Deyang Jiang, Jing Huang, Xuanle Zhao, Lei Chen, Liming Zheng, Fanfan Liu, Haibo Qiu, Peng Shi, Zhixiong Zeng\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.09662v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09662/"}
{"id": "arxiv:2602.09856", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55287e2fe8a5e94671e4aab8e9d4a907ba9b6c83306d9b3f61b74adccd5c5059", "submitted_at": "2026-02-10T14:56:19Z", "updated_at": "2026-02-10T14:56:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09856v1"}, "authors": ["Yuhao Zheng", "Li'an Zhong", "Yi Wang", "Rui Dai", "Kaikui Liu", "Xiangxiang Chu", "Linyuan Lv", "Philip Torr", "Kevin Qinghong Lin"], "canonical_source_url": "https://arxiv.org/abs/2602.09856", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-09856/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55287e2fe8a5e94671e4aab8e9d4a907ba9b6c83306d9b3f61b74adccd5c5059", "submitted_at": "2026-02-10T14:56:19Z", "updated_at": "2026-02-10T14:56:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.09856v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2602.09856"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.09856"}], "name": "Code2World: A GUI World Model via Renderable Code Generation", "publication": null, "published_at": "2026-02-10", "record_sha256": "ffabbdf363a8d14c7258da15fb57732563fe9319295160754198769299a8d8ac", "relations": [], "resource_id": "arxiv:2602.09856", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.09856v1", "topics": ["world-models", "data-generation", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-02-10"}, "text": "Code2World: A GUI World Model via Renderable Code Generation\nResource type: paper\nReview status: curated\nScope: core\n\nCode2World predicts future GUI screens through renderable code, trained on AndroidCode trajectories with supervised learning and render-aware reinforcement learning. It evaluates visual prediction and downstream AndroidWorld navigation with the learned world model.\nPlatforms: android\nTopics: world-models, data-generation, reinforcement-learning, mobile-navigation\nPublished: 2026-02-10\nAuthors: Yuhao Zheng, Li'an Zhong, Yi Wang, Rui Dai, Kaikui Liu, Xiangxiang Chu, Linyuan Lv, Philip Torr, Kevin Qinghong Lin\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.09856v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-09856/"}
{"id": "arxiv:2602.10139", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "727279bf83a67f66fa8f173d028936fe0bdae63e9f23800d98476da10201ad0d", "submitted_at": "2026-02-08T15:50:04Z", "updated_at": "2026-04-26T01:34:23Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.10139v3"}, "authors": ["Lepeng Zhao", "Zhenhua Zou", "Shuo Li", "Zhuotao Liu"], "canonical_source_url": "https://arxiv.org/abs/2602.10139", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-10139/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "727279bf83a67f66fa8f173d028936fe0bdae63e9f23800d98476da10201ad0d", "submitted_at": "2026-02-08T15:50:04Z", "updated_at": "2026-04-26T01:34:23Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.10139v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.10139"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10139"}], "name": "Anonymization-Enhanced Privacy Protection for Mobile GUI Agents: Available but Invisible", "publication": null, "published_at": "2026-02-08", "record_sha256": "6349620003b8b26c94c318e6a66b0a44d7f4c4370ca1cee6f3f8f504e27434f0", "relations": [], "resource_id": "arxiv:2602.10139", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.10139v3", "topics": ["safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-04-26"}, "text": "Anonymization-Enhanced Privacy Protection for Mobile GUI Agents: Available but Invisible\nResource type: paper\nReview status: curated\nScope: core\n\nAn anonymization framework replaces sensitive screen content and instructions with consistent, type-preserving placeholders while mediating actions through a local interaction proxy. Experiments on AndroidLab and PrivScreen examine the privacy–utility trade-off when mobile GUI agents operate over these transformed interfaces.\nPlatforms: android\nTopics: safety-security, screen-understanding, mobile-navigation\nPublished: 2026-02-08\nAuthors: Lepeng Zhao, Zhenhua Zou, Shuo Li, Zhuotao Liu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.10139v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-10139/"}
{"id": "arxiv:2602.10356", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3385966e7b483b33738b3c8e6c9a4d9cf17afa671e1f2e49de4f946cbc5f251", "submitted_at": "2026-02-10T23:06:02Z", "updated_at": "2026-05-11T00:11:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.10356v2"}, "authors": ["Tianci Xue", "Zeyi Liao", "Tianneng Shi", "Zilu Wang", "Kai Zhang", "Dawn Song", "Yu Su", "Huan Sun"], "canonical_source_url": "https://arxiv.org/abs/2602.10356", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-10356/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3385966e7b483b33738b3c8e6c9a4d9cf17afa671e1f2e49de4f946cbc5f251", "submitted_at": "2026-02-10T23:06:02Z", "updated_at": "2026-05-11T00:11:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.10356v2"}}, "environments": ["desktop", "linux", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:29:29+00:00", "role": "primary", "sha256": "fba520c76ca9268af9239fbeaf5836e21932ab41ce53ccba71f47e3fc767028a", "supports": ["summary", "environments", "scope"], "url": "https://arxiv.org/html/2602.10356v2"}], "identifiers": {"arxiv": "2602.10356"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10356"}], "name": "Autonomous Continual Learning for Environment Adaptation of Computer-Use Agents", "publication": null, "published_at": "2026-02-10", "record_sha256": "335f896f877231f7880a700bda6179d432aa458899864618f7e4b2c233b98303", "relations": [], "resource_id": "arxiv:2602.10356", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.10356v2", "topics": ["reinforcement-learning", "data-generation", "reward-models", "robustness"], "updated_at": "2026-05-11"}, "text": "Autonomous Continual Learning for Environment Adaptation of Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nACuRL uses autonomous interface exploration and an adaptive task curriculum to continually train computer-use policies with reinforcement learning. CUAJudge supplies outcome feedback, and experiments study adaptation across desktop applications and environment changes.\nPlatforms: desktop, linux, windows\nTopics: reinforcement-learning, data-generation, reward-models, robustness\nPublished: 2026-02-10\nAuthors: Tianci Xue, Zeyi Liao, Tianneng Shi, Zilu Wang, Kai Zhang, Dawn Song, Yu Su, Huan Sun\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract and the specified sections of the same recorded paper version. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.10356v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-10356/"}
{"id": "arxiv:2602.10814", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86a47f084b23cc70e754808f6404bed0bbd500f78fcac4ccc104f95052a49868", "submitted_at": "2026-02-11T12:54:53Z", "updated_at": "2026-02-11T12:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.10814v1"}, "authors": ["Xingyi Zhang", "Yulei Ye", "Kaifeng Huang", "Wenhao Li", "Xiangfeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2602.10814", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-10814/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "86a47f084b23cc70e754808f6404bed0bbd500f78fcac4ccc104f95052a49868", "submitted_at": "2026-02-11T12:54:53Z", "updated_at": "2026-02-11T12:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.10814v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.10814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.10814"}], "name": "See, Plan, Snap: Evaluating Multimodal GUI Agents in Scratch", "publication": null, "published_at": "2026-02-11", "record_sha256": "e7b818ddcf22a2aed3b30247ba0f01acfc473cf62f6f1f3813e08325203911cf", "relations": [], "resource_id": "arxiv:2602.10814", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.10814v1", "topics": ["evaluation", "grounding", "automation"], "updated_at": "2026-02-11"}, "text": "See, Plan, Snap: Evaluating Multimodal GUI Agents in Scratch\nResource type: paper\nReview status: curated\nScope: core\n\nScratchWorld evaluates program construction through the Scratch interface using fine-grained drag-and-drop actions and a complementary semantic-API mode. Runtime tests assess the resulting programs, allowing analysis of the gap between program reasoning and GUI manipulation.\nPlatforms: web\nTopics: evaluation, grounding, automation\nPublished: 2026-02-11\nAuthors: Xingyi Zhang, Yulei Ye, Kaifeng Huang, Wenhao Li, Xiangfeng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.10814v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-10814/"}
{"id": "arxiv:2602.11524", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ffec5a47e31b541cfdf0fb3c143b55cbb28119463731d9a6f3add0fe116caea9", "submitted_at": "2026-02-12T03:31:40Z", "updated_at": "2026-02-12T03:31:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11524v1"}, "authors": ["Congmin Zheng", "Xiaoyun Mo", "Xinbei Ma", "Qiqiang Lin", "Yin Zhao", "Jiachen Zhu", "Xingyu Lou", "Jun Wang", "Zhaoxiang Wang", "Weiwen Liu", "Zhuosheng Zhang", "Yong Yu", "Weinan Zhang"], "canonical_source_url": "https://arxiv.org/abs/2602.11524", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-11524/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ffec5a47e31b541cfdf0fb3c143b55cbb28119463731d9a6f3add0fe116caea9", "submitted_at": "2026-02-12T03:31:40Z", "updated_at": "2026-02-12T03:31:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11524v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.11524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11524"}], "name": "Adaptive Milestone Reward for GUI Agents", "publication": null, "published_at": "2026-02-12", "record_sha256": "6b110fd5892183f88654d1126d5bff6185eb249f4b3a97d2fe3040d8424142ea", "relations": [], "resource_id": "arxiv:2602.11524", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.11524v1", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": "2026-02-12"}, "text": "Adaptive Milestone Reward for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nADMIRE extracts milestones from successful GUI explorations to provide adaptive intermediate rewards. It assigns credit differently to successful and failed trajectories and evaluates training on AndroidWorld, with additional transfer studies in web and embodied settings.\nPlatforms: android, web\nTopics: reinforcement-learning, reward-models, mobile-navigation\nPublished: 2026-02-12\nAuthors: Congmin Zheng, Xiaoyun Mo, Xinbei Ma, Qiqiang Lin, Yin Zhao, Jiachen Zhu, Xingyu Lou, Jun Wang, Zhaoxiang Wang, Weiwen Liu, Zhuosheng Zhang, Yong Yu, Weinan Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.11524v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-11524/"}
{"id": "arxiv:2602.11724", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2076fc8b0b7b37bdf43fe2e0b0cc5fa4c3a8cb7b44a41fdf80f59a8c05c7e1b3", "submitted_at": "2026-02-12T08:51:07Z", "updated_at": "2026-04-23T15:00:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.11724v3"}, "authors": ["Xiwen Teoh", "Yun Lin", "Duc-Minh Nguyen", "Ruofei Ren", "Wenjie Zhang", "Jin Song Dong"], "canonical_source_url": "https://arxiv.org/abs/2602.11724", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-11724/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2076fc8b0b7b37bdf43fe2e0b0cc5fa4c3a8cb7b44a41fdf80f59a8c05c7e1b3", "submitted_at": "2026-02-12T08:51:07Z", "updated_at": "2026-04-23T15:00:58Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.11724v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.11724"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11724"}], "name": "WebTestPilot: Agentic End-to-End Web Testing against Natural Language Specification by Inferring Oracles with Symbolized GUI Elements", "publication": null, "published_at": "2026-02-12", "record_sha256": "f0dde059800d0f85be405a86d624ca985d820121038a28661f26dc4997a40418", "relations": [], "resource_id": "arxiv:2602.11724", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.11724v3", "topics": ["automation", "evaluation", "planning"], "updated_at": "2026-04-23"}, "text": "WebTestPilot: Agentic End-to-End Web Testing against Natural Language Specification by Inferring Oracles with Symbolized GUI Elements\nResource type: paper\nReview status: curated\nScope: core\n\nWebTestPilot turns natural-language requirements into web test steps with inferred preconditions and postconditions over symbolized GUI elements. A benchmark of bug-injected web applications tests whether these context-aware oracles distinguish application defects from agent mistakes.\nPlatforms: web\nTopics: automation, evaluation, planning\nPublished: 2026-02-12\nAuthors: Xiwen Teoh, Yun Lin, Duc-Minh Nguyen, Ruofei Ren, Wenjie Zhang, Jin Song Dong\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.11724v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-11724/"}
{"id": "arxiv:2602.11750", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbba09ca845f4fb772516d26d59fc3d99f8e87ba3657bb29c3c062b50925ec5", "submitted_at": "2026-02-12T09:25:15Z", "updated_at": "2026-02-12T09:25:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11750v1"}, "authors": ["Jiazheng Sun", "Mingxuan Li", "Yingying Zhang", "Jiayang Niu", "Yachen Wu", "Ruihan Jin", "Shuyu Lei", "Pengrongrui Tan", "Zongyu Zhang", "Ruoyi Wang", "Jiachen Yang", "Boyu Yang", "Jiacheng Liu", "Xin Peng"], "canonical_source_url": "https://arxiv.org/abs/2602.11750", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-11750/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2bbba09ca845f4fb772516d26d59fc3d99f8e87ba3657bb29c3c062b50925ec5", "submitted_at": "2026-02-12T09:25:15Z", "updated_at": "2026-02-12T09:25:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.11750v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.11750"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.11750"}], "name": "AmbiBench: Benchmarking Mobile GUI Agents Beyond One-Shot Instructions in the Wild", "publication": null, "published_at": "2026-02-12", "record_sha256": "d92d26f9d14552b0ad8cfee957487928d327a7f0fef51f624e8bf3766a03f8ad", "relations": [], "resource_id": "arxiv:2602.11750", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.11750v1", "topics": ["evaluation", "human-in-the-loop", "mobile-navigation"], "updated_at": "2026-02-12"}, "text": "AmbiBench: Benchmarking Mobile GUI Agents Beyond One-Shot Instructions in the Wild\nResource type: paper\nReview status: curated\nScope: core\n\nAmbiBench evaluates mobile GUI agents under four levels of instruction clarity, including tasks requiring active clarification. Its MUSE evaluator separately audits outcomes, execution quality and interaction quality to study alignment with user intent.\nPlatforms: unspecified\nTopics: evaluation, human-in-the-loop, mobile-navigation\nPublished: 2026-02-12\nAuthors: Jiazheng Sun, Mingxuan Li, Yingying Zhang, Jiayang Niu, Yachen Wu, Ruihan Jin, Shuyu Lei, Pengrongrui Tan, Zongyu Zhang, Ruoyi Wang, Jiachen Yang, Boyu Yang, Jiacheng Liu, Xin Peng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.11750v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-11750/"}
{"id": "arxiv:2602.12276", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c64cd65fc7fec601658baf7836bdbe9c0e494bde2b0a4b66ec9d3090ee106b1f", "submitted_at": "2026-02-12T18:58:30Z", "updated_at": "2026-08-14T21:25:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.12276v2"}, "authors": ["Nicholas Lee", "Lutfi Eren Erdogan", "Chris Joseph John", "Surya Krishnapillai", "Michael W. Mahoney", "Kurt Keutzer", "Amir Gholami"], "canonical_source_url": "https://arxiv.org/abs/2602.12276", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-12276/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c64cd65fc7fec601658baf7836bdbe9c0e494bde2b0a4b66ec9d3090ee106b1f", "submitted_at": "2026-02-12T18:58:30Z", "updated_at": "2026-08-14T21:25:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.12276v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2602.12276"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12276"}], "name": "Agentic Test-Time Scaling for WebAgents", "publication": null, "published_at": "2026-02-12", "record_sha256": "95d102f4e1da27adce453f555731351edde544a317db038bd5f3b07148c50f8e", "relations": [], "resource_id": "arxiv:2602.12276", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.12276v2", "topics": ["efficiency", "planning", "evaluation", "web-navigation"], "updated_at": "2026-08-14"}, "text": "Agentic Test-Time Scaling for WebAgents\nResource type: paper\nReview status: curated\nScope: core\n\nCATTS allocates additional inference computation when an agent’s action votes show uncertainty. The paper compares voting and arbiter strategies and evaluates this selective scaling on WebArena-Lite, Online-Mind2Web and GoBrowse.\nPlatforms: web\nTopics: efficiency, planning, evaluation, web-navigation\nPublished: 2026-02-12\nAuthors: Nicholas Lee, Lutfi Eren Erdogan, Chris Joseph John, Surya Krishnapillai, Michael W. Mahoney, Kurt Keutzer, Amir Gholami\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.12276v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-12276/"}
{"id": "arxiv:2602.12430", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0edc70265a8d46a7c22a9c6384de8aa19f79dbb6987f8c23f4a260647a6a3f0", "submitted_at": "2026-02-12T21:33:25Z", "updated_at": "2026-06-02T16:14:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.12430v4"}, "authors": ["Renjun Xu", "Yang Yan"], "canonical_source_url": "https://arxiv.org/abs/2602.12430", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-12430/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f0edc70265a8d46a7c22a9c6384de8aa19f79dbb6987f8c23f4a260647a6a3f0", "submitted_at": "2026-02-12T21:33:25Z", "updated_at": "2026-06-02T16:14:54Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.12430v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "identifiers": {"arxiv": "2602.12430"}, "kind": "survey", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12430"}], "name": "Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward", "publication": null, "published_at": "2026-02-12", "record_sha256": "d738a59cfefe564cf59fedbd4db58229a40d91150ea870b8ebbdda9e09721866", "relations": [], "resource_id": "arxiv:2602.12430", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2602.12430v4", "topics": ["survey", "tool-integration", "safety-security", "planning"], "updated_at": "2026-06-02"}, "text": "Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward\nResource type: survey\nReview status: curated\nScope: adjacent\n\nThis survey organizes agent skills by architecture, acquisition, deployment and security, including their role in computer-use systems and GUI grounding. It also proposes a permission and lifecycle governance framework for skill deployment.\nPlatforms: unspecified\nTopics: survey, tool-integration, safety-security, planning\nPublished: 2026-02-12\nAuthors: Renjun Xu, Yang Yan\nRecorded arXiv version: v4\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent survey of the skill layer: CUA deployment and GUI grounding are explicit parts of the review, but its overall scope extends well beyond interface agents.\n\nPrimary source: https://arxiv.org/abs/2602.12430v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-12430/"}
{"id": "arxiv:2602.12544", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "555285a1ab51e02661067bdf8bcc3fa7f6ee0fdd63f52f786a03eb2483bd1a31", "submitted_at": "2026-02-13T02:52:18Z", "updated_at": "2026-02-13T02:52:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.12544v1"}, "authors": ["Lajanugen Logeswaran", "Jaekyeom Kim", "Sungryull Sohn", "Creighton Glasscock", "Honglak Lee"], "canonical_source_url": "https://arxiv.org/abs/2602.12544", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-12544/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "555285a1ab51e02661067bdf8bcc3fa7f6ee0fdd63f52f786a03eb2483bd1a31", "submitted_at": "2026-02-13T02:52:18Z", "updated_at": "2026-02-13T02:52:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.12544v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.12544"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.12544"}], "name": "Scaling Web Agent Training through Automatic Data Generation and Fine-grained Evaluation", "publication": null, "published_at": "2026-02-13", "record_sha256": "2f63fb6db303acf3f95a99483ec31e933965d17846b45a2b8fce5ecce82b0a92", "relations": [], "resource_id": "arxiv:2602.12544", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.12544v1", "topics": ["data-generation", "evaluation", "web-navigation"], "updated_at": "2026-02-13"}, "text": "Scaling Web Agent Training through Automatic Data Generation and Fine-grained Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nThis work uses constraint-based progress evaluation to retain partially successful web trajectories for training. It introduces BookingArena for structured booking tasks and studies a distilled agent trained with the resulting data.\nPlatforms: web\nTopics: data-generation, evaluation, web-navigation\nPublished: 2026-02-13\nAuthors: Lajanugen Logeswaran, Jaekyeom Kim, Sungryull Sohn, Creighton Glasscock, Honglak Lee\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.12544v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-12544/"}
{"id": "arxiv:2602.13516", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb048b4e5d1d244d689957f30cdcb93cf4dbb750b24ae527b62ecfb04b554eb4", "submitted_at": "2026-02-13T23:02:50Z", "updated_at": "2026-02-13T23:02:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13516v1"}, "authors": ["Jaechul Roh", "Eugene Bagdasarian", "Hamed Haddadi", "Ali Shahin Shamsabadi"], "canonical_source_url": "https://arxiv.org/abs/2602.13516", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-13516/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb048b4e5d1d244d689957f30cdcb93cf4dbb750b24ae527b62ecfb04b554eb4", "submitted_at": "2026-02-13T23:02:50Z", "updated_at": "2026-02-13T23:02:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2602.13516"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13516"}], "name": "SPILLage: Agentic Oversharing on the Web", "publication": null, "published_at": "2026-02-13", "record_sha256": "999c0651d6b977279e4f0ab109c83a9335d59b59524a3b71825bb77b352cef32", "relations": [], "resource_id": "arxiv:2602.13516", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.13516v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-02-13"}, "text": "SPILLage: Agentic Oversharing on the Web\nResource type: paper\nReview status: curated\nScope: core\n\nFormalizes and measures natural agentic oversharing when browser agents handle private user emails and calendars across live websites.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2026-02-13\nAuthors: Jaechul Roh, Eugene Bagdasarian, Hamed Haddadi, Ali Shahin Shamsabadi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.13516v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-13516/"}
{"id": "arxiv:2602.13559", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04298ed8a4bc0f2811f52eeee6da0a13faa265bb3c5cbc5ea6be573a77c21f27", "submitted_at": "2026-02-14T02:33:55Z", "updated_at": "2026-04-30T02:05:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.13559v2"}, "authors": ["Yuyu Guo", "Wenjie Yang", "Siyuan Yang", "Ziyang Liu", "Cheng Chen", "Yuan Wei", "Yun Hu", "Yang Huang", "Guoliang Hao", "Dongsheng Yuan", "Jianming Wang", "Xin Chen", "Hang Yu", "Lei Lei", "Peng Di"], "canonical_source_url": "https://arxiv.org/abs/2602.13559", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-13559/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "04298ed8a4bc0f2811f52eeee6da0a13faa265bb3c5cbc5ea6be573a77c21f27", "submitted_at": "2026-02-14T02:33:55Z", "updated_at": "2026-04-30T02:05:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.13559v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2602.13559"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13559"}], "name": "OpAgent: Operator Agent for Web Navigation", "publication": null, "published_at": "2026-02-14", "record_sha256": "2a7f3a92007a329a24935d65970613d229f7739dcc985a77f85339901b253b14", "relations": [], "resource_id": "arxiv:2602.13559", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.13559v2", "topics": ["web-navigation", "reinforcement-learning", "planning", "self-reflection"], "updated_at": "2026-04-30"}, "text": "OpAgent: Operator Agent for Web Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nOpAgent combines multi-task GUI fine-tuning with online web reinforcement learning using outcome judgments and rule-based progress rewards. A modular planner, grounder, reflector and summarizer support navigation and recovery, evaluated separately from the underlying trained policy.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, planning, self-reflection\nPublished: 2026-02-14\nAuthors: Yuyu Guo, Wenjie Yang, Siyuan Yang, Ziyang Liu, Cheng Chen, Yuan Wei, Yun Hu, Yang Huang, Guoliang Hao, Dongsheng Yuan, Jianming Wang, Xin Chen, Hang Yu, Lei Lei, Peng Di\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.13559v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-13559/"}
{"id": "arxiv:2602.13653", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "573d859133ec9986af92e1943d0ec913d914acf4e54a030a4b103181ba505ae0", "submitted_at": "2026-02-14T07:44:47Z", "updated_at": "2026-02-14T07:44:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13653v1"}, "authors": ["Yibo Wang", "Guangda Huzhang", "Yuwei Hu", "Yu Xia", "Shiyin Lu", "Qing-Guo Chen", "Zhao Xu", "Weihua Luo", "Kaifu Zhang", "Lijun Zhang"], "canonical_source_url": "https://arxiv.org/abs/2602.13653", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-13653/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "573d859133ec9986af92e1943d0ec913d914acf4e54a030a4b103181ba505ae0", "submitted_at": "2026-02-14T07:44:47Z", "updated_at": "2026-02-14T07:44:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.13653v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2602.13653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.13653"}], "name": "Building Autonomous GUI Navigation via Agentic-Q Estimation and Step-Wise Policy Optimization", "publication": null, "published_at": "2026-02-14", "record_sha256": "25e096136f5e994a6e9617ddef070c29e75fc15b39283e0a25e8bdeb3db95d28", "relations": [], "resource_id": "arxiv:2602.13653", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.13653v1", "topics": ["reinforcement-learning", "reward-models", "grounding"], "updated_at": "2026-02-14"}, "text": "Building Autonomous GUI Navigation via Agentic-Q Estimation and Step-Wise Policy Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nThis framework learns an agentic-Q model to estimate individual actions’ contributions to GUI task completion. It then optimizes a policy with step-level samples from its own trajectories, decoupling policy updates from environment interaction.\nPlatforms: unspecified\nTopics: reinforcement-learning, reward-models, grounding\nPublished: 2026-02-14\nAuthors: Yibo Wang, Guangda Huzhang, Yuwei Hu, Yu Xia, Shiyin Lu, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Lijun Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.13653v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-13653/"}
{"id": "arxiv:2602.14093", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b4c8a5bda1df237fad33ce483018b3476684f904841ed0598ec4286234b858", "submitted_at": "2026-02-15T10:58:01Z", "updated_at": "2026-02-15T10:58:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14093v1"}, "authors": ["Yuan Cao", "Dezhi Ran", "Mengzhou Wu", "Yuzhe Guo", "Xin Chen", "Ang Li", "Gang Cao", "Gong Zhi", "Hao Yu", "Linyi Li", "Wei Yang", "Tao Xie"], "canonical_source_url": "https://arxiv.org/abs/2602.14093", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-14093/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b4c8a5bda1df237fad33ce483018b3476684f904841ed0598ec4286234b858", "submitted_at": "2026-02-15T10:58:01Z", "updated_at": "2026-02-15T10:58:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14093v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.14093"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14093"}], "name": "GUI-GENESIS: Automated Synthesis of Efficient Environments with Verifiable Rewards for GUI Agent Post-Training", "publication": null, "published_at": "2026-02-15", "record_sha256": "1723a3af23fc7cf2b36501a8c8be9246a26ef9d03a9f56202fc4271b43df2c3c", "relations": [], "resource_id": "arxiv:2602.14093", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.14093v1", "topics": ["data-generation", "sandbox", "reinforcement-learning", "evaluation"], "updated_at": "2026-02-15"}, "text": "GUI-GENESIS: Automated Synthesis of Efficient Environments with Verifiable Rewards for GUI Agent Post-Training\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-GENESIS reconstructs applications as lightweight web environments with executable assertions for reward computation. It evaluates the cost of these training environments and transfer from synthetic interaction training to held-out real tasks.\nPlatforms: web\nTopics: data-generation, sandbox, reinforcement-learning, evaluation\nPublished: 2026-02-15\nAuthors: Yuan Cao, Dezhi Ran, Mengzhou Wu, Yuzhe Guo, Xin Chen, Ang Li, Gang Cao, Gong Zhi, Hao Yu, Linyi Li, Wei Yang, Tao Xie\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.14093v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-14093/"}
{"id": "arxiv:2602.14276", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7c6e23f704562c4ae4db9e0b4f8f2a8dde1318efc08876ee57f5967b178257e", "submitted_at": "2026-02-15T19:00:02Z", "updated_at": "2026-05-01T17:32:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.14276v2"}, "authors": ["A. Said Gurbuz", "Sunghwan Hong", "Ahmed Nassar", "Marc Pollefeys", "Peter Staar"], "canonical_source_url": "https://arxiv.org/abs/2602.14276", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-14276/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7c6e23f704562c4ae4db9e0b4f8f2a8dde1318efc08876ee57f5967b178257e", "submitted_at": "2026-02-15T19:00:02Z", "updated_at": "2026-05-01T17:32:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.14276v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.14276"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14276"}], "name": "ScreenParse: Moving Beyond Sparse Grounding with Complete Screen Parsing Supervision", "publication": null, "published_at": "2026-02-15", "record_sha256": "e045d2b3e1eeba0dae3bfd5edb06232869df066818930bda0065cf68b8518688", "relations": [], "resource_id": "arxiv:2602.14276", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.14276v2", "topics": ["screen-understanding", "data-generation", "grounding", "evaluation"], "updated_at": "2026-05-01"}, "text": "ScreenParse: Moving Beyond Sparse Grounding with Complete Screen Parsing Supervision\nResource type: paper\nReview status: curated\nScope: core\n\nScreenParse supplies dense annotations of UI elements, types and text in web screenshots through an automated rendering and filtering pipeline. ScreenVLM learns to output structured screen markup, and the study evaluates dense parsing and transfer to GUI grounding.\nPlatforms: web\nTopics: screen-understanding, data-generation, grounding, evaluation\nPublished: 2026-02-15\nAuthors: A. Said Gurbuz, Sunghwan Hong, Ahmed Nassar, Marc Pollefeys, Peter Staar\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.14276v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-14276/"}
{"id": "arxiv:2602.14296", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3773ec51f5316b9006f93aeafd7b654bf0e7418253c0af16ef7b7c00fddea102", "submitted_at": "2026-02-15T20:03:19Z", "updated_at": "2026-02-15T20:03:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14296v1"}, "authors": ["Yifan Wu", "Yiran Peng", "Yiyu Chen", "Jianhao Ruan", "Zijie Zhuang", "Cheng Yang", "Jiayi Zhang", "Man Chen", "Yenchi Tseng", "Zhaoyang Yu", "Liang Chen", "Yuyao Zhai", "Bang Liu", "Chenglin Wu", "Yuyu Luo"], "canonical_source_url": "https://arxiv.org/abs/2602.14296", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-14296/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3773ec51f5316b9006f93aeafd7b654bf0e7418253c0af16ef7b7c00fddea102", "submitted_at": "2026-02-15T20:03:19Z", "updated_at": "2026-02-15T20:03:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14296v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2602.14296"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14296"}], "name": "AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines", "publication": null, "published_at": "2026-02-15", "record_sha256": "aecba5d8f6d0b787970d114d58055537098955b0b090ffb181624b5935cb3c51", "relations": [], "resource_id": "arxiv:2602.14296", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.14296v1", "topics": ["data-generation", "sandbox", "web-navigation", "evaluation"], "updated_at": "2026-02-15"}, "text": "AutoWebWorld: Synthesizing Infinite Verifiable Web Environments via Finite State Machines\nResource type: paper\nReview status: curated\nScope: core\n\nAutoWebWorld turns finite-state task models into interactive websites whose actions and goal states can be checked programmatically. It generates verified trajectories for training and evaluates transfer to WebVoyager and Online-Mind2Web.\nPlatforms: web\nTopics: data-generation, sandbox, web-navigation, evaluation\nPublished: 2026-02-15\nAuthors: Yifan Wu, Yiran Peng, Yiyu Chen, Jianhao Ruan, Zijie Zhuang, Cheng Yang, Jiayi Zhang, Man Chen, Yenchi Tseng, Zhaoyang Yu, Liang Chen, Yuyao Zhai, Bang Liu, Chenglin Wu, Yuyu Luo\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.14296v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-14296/"}
{"id": "arxiv:2602.14721", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dca0e4219a788205bbb060911d0f004d1db4cf2ff8a4962c04fd5777ea194d7", "submitted_at": "2026-02-16T13:06:49Z", "updated_at": "2026-02-16T13:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14721v1"}, "authors": ["Zikai Xiao", "Jianhong Tu", "Chuhang Zou", "Yuxin Zuo", "Zhi Li", "Peng Wang", "Bowen Yu", "Fei Huang", "Junyang Lin", "Zuozhu Liu"], "canonical_source_url": "https://arxiv.org/abs/2602.14721", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-14721/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dca0e4219a788205bbb060911d0f004d1db4cf2ff8a4962c04fd5777ea194d7", "submitted_at": "2026-02-16T13:06:49Z", "updated_at": "2026-02-16T13:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14721v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.14721"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14721"}], "name": "WebWorld: A Large-Scale World Model for Web Agent Training", "publication": null, "published_at": "2026-02-16", "record_sha256": "bb356e5ed0f5b5897b6fe4732380a0035b02a28a34101e7e6f5ea24b24c99b35", "relations": [], "resource_id": "arxiv:2602.14721", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.14721v1", "topics": ["world-models", "data-generation", "evaluation"], "updated_at": "2026-02-16"}, "text": "WebWorld: A Large-Scale World Model for Web Agent Training\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebWorld, an open-web simulator trained on interaction trajectories for generating agent training data and testing candidate actions. WebWorld-Bench evaluates simulated transitions, alongside downstream web-agent training experiments.\nPlatforms: web\nTopics: world-models, data-generation, evaluation\nPublished: 2026-02-16\nAuthors: Zikai Xiao, Jianhong Tu, Chuhang Zou, Yuxin Zuo, Zhi Li, Peng Wang, Bowen Yu, Fei Huang, Junyang Lin, Zuozhu Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.14721v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-14721/"}
{"id": "arxiv:2602.14865", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "353036db8a5609d64792ea49523ca424cf265d981ff34d1e7bc8fed35ffb21c0", "submitted_at": "2026-02-16T15:59:56Z", "updated_at": "2026-02-16T15:59:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14865v1"}, "authors": ["Chenyang Ma", "Clyde Fare", "Matthew Wilson", "Dave Braines"], "canonical_source_url": "https://arxiv.org/abs/2602.14865", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-14865/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "353036db8a5609d64792ea49523ca424cf265d981ff34d1e7bc8fed35ffb21c0", "submitted_at": "2026-02-16T15:59:56Z", "updated_at": "2026-02-16T15:59:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.14865v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.14865"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.14865"}], "name": "EmbeWebAgent: Embedding Web Agents into Any Customized UI", "publication": null, "published_at": "2026-02-16", "record_sha256": "64db2e5becb155e3c7791c0806622c11d034904b23f2d2c39e18c2007ada1327", "relations": [], "resource_id": "arxiv:2602.14865", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.14865v1", "topics": ["web-navigation", "tool-integration", "automation", "workflow"], "updated_at": "2026-02-16"}, "text": "EmbeWebAgent: Embedding Web Agents into Any Customized UI\nResource type: paper\nReview status: curated\nScope: core\n\nEmbeWebAgent embeds agent access into existing web applications through frontend observations and a page-specific function registry. Its workflow combines GUI primitives, composite actions and MCP tools, demonstrated in a live interface.\nPlatforms: web\nTopics: web-navigation, tool-integration, automation, workflow\nPublished: 2026-02-16\nAuthors: Chenyang Ma, Clyde Fare, Matthew Wilson, Dave Braines\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.14865v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-14865/"}
{"id": "arxiv:2602.16313", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2317a94d54b4458a1b7c5e92754aeaefcbf270aabfa0bd8ea333c673cd4596b3", "submitted_at": "2026-02-18T09:49:14Z", "updated_at": "2026-09-17T22:55:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.16313v2"}, "authors": ["Zexue He", "Yu Wang", "Churan Zhi", "Yuanzhe Hu", "Tzu-Ping Chen", "Lang Yin", "Ze Chen", "Tong Arthur Wu", "Siru Ouyang", "Zihan Wang", "Jiaxin Pei", "Julian McAuley", "Yejin Choi", "Alex Pentland"], "canonical_source_url": "https://arxiv.org/abs/2602.16313", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-16313/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2317a94d54b4458a1b7c5e92754aeaefcbf270aabfa0bd8ea333c673cd4596b3", "submitted_at": "2026-02-18T09:49:14Z", "updated_at": "2026-09-17T22:55:03Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.16313v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:12Z", "role": "primary", "sha256": "c8c95278aa5dcc6a8f2eda3f3b44d61ec1f35adcb6a4c6c9ab328691a1a4f102", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.13462%2C2601.14232%2C2601.15331%2C2601.17548%2C2601.18418%2C2601.18491%2C2601.18827%2C2601.20171%2C2601.20613%2C2601.20742%2C2601.21233%2C2601.21253%2C2601.22037%2C2601.22154%2C2601.22511%2C2601.22607%2C2602.01550%2C2602.03117%2C2602.03361%2C2602.05665%2C2602.05746%2C2602.05877%2C2602.06052%2C2602.07398%2C2602.07458%2C2602.07918%2C2602.09007%2C2602.09463%2C2602.09514%2C2602.10122%2C2602.10525%2C2602.11224%2C2602.12268%2C2602.12430%2C2602.13318%2C2602.13543%2C2602.13692%2C2602.14038%2C2602.14337%2C2602.14364%2C2602.16313%2C2602.16317%2C2602.16666%2C2602.16901%2C2602.16930%2C2602.17547%2C2602.17646%2C2602.18846%2C2602.19320%2C2602.19594&max_results=50"}], "identifiers": {"arxiv": "2602.16313"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.16313"}], "name": "MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks", "publication": null, "published_at": "2026-02-18", "record_sha256": "150346314db54af91f10a963be6d0a90cce84be97489da989357e3d01dc042e4", "relations": [], "resource_id": "arxiv:2602.16313", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.16313v2", "topics": ["memory", "interaction-history", "planning", "evaluation", "long-horizon"], "updated_at": "2026-09-17"}, "text": "MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nMemoryArena links tasks across sessions so that agents must distill earlier interaction experience and use it in later decisions. Its evaluation includes web navigation alongside planning, search and formal reasoning, testing memory through subsequent actions rather than recall alone.\nPlatforms: web\nTopics: memory, interaction-history, planning, evaluation, long-horizon\nPublished: 2026-02-18\nAuthors: Zexue He, Yu Wang, Churan Zhi, Yuanzhe Hu, Tzu-Ping Chen, Lang Yin, Ze Chen, Tong Arthur Wu, Siru Ouyang, Zihan Wang, Jiaxin Pei, Julian McAuley, Yejin Choi, Alex Pentland\nRecorded arXiv version: v2\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2602.16313v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-16313/"}
{"id": "arxiv:2602.16855", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eeaddd8a32bdd4b53356ca43e849ea3697ac3d40a05b32e6c930dc445c243898", "submitted_at": "2026-02-15T01:52:19Z", "updated_at": "2026-02-15T01:52:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16855v1"}, "authors": ["Haiyang Xu", "Xi Zhang", "Haowei Liu", "Junyang Wang", "Zhaozai Zhu", "Shengjie Zhou", "Xuhao Hu", "Feiyu Gao", "Junjie Cao", "Zihua Wang", "Zhiyuan Chen", "Jitong Liao", "Qi Zheng", "Jiahui Zeng", "Ze Xu", "Shuai Bai", "Junyang Lin", "Jingren Zhou", "Ming Yan"], "canonical_source_url": "https://arxiv.org/abs/2602.16855", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-16855/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eeaddd8a32bdd4b53356ca43e849ea3697ac3d40a05b32e6c930dc445c243898", "submitted_at": "2026-02-15T01:52:19Z", "updated_at": "2026-02-15T01:52:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.16855v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.16855"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.16855"}], "name": "Mobile-Agent-v3.5: Multi-platform Fundamental GUI Agents", "publication": null, "published_at": "2026-02-15", "record_sha256": "0ab27a628bcc5e27d6446a37291b2b292f006032531e63c51b4d893a905d807f", "relations": [], "resource_id": "arxiv:2602.16855", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.16855v1", "topics": ["tool-integration", "memory", "reinforcement-learning"], "updated_at": "2026-02-15"}, "text": "Mobile-Agent-v3.5: Multi-platform Fundamental GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents GUI-Owl-1.5 variants and a hybrid data pipeline for multimodal GUI interaction, tool use, memory, and multi-platform tasks.\nPlatforms: cross-platform\nTopics: tool-integration, memory, reinforcement-learning\nPublished: 2026-02-15\nAuthors: Haiyang Xu, Xi Zhang, Haowei Liu, Junyang Wang, Zhaozai Zhu, Shengjie Zhou, Xuhao Hu, Feiyu Gao, Junjie Cao, Zihua Wang, Zhiyuan Chen, Jitong Liao, Qi Zheng, Jiahui Zeng, Ze Xu, Shuai Bai, Junyang Lin, Jingren Zhou, Ming Yan\nRecorded arXiv version: v1\n\nSummary and topic tags reviewed against the primary abstract. Reported experiments, task counts, and artifact availability are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.16855v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-16855/"}
{"id": "arxiv:2602.17003", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28a23bf22920f8c33317e9d16f59d7085f188e79157301ff573abe773cf562ce", "submitted_at": "2026-02-19T01:54:26Z", "updated_at": "2026-05-27T12:15:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.17003v3"}, "authors": ["Serin Kim", "Sangam Lee", "Dongha Lee"], "canonical_source_url": "https://arxiv.org/abs/2602.17003", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-17003/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "28a23bf22920f8c33317e9d16f59d7085f188e79157301ff573abe773cf562ce", "submitted_at": "2026-02-19T01:54:26Z", "updated_at": "2026-05-27T12:15:49Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2602.17003v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.17003"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17003"}], "name": "Persona2Web: Benchmarking Personalized Web Agents for Contextual Reasoning with User History", "publication": null, "published_at": "2026-02-19", "record_sha256": "05cc8e916d3dffdb836ce2b510d94d5867c37decccd683350ff697a39b16612f", "relations": [], "resource_id": "arxiv:2602.17003", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.17003v3", "topics": ["evaluation", "memory", "web-navigation"], "updated_at": "2026-05-27"}, "text": "Persona2Web: Benchmarking Personalized Web Agents for Contextual Reasoning with User History\nResource type: paper\nReview status: curated\nScope: core\n\nPersona2Web tests whether web agents can resolve ambiguous requests using preferences implicit in a user’s history. Its reasoning-aware evaluation compares agent architectures and history-access strategies on personalized tasks on the open web.\nPlatforms: web\nTopics: evaluation, memory, web-navigation\nPublished: 2026-02-19\nAuthors: Serin Kim, Sangam Lee, Dongha Lee\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.17003v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-17003/"}
{"id": "arxiv:2602.17365", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2237dd43f4c876ba69b68c1871ca3efd3aa501e93c86655eb2de6d9f35d2e8", "submitted_at": "2026-02-19T13:48:29Z", "updated_at": "2026-02-19T13:48:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.17365v1"}, "authors": ["Yiming Guan", "Rui Yu", "John Zhang", "Lu Wang", "Chaoyun Zhang", "Liqun Li", "Bo Qiao", "Si Qin", "He Huang", "Fangkai Yang", "Pu Zhao", "Lukas Wutschitz", "Samuel Kessler", "Huseyin A Inan", "Robert Sim", "Saravan Rajmohan", "Qingwei Lin", "Dongmei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2602.17365", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-17365/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3d2237dd43f4c876ba69b68c1871ca3efd3aa501e93c86655eb2de6d9f35d2e8", "submitted_at": "2026-02-19T13:48:29Z", "updated_at": "2026-02-19T13:48:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2602.17365v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.17365"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17365"}], "name": "Computer-Using World Model", "publication": null, "published_at": "2026-02-19", "record_sha256": "5caca9c5ec17739eeee9db3e2a3255d8da735b6da3ecda544c69f4d5f05090fc", "relations": [], "resource_id": "arxiv:2602.17365", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.17365v1", "topics": ["world-models", "planning", "action-prediction"], "updated_at": "2026-02-19"}, "text": "Computer-Using World Model\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a desktop world model that predicts textual UI changes before synthesizing the next screenshot. It learns from Microsoft Office transitions and is evaluated by letting a frozen agent compare simulated actions before execution.\nPlatforms: desktop\nTopics: world-models, planning, action-prediction\nPublished: 2026-02-19\nAuthors: Yiming Guan, Rui Yu, John Zhang, Lu Wang, Chaoyun Zhang, Liqun Li, Bo Qiao, Si Qin, He Huang, Fangkai Yang, Pu Zhao, Lukas Wutschitz, Samuel Kessler, Huseyin A Inan, Robert Sim, Saravan Rajmohan, Qingwei Lin, Dongmei Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.17365v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-17365/"}
{"id": "arxiv:2602.17588", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe83d38ece7cd8f39d932611abaf7a170e7cb9b7c9744bd1176eb42e73be85b5", "submitted_at": "2026-02-19T18:11:28Z", "updated_at": "2026-07-07T23:37:46Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.17588v4"}, "authors": ["Faria Huq", "Zora Zhiruo Wang", "Zhanqiu Guo", "Venu Arvind Arangarajan", "Tianyue Ou", "Frank Xu", "Shuyan Zhou", "Graham Neubig", "Jeffrey P. Bigham"], "canonical_source_url": "https://arxiv.org/abs/2602.17588", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-17588/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fe83d38ece7cd8f39d932611abaf7a170e7cb9b7c9744bd1176eb42e73be85b5", "submitted_at": "2026-02-19T18:11:28Z", "updated_at": "2026-07-07T23:37:46Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2602.17588v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.17588"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.17588"}], "name": "Modeling Distinct Human Interaction in Web Agents", "publication": null, "published_at": "2026-02-19", "record_sha256": "581b67338f289014d1a678d244c98b7683ddcd87f1fc60545fb282a619424b83", "relations": [], "resource_id": "arxiv:2602.17588", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.17588v4", "topics": ["human-in-the-loop", "web-navigation", "data-generation", "action-prediction"], "updated_at": "2026-07-07"}, "text": "Modeling Distinct Human Interaction in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCowCorpus records interleaved human and agent actions during web navigation and identifies distinct patterns of user intervention. Models trained to anticipate these interventions are integrated into live web agents and assessed in a user study.\nPlatforms: web\nTopics: human-in-the-loop, web-navigation, data-generation, action-prediction\nPublished: 2026-02-19\nAuthors: Faria Huq, Zora Zhiruo Wang, Zhanqiu Guo, Venu Arvind Arangarajan, Tianyue Ou, Frank Xu, Shuyan Zhou, Graham Neubig, Jeffrey P. Bigham\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.17588v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-17588/"}
{"id": "arxiv:2602.20502", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e179c833d43a88098d576e31e726eb3e4289f728659697364d0b448282df3fc1", "submitted_at": "2026-02-24T03:03:18Z", "updated_at": "2026-09-28T17:40:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.20502v2"}, "authors": ["Hongbin Zhong", "Fazle Faisal", "Luis França", "Tanakorn Leesatapornwongsa", "Adriana Szekeres", "Kexin Rong", "Suman Nath"], "canonical_source_url": "https://arxiv.org/abs/2602.20502", "citation_url": "https://awesomecua.github.io/resources/arxiv-2602-20502/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e179c833d43a88098d576e31e726eb3e4289f728659697364d0b448282df3fc1", "submitted_at": "2026-02-24T03:03:18Z", "updated_at": "2026-09-28T17:40:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2602.20502v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2602.20502"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2602.20502"}], "name": "ActionEngine: From Reactive to Programmatic Web Agents via State Machine Memory", "publication": null, "published_at": "2026-02-24", "record_sha256": "d18dded34c637ee5bc0945e640baa1e530713fae6772e088ab55d688eb056e95", "relations": [], "resource_id": "arxiv:2602.20502", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2602.20502v2", "topics": ["memory", "planning", "efficiency", "web-navigation"], "updated_at": "2026-09-28"}, "text": "ActionEngine: From Reactive to Programmatic Web Agents via State Machine Memory\nResource type: paper\nReview status: curated\nScope: core\n\nActionEngine explores web applications to build an updateable state-machine memory of GUI states, operations and transitions. It generates complete task programs from that memory, using reactive repair to handle changed interfaces or missing knowledge.\nPlatforms: web\nTopics: memory, planning, efficiency, web-navigation\nPublished: 2026-02-24\nAuthors: Hongbin Zhong, Fazle Faisal, Luis França, Tanakorn Leesatapornwongsa, Adriana Szekeres, Kexin Rong, Suman Nath\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2602.20502v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2602-20502/"}
{"id": "arxiv:2603.00476", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3b24ac636d86124a70b44e84657b963db4c5af3367ae70e5c35e0bebb3f600f", "submitted_at": "2026-02-28T05:25:03Z", "updated_at": "2026-02-28T05:25:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.00476v1"}, "authors": ["Linxi Jiang", "Zhijie Liu", "Haotian Luo", "Zhiqiang Lin"], "canonical_source_url": "https://arxiv.org/abs/2603.00476", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-00476/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d3b24ac636d86124a70b44e84657b963db4c5af3367ae70e5c35e0bebb3f600f", "submitted_at": "2026-02-28T05:25:03Z", "updated_at": "2026-02-28T05:25:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.00476v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2603.00476"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.00476"}], "name": "Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use Agents", "publication": null, "published_at": "2026-02-28", "record_sha256": "d1aa6d120e1c72d361bc98e4321d80c8c5f0c7372cf0e94de4afbda4fdb40a12", "relations": [], "resource_id": "arxiv:2603.00476", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.00476v1", "topics": ["safety-security", "web-navigation", "state-tracking"], "updated_at": "2026-02-28"}, "text": "Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIdentifies time-of-check to time-of-use (TOCTOU) race vulnerabilities when web pages mutate between agent planning and action execution.\nPlatforms: web\nTopics: safety-security, web-navigation, state-tracking\nPublished: 2026-02-28\nAuthors: Linxi Jiang, Zhijie Liu, Haotian Luo, Zhiqiang Lin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.00476v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-00476/"}
{"id": "arxiv:2603.02951", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75f8eb625e969c2b180eeb3e90aa4e513ab5a1dc15657d3849c68b69bc5edf7c", "submitted_at": "2026-03-03T13:02:20Z", "updated_at": "2026-03-07T10:11:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.02951v2"}, "authors": ["Zhenquan Yao", "Zitong Huang", "Yihan Zeng", "Jianhua Han", "Hang Xu", "Chun-Mei Feng", "Jianwei Ma", "Wangmeng Zuo"], "canonical_source_url": "https://arxiv.org/abs/2603.02951", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-02951/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "75f8eb625e969c2b180eeb3e90aa4e513ab5a1dc15657d3849c68b69bc5edf7c", "submitted_at": "2026-03-03T13:02:20Z", "updated_at": "2026-03-07T10:11:08Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.02951v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2603.02951"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.02951"}], "name": "CGL: Advancing Continual GUI Learning via Reinforcement Fine-Tuning", "publication": null, "published_at": "2026-03-03", "record_sha256": "12e57399732eb8b539495f0ad63bd17d23c28e396996736f9ed87773152af686", "relations": [], "resource_id": "arxiv:2603.02951", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.02951v2", "topics": ["reinforcement-learning", "robustness", "evaluation"], "updated_at": "2026-03-07"}, "text": "CGL: Advancing Continual GUI Learning via Reinforcement Fine-Tuning\nResource type: paper\nReview status: curated\nScope: core\n\nCombines supervised and reinforcement fine-tuning for continual GUI learning, using policy entropy to adjust their proportions and gradient projection to limit conflicting updates. AndroidControl-CL groups applications into sequential tasks to evaluate adaptation and retention.\nPlatforms: android\nTopics: reinforcement-learning, robustness, evaluation\nPublished: 2026-03-03\nAuthors: Zhenquan Yao, Zitong Huang, Yihan Zeng, Jianhua Han, Hang Xu, Chun-Mei Feng, Jianwei Ma, Wangmeng Zuo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract describes artifact release as future work; this annotation does not assert current availability.\n\nPrimary source: https://arxiv.org/abs/2603.02951v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-02951/"}
{"id": "arxiv:2603.04364", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "971770e61bf0af91ad1f17b197f09fc166b26bfb4953d3f2c74d457a87e25f3a", "submitted_at": "2026-03-04T18:29:54Z", "updated_at": "2026-10-05T23:01:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.04364v2"}, "authors": ["Haoyu Liu", "Dingcheng Li", "Lukas Rutishauser", "Zeyu Zheng"], "canonical_source_url": "https://arxiv.org/abs/2603.04364", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-04364/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "971770e61bf0af91ad1f17b197f09fc166b26bfb4953d3f2c74d457a87e25f3a", "submitted_at": "2026-03-04T18:29:54Z", "updated_at": "2026-10-05T23:01:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.04364v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.04364"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.04364"}], "name": "Dual-Modality Multi-Stage Adversarial Safety Training: Robustifying Multimodal Web Agents Against Cross-Modal Attacks", "publication": null, "published_at": "2026-03-04", "record_sha256": "2b03e4fb1d467f60bcfc9d8f54811bebd8bbca66af50eee82955bf1d0fffa64a", "relations": [], "resource_id": "arxiv:2603.04364", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.04364v2", "topics": ["safety-security", "reinforcement-learning", "evaluation"], "updated_at": "2026-10-05"}, "text": "Dual-Modality Multi-Stage Adversarial Safety Training: Robustifying Multimodal Web Agents Against Cross-Modal Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nStudies webpage injections that affect both screenshots and accessibility trees, then trains web agents through teacher imitation, supervised adversarial examples, and adversarial reinforcement learning. The evaluation tests defenses on MiniWoB++ and held-out tasks.\nPlatforms: web\nTopics: safety-security, reinforcement-learning, evaluation\nPublished: 2026-03-04\nAuthors: Haoyu Liu, Dingcheng Li, Lukas Rutishauser, Zeyu Zheng\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.04364v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-04364/"}
{"id": "arxiv:2603.04949", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1901125b21431e41eacb5d5f0fc7a439d9a637c43762d441ece46a90e135051b", "submitted_at": "2026-03-05T08:43:06Z", "updated_at": "2026-10-14T10:16:21Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.04949v4"}, "authors": ["Md Farhan Ishmam", "Kenneth Marino"], "canonical_source_url": "https://arxiv.org/abs/2603.04949", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-04949/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1901125b21431e41eacb5d5f0fc7a439d9a637c43762d441ece46a90e135051b", "submitted_at": "2026-03-05T08:43:06Z", "updated_at": "2026-10-14T10:16:21Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.04949v4"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.04949"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.04949"}], "name": "TimeWarp: Evaluating Web Agents by Revisiting the Past", "publication": null, "published_at": "2026-03-05", "record_sha256": "aec9851d399b426d8fcfb83a95366cfd3cd3ff2b973b68ce6a5696d7471d8e8c", "relations": [], "resource_id": "arxiv:2603.04949", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.04949v4", "topics": ["evaluation", "robustness", "web-navigation", "data-generation"], "updated_at": "2026-10-14"}, "text": "TimeWarp: Evaluating Web Agents by Revisiting the Past\nResource type: paper\nReview status: curated\nScope: core\n\nTimeWarp tests web agents across multiple versions of the same sites, varying visual design, frontend code and workflows. TimeTraj transfers plans across versions to collect training trajectories and study robustness to interface change.\nPlatforms: web\nTopics: evaluation, robustness, web-navigation, data-generation\nPublished: 2026-03-05\nAuthors: Md Farhan Ishmam, Kenneth Marino\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.04949v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-04949/"}
{"id": "arxiv:2603.05044", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcca36d9a28ace835ede3fa5a93cb9c54608f164240c18bfe007bc945273b460", "submitted_at": "2026-03-05T10:51:34Z", "updated_at": "2026-04-14T08:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.05044v2"}, "authors": ["Sicheng Fan", "Qingyun Shi", "Shengze Xu", "Shengbo Cai", "Tieyong Zeng", "Li Ling", "Yanyi Shang", "Dehan Kong"], "canonical_source_url": "https://arxiv.org/abs/2603.05044", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-05044/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fcca36d9a28ace835ede3fa5a93cb9c54608f164240c18bfe007bc945273b460", "submitted_at": "2026-03-05T10:51:34Z", "updated_at": "2026-04-14T08:14:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.05044v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.05044"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.05044"}], "name": "WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents", "publication": null, "published_at": "2026-03-05", "record_sha256": "9e5314c27e3929b8518f9085aec5a692e4685d39288def3b479ff42bd76b13d2", "relations": [], "resource_id": "arxiv:2603.05044", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.05044v2", "topics": ["data-generation", "reinforcement-learning", "web-navigation"], "updated_at": "2026-04-14"}, "text": "WebFactory: Automated Compression of Foundational Language Intelligence into Grounded Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebFactory, a GUI-agent training pipeline that synthesizes websites, generates tasks and trajectories, and trains with decomposed rewards. Offline and online transfer evaluations examine agents trained on the generated environments.\nPlatforms: web\nTopics: data-generation, reinforcement-learning, web-navigation\nPublished: 2026-03-05\nAuthors: Sicheng Fan, Qingyun Shi, Shengze Xu, Shengbo Cai, Tieyong Zeng, Li Ling, Yanyi Shang, Dehan Kong\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.05044v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-05044/"}
{"id": "arxiv:2603.05295", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db52246d9f5f305046e6dd14f2ef15440c371678e1db0d05282c84a3667c6f10", "submitted_at": "2026-03-05T15:37:34Z", "updated_at": "2026-04-14T08:41:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.05295v3"}, "authors": ["Sicheng Fan", "Rui Wan", "Yifei Leng", "Gaoning Liang", "Li Ling", "Yanyi Shang", "Dehan Kong"], "canonical_source_url": "https://arxiv.org/abs/2603.05295", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-05295/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db52246d9f5f305046e6dd14f2ef15440c371678e1db0d05282c84a3667c6f10", "submitted_at": "2026-03-05T15:37:34Z", "updated_at": "2026-04-14T08:41:56Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.05295v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:56Z", "role": "primary", "sha256": "c22be2d5c6a84b2c367759ce2b8f41f5a178fccd0534e832802c9d57b40918e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2601.16206%2C2601.17722%2C2601.18113%2C2601.18842%2C2601.20162%2C2601.20335%2C2601.20380%2C2601.20650%2C2601.21123%2C2601.22781%2C2602.00575%2C2602.01576%2C2602.01725%2C2602.02419%2C2602.02548%2C2602.02995%2C2602.03255%2C2602.03792%2C2602.05429%2C2602.05832%2C2602.06075%2C2602.06391%2C2602.07153%2C2602.07624%2C2602.08235%2C2602.08995%2C2602.09082%2C2602.09222%2C2602.09310%2C2602.09662%2C2602.10139%2C2602.11514%2C2602.11724%2C2602.11750%2C2602.14093%2C2602.14276%2C2602.14296%2C2602.14721%2C2602.16855%2C2602.17049%2C2602.17365%2C2602.20502%2C2602.21858%2C2602.22190%2C2603.05044%2C2603.05295%2C2603.07978%2C2603.08533%2C2603.10178%2C2603.12138&max_results=50"}], "identifiers": {"arxiv": "2603.05295"}, "kind": "paper", "last_verified_at": "2026-10-16T06:31:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.05295"}], "name": "WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces", "publication": null, "published_at": "2026-03-05", "record_sha256": "d93936277371d808aa4573ca03f429590ae69ea6fd3479f6584d9bb54fd7b42a", "relations": [], "resource_id": "arxiv:2603.05295", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.05295v3", "topics": ["data-generation", "evaluation", "grounding", "planning"], "updated_at": "2026-04-14"}, "text": "WebChain: A Large-Scale Human-Annotated Dataset of Real-World Web Interaction Traces\nResource type: paper\nReview status: curated\nScope: core\n\nWebChain aligns human web-interaction trajectories with visual, structural and action data for agent training. It also introduces WebChainBench and a mid-training approach that separates spatial grounding from planning.\nPlatforms: web\nTopics: data-generation, evaluation, grounding, planning\nPublished: 2026-03-05\nAuthors: Sicheng Fan, Rui Wan, Yifei Leng, Gaoning Liang, Li Ling, Yanyi Shang, Dehan Kong\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.05295v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-05295/"}
{"id": "arxiv:2603.07432", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68350312fecc36630ecaeec18a203e692e072d31ff80f442662d202b43ead788", "submitted_at": "2026-03-08T03:08:11Z", "updated_at": "2026-03-08T03:08:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07432v1"}, "authors": ["Li Gu", "Zihuan Jiang", "Zhixiang Chi", "Huan Liu", "Ziqiang Wang", "Yuanhao Yu", "Glen Berseth", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2603.07432", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-07432/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68350312fecc36630ecaeec18a203e692e072d31ff80f442662d202b43ead788", "submitted_at": "2026-03-08T03:08:11Z", "updated_at": "2026-03-08T03:08:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07432v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2603.07432"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.07432"}], "name": "Generalization in Online Reinforcement Learning for Mobile Agents", "publication": null, "published_at": "2026-03-08", "record_sha256": "bd4d74d3d3f7d35db3d066628443b34ca1d09c635ba42c7ef5e3c844e84103fe", "relations": [], "resource_id": "arxiv:2603.07432", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.07432v1", "topics": ["reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-03-08"}, "text": "Generalization in Online Reinforcement Learning for Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AndroidWorld-Generalization with separate tests for unseen task instances, templates, and applications. A GRPO training system combines containerized environments with asynchronous trajectory collection to study these generalization regimes.\nPlatforms: android\nTopics: reinforcement-learning, evaluation, robustness\nPublished: 2026-03-08\nAuthors: Li Gu, Zihuan Jiang, Zhixiang Chi, Huan Liu, Ziqiang Wang, Yuanhao Yu, Glen Berseth, Yang Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.07432v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-07432/"}
{"id": "arxiv:2603.07978", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77eb68f31d616b8e542dbb41ada67f28bdc21f29f7da3f04a0077c0fa2d10196", "submitted_at": "2026-03-09T05:27:56Z", "updated_at": "2026-03-09T05:27:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07978v1"}, "authors": ["Jiateng Liu", "Zhenhailong Wang", "Rushi Wang", "Bingxuan Li", "Jeonghwan Kim", "Aditi Tiwari", "Pengfei Yu", "Denghui Zhang", "Heng Ji"], "canonical_source_url": "https://arxiv.org/abs/2603.07978", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-07978/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77eb68f31d616b8e542dbb41ada67f28bdc21f29f7da3f04a0077c0fa2d10196", "submitted_at": "2026-03-09T05:27:56Z", "updated_at": "2026-03-09T05:27:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.07978v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.07978"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.07978"}], "name": "OSExpert: Computer-Use Agents Learning Professional Skills via Exploration", "publication": null, "published_at": "2026-03-09", "record_sha256": "5501eade05d0d88e146eded22de3f4a86cd4aa1e8aec09e8ba577978354584d5", "relations": [], "resource_id": "arxiv:2603.07978", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.07978v1", "topics": ["planning", "data-generation", "evaluation"], "updated_at": "2026-03-09"}, "text": "OSExpert: Computer-Use Agents Learning Professional Skills via Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nUses GUI-based depth-first exploration to verify elementary software functions, assemble action primitives, and construct curricula of composed skills. OSExpert-Eval examines how the resulting procedural knowledge supports complex computer-use tasks.\nPlatforms: unspecified\nTopics: planning, data-generation, evaluation\nPublished: 2026-03-09\nAuthors: Jiateng Liu, Zhenhailong Wang, Rushi Wang, Bingxuan Li, Jeonghwan Kim, Aditi Tiwari, Pengfei Yu, Denghui Zhang, Heng Ji\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.07978v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-07978/"}
{"id": "arxiv:2603.08013", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7334e01b1fbf5cdd02c275c73d2a1b5c2931718d6019fcf6a0ea72b449704d94", "submitted_at": "2026-03-09T06:41:32Z", "updated_at": "2026-03-09T06:41:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.08013v1"}, "authors": ["Yuxiang Chai", "Shunye Tang", "Han Xiao", "Rui Liu", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2603.08013", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-08013/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7334e01b1fbf5cdd02c275c73d2a1b5c2931718d6019fcf6a0ea72b449704d94", "submitted_at": "2026-03-09T06:41:32Z", "updated_at": "2026-03-09T06:41:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.08013v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.08013"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.08013"}], "name": "PIRA-Bench: A Transition from Reactive GUI Agents to GUI-based Proactive Intent Recommendation Agents", "publication": null, "published_at": "2026-03-09", "record_sha256": "08145b0c3189adb79cbd999c52e5b605e602cf7a6ee944d138a763ada6a07d63", "relations": [], "resource_id": "arxiv:2603.08013", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.08013v1", "topics": ["evaluation", "screen-understanding", "memory", "state-tracking", "human-in-the-loop"], "updated_at": "2026-03-09"}, "text": "PIRA-Bench: A Transition from Reactive GUI Agents to GUI-based Proactive Intent Recommendation Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPIRA-Bench evaluates proactive intent recommendations from noisy GUI screenshot sequences containing interleaved tasks and user context. Its PIRF baseline combines memory and state tracking to maintain multiple task threads and identify opportunities to assist.\nPlatforms: unspecified\nTopics: evaluation, screen-understanding, memory, state-tracking, human-in-the-loop\nPublished: 2026-03-09\nAuthors: Yuxiang Chai, Shunye Tang, Han Xiao, Rui Liu, Hongsheng Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.08013v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-08013/"}
{"id": "arxiv:2603.08533", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db31c6ca303f0c7dc43f24dd1eea69621b1e557af33122386945154e8236ab78", "submitted_at": "2026-03-09T16:04:08Z", "updated_at": "2026-03-31T03:08:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.08533v2"}, "authors": ["Yiping Xie", "Song Chen", "Jingxuan Xing", "Wei Jiang", "Zekun Zhu", "Yingyao Wang", "Pi Bu", "Jun Song", "Yuning Jiang", "Bo Zheng"], "canonical_source_url": "https://arxiv.org/abs/2603.08533", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-08533/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db31c6ca303f0c7dc43f24dd1eea69621b1e557af33122386945154e8236ab78", "submitted_at": "2026-03-09T16:04:08Z", "updated_at": "2026-03-31T03:08:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.08533v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.08533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.08533"}], "name": "SecAgent: Efficient Mobile GUI Agent with Semantic Context", "publication": null, "published_at": "2026-03-09", "record_sha256": "d7b7e1908b5fd8dffc9499be7b399687833983234b5b583553b6c584c24c1731", "relations": [], "resource_id": "arxiv:2603.08533", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.08533v2", "topics": ["mobile-navigation", "memory", "reinforcement-learning"], "updated_at": "2026-03-31"}, "text": "SecAgent: Efficient Mobile GUI Agent with Semantic Context\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a Chinese mobile GUI dataset and navigation benchmark together with SecAgent, which compresses previous screenshots and actions into semantic summaries. The agent is trained with supervised and reinforcement fine-tuning to use this compact interaction context.\nPlatforms: unspecified\nTopics: mobile-navigation, memory, reinforcement-learning\nPublished: 2026-03-09\nAuthors: Yiping Xie, Song Chen, Jingxuan Xing, Wei Jiang, Zekun Zhu, Yingyao Wang, Pi Bu, Jun Song, Yuning Jiang, Bo Zheng\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Mobile use is explicit, but the abstract does not identify a mobile operating system.\n\nPrimary source: https://arxiv.org/abs/2603.08533v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-08533/"}
{"id": "arxiv:2603.10178", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22b3001540ac8cc75117cdfb4f381462fe7df3fe7700a86e14b9f522803316ac", "submitted_at": "2026-03-10T19:17:22Z", "updated_at": "2026-03-10T19:17:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10178v1"}, "authors": ["Linxin Song", "Jieyu Zhang", "Huanxin Sheng", "Taiwei Shi", "Gupta Rahul", "Yang Liu", "Ranjay Krishna", "Jian Kang", "Jieyu Zhao"], "canonical_source_url": "https://arxiv.org/abs/2603.10178", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-10178/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "22b3001540ac8cc75117cdfb4f381462fe7df3fe7700a86e14b9f522803316ac", "submitted_at": "2026-03-10T19:17:22Z", "updated_at": "2026-03-10T19:17:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10178v1"}}, "environments": ["desktop", "linux", "macos", "windows", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.10178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10178"}], "name": "Video-Based Reward Modeling for Computer-Use Agents", "publication": null, "published_at": "2026-03-10", "record_sha256": "65398d25dc6044045f4ebf6ac33ea387a73b2db051e8171ed60552d7e5a511a6", "relations": [], "resource_id": "arxiv:2603.10178", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.10178v1", "topics": ["reward-models", "video", "evaluation"], "updated_at": "2026-03-10"}, "text": "Video-Based Reward Modeling for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds ExeVR-53k from execution-video, instruction and reward triplets, including negative instructions with step annotations. A reward model uses spatiotemporal token pruning to judge task completion from the instruction and execution keyframes.\nPlatforms: desktop, linux, macos, windows, android\nTopics: reward-models, video, evaluation\nPublished: 2026-03-10\nAuthors: Linxin Song, Jieyu Zhang, Huanxin Sheng, Taiwei Shi, Gupta Rahul, Yang Liu, Ranjay Krishna, Jian Kang, Jieyu Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.10178v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-10178/"}
{"id": "arxiv:2603.10268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ad18d7fe815f5529fef663130b7ce8841763589e7f4319c0c7556ec2df1da03", "submitted_at": "2026-03-10T22:56:03Z", "updated_at": "2026-03-10T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10268v1"}, "authors": ["Syed Yusuf Ahmed", "Shiwei Feng", "Chanwoo Bae", "Calix Barrus Xiangyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2603.10268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-10268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8ad18d7fe815f5529fef663130b7ce8841763589e7f4319c0c7556ec2df1da03", "submitted_at": "2026-03-10T22:56:03Z", "updated_at": "2026-03-10T22:56:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10268v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:59Z", "role": "primary", "sha256": "d494c76e512656f76a5912eac0768bbbdd50014387c779a0618a40c1722c399b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.07008%2C2602.07283%2C2602.07787%2C2602.09856%2C2602.10356%2C2602.10814%2C2602.10915%2C2602.11524%2C2602.11858%2C2602.12276%2C2602.12544%2C2602.12852%2C2602.13559%2C2602.13653%2C2602.14865%2C2602.15384%2C2602.17003%2C2602.17245%2C2602.17588%2C2602.24142%2C2603.00503%2C2603.00676%2C2603.02951%2C2603.04364%2C2603.04949%2C2603.07024%2C2603.07432%2C2603.08013%2C2603.08316%2C2603.08938%2C2603.10268%2C2603.10291%2C2603.10505%2C2603.10577%2C2603.12710%2C2603.14248%2C2603.14448%2C2603.17441%2C2603.17826%2C2603.20278%2C2603.20340%2C2603.22529%2C2603.23007%2C2603.23559%2C2603.24160%2C2603.24533%2C2603.26041%2C2603.26266%2C2604.00073%2C2604.00842&max_results=50"}], "identifiers": {"arxiv": "2603.10268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10268"}], "name": "SpecOps: A Fully Automated AI Agent Testing Framework in Real-World GUI Environments", "publication": null, "published_at": "2026-03-10", "record_sha256": "153e86cc63078e55e6b35e8e60da7c6d76a078ae945fbb30d8f26d7ebf8fcae3", "relations": [], "resource_id": "arxiv:2603.10268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.10268v1", "topics": ["automation", "evaluation", "planning"], "updated_at": "2026-03-10"}, "text": "SpecOps: A Fully Automated AI Agent Testing Framework in Real-World GUI Environments\nResource type: paper\nReview status: curated\nScope: core\n\nSpecOps divides testing of GUI-based agents into test generation, environment setup, execution and validation handled by specialist agents. Its evaluation spans five real-world agents, including web applications and browser extensions, and measures execution reliability and bug discovery.\nPlatforms: web\nTopics: automation, evaluation, planning\nPublished: 2026-03-10\nAuthors: Syed Yusuf Ahmed, Shiwei Feng, Chanwoo Bae, Calix Barrus Xiangyu Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.10268v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-10268/"}
{"id": "arxiv:2603.10505", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "178b2140d0c70ded39fede20bc91d1459d0a4bf65c5cb6a63c09c9db73828513", "submitted_at": "2026-03-11T07:58:53Z", "updated_at": "2026-03-11T07:58:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10505v1"}, "authors": ["Hyungjoo Chae", "Jungsoo Park", "Alan Ritter"], "canonical_source_url": "https://arxiv.org/abs/2603.10505", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-10505/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "178b2140d0c70ded39fede20bc91d1459d0a4bf65c5cb6a63c09c9db73828513", "submitted_at": "2026-03-11T07:58:53Z", "updated_at": "2026-03-11T07:58:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10505v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.10505"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10505"}], "name": "Safe and Scalable Web Agent Learning via Recreated Websites", "publication": null, "published_at": "2026-03-11", "record_sha256": "01fa65ee13d9467daa1d72019138afe2c55b2d9510069fa154992c6891a1c582", "relations": [], "resource_id": "arxiv:2603.10505", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.10505v1", "topics": ["data-generation", "sandbox", "reinforcement-learning"], "updated_at": "2026-03-11"}, "text": "Safe and Scalable Web Agent Learning via Recreated Websites\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VeriEnv, which recreates websites as executable training environments and exposes internal state through a Python SDK for task generation and deterministic reward checks. Experiments examine web-agent learning across new and expanded site environments.\nPlatforms: web\nTopics: data-generation, sandbox, reinforcement-learning\nPublished: 2026-03-11\nAuthors: Hyungjoo Chae, Jungsoo Park, Alan Ritter\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract promises a release upon acceptance; this annotation does not verify artifact availability.\n\nPrimary source: https://arxiv.org/abs/2603.10505v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-10505/"}
{"id": "arxiv:2603.10577", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "048dc7332372585872f5fd3da47a10bd999c7520b2999e45ea6f578b7c856400", "submitted_at": "2026-03-11T09:28:41Z", "updated_at": "2026-03-12T07:30:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.10577v2"}, "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "canonical_source_url": "https://arxiv.org/abs/2603.10577", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-10577/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "048dc7332372585872f5fd3da47a10bd999c7520b2999e45ea6f578b7c856400", "submitted_at": "2026-03-11T09:28:41Z", "updated_at": "2026-03-12T07:30:09Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2603.10577v2"}}, "environments": ["desktop", "macos", "windows", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.10577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10577"}], "name": "CUAAudit: Meta-Evaluation of Vision-Language Models as Auditors of Autonomous Computer-Use Agents", "publication": null, "published_at": "2026-03-11", "record_sha256": "592814cc486450db4c39f2dffdf5f11202bd23a15b09c50641419a59b1618465", "relations": [], "resource_id": "arxiv:2603.10577", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.10577v2", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-03-12"}, "text": "CUAAudit: Meta-Evaluation of Vision-Language Models as Auditors of Autonomous Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCUAAudit examines vision-language models that judge computer-use task completion from an instruction and final environment state. Across macOS, Windows and Linux benchmarks, it measures accuracy, confidence calibration and disagreement between auditors.\nPlatforms: desktop, macos, windows, linux\nTopics: evaluation, observability, robustness\nPublished: 2026-03-11\nAuthors: Marta Sumyk, Oleksandr Kosovan\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.10577v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-10577/"}
{"id": "arxiv:2603.10664", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5944045ee4562641ee02612969d1c2d038e019fa02cc8d48c671c1d5397557ad", "submitted_at": "2026-03-11T11:24:31Z", "updated_at": "2026-03-11T11:24:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10664v1"}, "authors": ["Alexandre De Masi"], "canonical_source_url": "https://arxiv.org/abs/2603.10664", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-10664/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5944045ee4562641ee02612969d1c2d038e019fa02cc8d48c671c1d5397557ad", "submitted_at": "2026-03-11T11:24:31Z", "updated_at": "2026-03-11T11:24:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.10664v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2603.10664"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.10664"}], "name": "Terminal Is All You Need: Design Properties for Human-AI Agent Collaboration", "publication": null, "published_at": "2026-03-11", "record_sha256": "d1882720486881b23c852bf7bbb9e445851f61d581f817c188b831afd871cea7", "relations": [], "resource_id": "arxiv:2603.10664", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2603.10664v1", "topics": ["human-in-the-loop", "ui-design", "observability", "tool-integration"], "updated_at": "2026-03-11"}, "text": "Terminal Is All You Need: Design Properties for Human-AI Agent Collaboration\nResource type: paper\nReview status: curated\nScope: adjacent\n\nThis position paper derives interface design principles from terminal-based human-agent collaboration: representational compatibility, transparent actions and low participation barriers. It argues that graphical and other agent-facing modalities should provide these properties explicitly.\nPlatforms: unspecified\nTopics: human-in-the-loop, ui-design, observability, tool-integration\nPublished: 2026-03-11\nAuthors: Alexandre De Masi\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent conceptual HCI work about agent-facing interfaces; the abstract does not report GUI task experiments or a new evaluated CUA.\n\nPrimary source: https://arxiv.org/abs/2603.10664v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-10664/"}
{"id": "arxiv:2603.12138", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6421a7928b502cd431964e6f84a1a58d4cd16dbbf1d9b3d1ecc1c2b3dae863bf", "submitted_at": "2026-03-12T16:40:59Z", "updated_at": "2026-03-12T16:40:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12138v1"}, "authors": ["Rui Shao", "Ruize Gao", "Bin Xie", "Yixing Li", "Kaiwen Zhou", "Shuai Wang", "Weili Guan", "Gongwei Chen"], "canonical_source_url": "https://arxiv.org/abs/2603.12138", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-12138/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6421a7928b502cd431964e6f84a1a58d4cd16dbbf1d9b3d1ecc1c2b3dae863bf", "submitted_at": "2026-03-12T16:40:59Z", "updated_at": "2026-03-12T16:40:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12138v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.12138"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.12138"}], "name": "HATS: Hardness-Aware Trajectory Synthesis for GUI Agents", "publication": null, "published_at": "2026-03-12", "record_sha256": "a3fc06fa92566ac0830f76b8fd8c8e24e463c9f553e3ab8bc71149f925cfc154", "relations": [], "resource_id": "arxiv:2603.12138", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.12138v1", "topics": ["data-generation", "robustness", "action-prediction"], "updated_at": "2026-03-12"}, "text": "HATS: Hardness-Aware Trajectory Synthesis for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces HATS, which directs GUI trajectory exploration toward semantically ambiguous actions and iteratively checks instruction-execution alignment. Feedback from trajectory refinement updates the exploration signal used to collect subsequent training examples.\nPlatforms: unspecified\nTopics: data-generation, robustness, action-prediction\nPublished: 2026-03-12\nAuthors: Rui Shao, Ruize Gao, Bin Xie, Yixing Li, Kaiwen Zhou, Shuai Wang, Weili Guan, Gongwei Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.12138v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-12138/"}
{"id": "arxiv:2603.12823", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "274deff6cacfb5f1b7b620a86abe7ef6496b673d929de39376a9ad42dd3c4c02", "submitted_at": "2026-03-13T09:21:25Z", "updated_at": "2026-03-13T09:21:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12823v1"}, "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "canonical_source_url": "https://arxiv.org/abs/2603.12823", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-12823/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "274deff6cacfb5f1b7b620a86abe7ef6496b673d929de39376a9ad42dd3c4c02", "submitted_at": "2026-03-13T09:21:25Z", "updated_at": "2026-03-13T09:21:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.12823v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.12823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.12823"}], "name": "Adaptive Vision-Language Model Routing for Computer Use Agents", "publication": null, "published_at": "2026-03-13", "record_sha256": "518a80de5d1bbb94fb757cb4eb6098434c9ca3a737e1f7aae601fbd475bcca7d", "relations": [], "resource_id": "arxiv:2603.12823", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.12823v1", "topics": ["efficiency", "grounding", "memory", "evaluation"], "updated_at": "2026-03-13"}, "text": "Adaptive Vision-Language Model Routing for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAdaptive VLM Routing estimates GUI-action difficulty and small-model confidence to select a model under a predicted reliability threshold. Grounding and agent-routing benchmarks assess the cost–accuracy trade-off, with prior interaction context used for warm agents.\nPlatforms: unspecified\nTopics: efficiency, grounding, memory, evaluation\nPublished: 2026-03-13\nAuthors: Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.12823v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-12823/"}
{"id": "arxiv:2603.14448", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4507a652e2b3c3857289763442fffd8db21726b97fb7b6304ad61a2139ef5af", "submitted_at": "2026-03-15T15:47:47Z", "updated_at": "2026-03-15T15:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14448v1"}, "authors": ["Ziwei Liu", "Tao Feng", "Borui Kang", "Yanbing Yang", "Jun Luo"], "canonical_source_url": "https://arxiv.org/abs/2603.14448", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-14448/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4507a652e2b3c3857289763442fffd8db21726b97fb7b6304ad61a2139ef5af", "submitted_at": "2026-03-15T15:47:47Z", "updated_at": "2026-03-15T15:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.14448"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14448"}], "name": "Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface Elements", "publication": null, "published_at": "2026-03-15", "record_sha256": "fcb71cac9a9df13effd00d38262d70cc3bbf55c6ca09f822c54d33de5f1e93b2", "relations": [], "resource_id": "arxiv:2603.14448", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.14448v1", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-03-15"}, "text": "Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface Elements\nResource type: paper\nReview status: curated\nScope: core\n\nPresents ZoomUI, a training-free grounding method that expresses instructions as visual element descriptions and uses model attention to zoom progressively into relevant interface regions. Evaluation examines instruction-to-element localization on GUI benchmarks.\nPlatforms: unspecified\nTopics: grounding, training-free, screen-understanding\nPublished: 2026-03-15\nAuthors: Ziwei Liu, Tao Feng, Borui Kang, Yanbing Yang, Jun Luo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.14448v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-14448/"}
{"id": "arxiv:2603.14707", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c7cb0e4784c2a58feeca27a1687d1f73880d9713c16387f0bbff9ed34c427c7", "submitted_at": "2026-03-16T01:31:32Z", "updated_at": "2026-03-16T01:31:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14707v1"}, "authors": ["Xunzhuo Liu", "Bowei He", "Xue Liu", "Andy Luo", "Haichen Zhang", "Huamin Chen"], "canonical_source_url": "https://arxiv.org/abs/2603.14707", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-14707/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c7cb0e4784c2a58feeca27a1687d1f73880d9713c16387f0bbff9ed34c427c7", "submitted_at": "2026-03-16T01:31:32Z", "updated_at": "2026-03-16T01:31:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.14707v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.14707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.14707"}], "name": "Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents", "publication": null, "published_at": "2026-03-16", "record_sha256": "13ea476faf608c64796767de5c4879635d223c54f11288d9cc48051ec7a4122e", "relations": [], "resource_id": "arxiv:2603.14707", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.14707v1", "topics": ["safety-security", "grounding", "robustness"], "updated_at": "2026-03-16"}, "text": "Visual Confused Deputy: Exploiting and Defending Perception Failures in Computer-Using Agents\nResource type: paper\nReview status: curated\nScope: core\n\nFormalizes a visual confused-deputy failure in which a computer agent acts on a misperceived UI object. A separate guardrail checks both the visual target and the agent's action reasoning against deployment-specific knowledge before permitting execution.\nPlatforms: unspecified\nTopics: safety-security, grounding, robustness\nPublished: 2026-03-16\nAuthors: Xunzhuo Liu, Bowei He, Xue Liu, Andy Luo, Haichen Zhang, Huamin Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.14707v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-14707/"}
{"id": "arxiv:2603.15039", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf6c3df80e405968d9ae2353639f3e039d3fcf8351d89db212bda038e3dfac9b", "submitted_at": "2026-03-16T09:45:33Z", "updated_at": "2026-03-16T09:45:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15039v1"}, "authors": ["Yang Li", "Yuchen Liu", "Haoyu Lu", "Zhiqiang Xia", "Hongzhen Wang", "Kaiyang Han", "Changpeng Yang", "Jinyang Wu", "Jiaming Xu", "Runyu Shi", "Ying Huang"], "canonical_source_url": "https://arxiv.org/abs/2603.15039", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-15039/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf6c3df80e405968d9ae2353639f3e039d3fcf8351d89db212bda038e3dfac9b", "submitted_at": "2026-03-16T09:45:33Z", "updated_at": "2026-03-16T09:45:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15039v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.15039"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.15039"}], "name": "GUI-CEval: A Hierarchical and Comprehensive Chinese Benchmark for Mobile GUI Agents", "publication": null, "published_at": "2026-03-16", "record_sha256": "305fae851ee697268b5efdd4d3cfc795c28a77fc2287f68251e6d44620f31f2d", "relations": [], "resource_id": "arxiv:2603.15039", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.15039v1", "topics": ["evaluation", "mobile-navigation", "self-reflection", "screen-understanding"], "updated_at": "2026-03-16"}, "text": "GUI-CEval: A Hierarchical and Comprehensive Chinese Benchmark for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-CEval evaluates Chinese mobile GUI agents through a hierarchy of atomic abilities and application-level tasks collected on physical devices. Its diagnostics cover perception, planning, reflection, execution and post-action evaluation across a broad set of apps.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, self-reflection, screen-understanding\nPublished: 2026-03-16\nAuthors: Yang Li, Yuchen Liu, Haoyu Lu, Zhiqiang Xia, Hongzhen Wang, Kaiyang Han, Changpeng Yang, Jinyang Wu, Jiaming Xu, Runyu Shi, Ying Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.15039v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-15039/"}
{"id": "arxiv:2603.15714", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fe52247e01f331021013dbc3003b3b828e134d7706ec1f587210ad286eee1d", "submitted_at": "2026-03-16T14:49:36Z", "updated_at": "2026-03-16T14:49:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15714v1"}, "authors": ["Mateusz Dziemian", "Maxwell Lin", "Xiaohan Fu", "Micha Nowak", "Nick Winter", "Eliot Jones", "Andy Zou", "Lama Ahmad", "Kamalika Chaudhuri", "Sahana Chennabasappa", "Xander Davies", "Lauren Deason", "Benjamin L. Edelman", "Tanner Emek", "Ivan Evtimov", "Jim Gust", "Maia Hamin", "Kat He", "Klaudia Krawiecka", "Riccardo Patana", "Neil Perry", "Troy Peterson", "Xiangyu Qi", "Javier Rando", "Zifan Wang", "Zihan Wang", "Spencer Whitman", "Eric Winsor", "Arman Zharmagambetov", "Matt Fredrikson", "Zico Kolter"], "canonical_source_url": "https://arxiv.org/abs/2603.15714", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-15714/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fe52247e01f331021013dbc3003b3b828e134d7706ec1f587210ad286eee1d", "submitted_at": "2026-03-16T14:49:36Z", "updated_at": "2026-03-16T14:49:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.15714v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2603.15714"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.15714"}], "name": "How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition", "publication": null, "published_at": "2026-03-16", "record_sha256": "72d5e33a01c4684b6462e36f407d6778b0d1b548c30c82e29c916e86367bc6c2", "relations": [], "resource_id": "arxiv:2603.15714", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.15714v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-03-16"}, "text": "How Vulnerable Are AI Agents to Indirect Prompt Injections? Insights from a Large-Scale Public Competition\nResource type: paper\nReview status: curated\nScope: core\n\nThis public red-teaming study evaluates indirect prompt injection that both changes agent behavior and hides the compromise from the final response. Its scenarios include computer use alongside tool calling and coding, with analyses of transfer and model-dependent vulnerability.\nPlatforms: unspecified\nTopics: safety-security, evaluation, robustness\nPublished: 2026-03-16\nAuthors: Mateusz Dziemian, Maxwell Lin, Xiaohan Fu, Micha Nowak, Nick Winter, Eliot Jones, Andy Zou, Lama Ahmad, Kamalika Chaudhuri, Sahana Chennabasappa, Xander Davies, Lauren Deason, Benjamin L. Edelman, Tanner Emek, Ivan Evtimov, Jim Gust, Maia Hamin, Kat He, Klaudia Krawiecka, Riccardo Patana, Neil Perry, Troy Peterson, Xiangyu Qi, Javier Rando, Zifan Wang, Zihan Wang, Spencer Whitman, Eric Winsor, Arman Zharmagambetov, Matt Fredrikson, Zico Kolter\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2603.15714v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-15714/"}
{"id": "arxiv:2603.17357", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3b429d277ec776af18977a72aaeb95000a4fedf66aa89c0366ab02120d4cdbe", "submitted_at": "2026-03-18T04:41:16Z", "updated_at": "2026-03-18T04:41:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17357v1"}, "authors": ["Nathan Zhao"], "canonical_source_url": "https://arxiv.org/abs/2603.17357", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-17357/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3b429d277ec776af18977a72aaeb95000a4fedf66aa89c0366ab02120d4cdbe", "submitted_at": "2026-03-18T04:41:16Z", "updated_at": "2026-03-18T04:41:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17357v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.17357"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.17357"}], "name": "WebPII: Benchmarking Visual PII Detection for Computer-Use Agents", "publication": null, "published_at": "2026-03-18", "record_sha256": "40c47ccb3464283960ede540d0d96dd676aa59941a51ff0e21187b747c550340", "relations": [], "resource_id": "arxiv:2603.17357", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.17357v1", "topics": ["evaluation", "data-generation", "safety-security", "screen-understanding"], "updated_at": "2026-03-18"}, "text": "WebPII: Benchmarking Visual PII Detection for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebPII provides synthetic e-commerce screenshots with fine-grained annotations for sensitive information, including partially filled forms and transaction identifiers. It also trains WebRedact to detect visual PII as a privacy component for computer-use agents.\nPlatforms: web\nTopics: evaluation, data-generation, safety-security, screen-understanding\nPublished: 2026-03-18\nAuthors: Nathan Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.17357v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-17357/"}
{"id": "arxiv:2603.17441", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2101a32746d6c668bff79461983b2e12138959f49ab3cfcb542ecbd629317f1d", "submitted_at": "2026-03-18T07:26:18Z", "updated_at": "2026-03-18T07:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17441v1"}, "authors": ["Siqi Pei", "Liang Tang", "Tiaonan Duan", "Long Chen", "Shuxian Li", "Kaer Huang", "Yanzhe Jing", "Yiqiang Yan", "Bo Zhang", "Chenghao Jiang", "Borui Zhang", "Jiwen Lu"], "canonical_source_url": "https://arxiv.org/abs/2603.17441", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-17441/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2101a32746d6c668bff79461983b2e12138959f49ab3cfcb542ecbd629317f1d", "submitted_at": "2026-03-18T07:26:18Z", "updated_at": "2026-03-18T07:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.17441v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.17441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.17441"}], "name": "AdaZoom-GUI: Adaptive Zoom-based GUI Grounding with Instruction Refinement", "publication": null, "published_at": "2026-03-18", "record_sha256": "bf83694e2efad7ccb649aca693a8313d9a55153cba2bd41e93dca1e84126100d", "relations": [], "resource_id": "arxiv:2603.17441", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.17441v1", "topics": ["grounding", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-03-18"}, "text": "AdaZoom-GUI: Adaptive Zoom-based GUI Grounding with Instruction Refinement\nResource type: paper\nReview status: curated\nScope: core\n\nCombines instruction rewriting with conditional second-stage zooming for small GUI targets. A grounding dataset and GRPO training support joint prediction of click coordinates and element bounding boxes.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, screen-understanding\nPublished: 2026-03-18\nAuthors: Siqi Pei, Liang Tang, Tiaonan Duan, Long Chen, Shuxian Li, Kaer Huang, Yanzhe Jing, Yiqiang Yan, Bo Zhang, Chenghao Jiang, Borui Zhang, Jiwen Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.17441v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-17441/"}
{"id": "arxiv:2603.18429", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce6da91e59ac57fa24ff991ca023b84448c54d477cacb2d6fded9bf527dd29d", "submitted_at": "2026-03-19T02:45:21Z", "updated_at": "2026-03-19T02:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18429v1"}, "authors": ["Yibo Shi", "Jungang Li", "Linghao Zhang", "Zihao Dongfang", "Biao Wu", "Sicheng Tao", "Yibo Yan", "Chenxi Qin", "Weiting Liu", "Zhixin Lin", "Hanqian Li", "Yu Huang", "Song Dai", "Yonghua Hei", "Yue Ding", "Xiang Li", "Shikang Wang", "Chengdong Xu", "Jingqi Liu", "Xueying Ma", "Zhiwen Zheng", "Xiaofei Zhang", "Bincheng Wang", "Nichen Yang", "Jie Wu", "Lihua Tian", "Chen Li", "Xuming Hu"], "canonical_source_url": "https://arxiv.org/abs/2603.18429", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-18429/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce6da91e59ac57fa24ff991ca023b84448c54d477cacb2d6fded9bf527dd29d", "submitted_at": "2026-03-19T02:45:21Z", "updated_at": "2026-03-19T02:45:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18429v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.18429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.18429"}], "name": "AndroTMem: From Interaction Trajectories to Anchored Memory in Long-Horizon GUI Agents", "publication": null, "published_at": "2026-03-19", "record_sha256": "2d0a96deb786366566f5d18ac49d277f1c3d2da6deafb02ced3c3bad6d33e8af", "relations": [], "resource_id": "arxiv:2603.18429", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.18429v1", "topics": ["memory", "long-horizon", "evaluation"], "updated_at": "2026-03-19"}, "text": "AndroTMem: From Interaction Trajectories to Anchored Memory in Long-Horizon GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AndroTMem-Bench for Android tasks whose later actions depend on earlier UI states. Anchored State Memory retains causally linked intermediate states for subgoal-specific retrieval and evaluates them against full-history and summary-based memory.\nPlatforms: android\nTopics: memory, long-horizon, evaluation\nPublished: 2026-03-19\nAuthors: Yibo Shi, Jungang Li, Linghao Zhang, Zihao Dongfang, Biao Wu, Sicheng Tao, Yibo Yan, Chenxi Qin, Weiting Liu, Zhixin Lin, Hanqian Li, Yu Huang, Song Dai, Yonghua Hei, Yue Ding, Xiang Li, Shikang Wang, Chengdong Xu, Jingqi Liu, Xueying Ma, Zhiwen Zheng, Xiaofei Zhang, Bincheng Wang, Nichen Yang, Jie Wu, Lihua Tian, Chen Li, Xuming Hu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.18429v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-18429/"}
{"id": "arxiv:2603.18762", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0beb3fe342f080f0727181aba55fc1167d2512dd04542986448c5006b6b8213b", "submitted_at": "2026-03-19T11:14:45Z", "updated_at": "2026-03-19T11:14:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18762v1"}, "authors": ["Haochen Zhao", "Shaoyang Cui"], "canonical_source_url": "https://arxiv.org/abs/2603.18762", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-18762/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0beb3fe342f080f0727181aba55fc1167d2512dd04542986448c5006b6b8213b", "submitted_at": "2026-03-19T11:14:45Z", "updated_at": "2026-03-19T11:14:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.18762v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2603.18762"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.18762"}], "name": "ClawTrap: A MITM-Based Red-Teaming Framework for Real-World OpenClaw Security Evaluation", "publication": null, "published_at": "2026-03-19", "record_sha256": "cf402332a3e2c9d20efb9df3597f8924f5a24f9cbb449b3688d98bc2a8d72f88", "relations": [], "resource_id": "arxiv:2603.18762", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.18762v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-03-19"}, "text": "ClawTrap: A MITM-Based Red-Teaming Framework for Real-World OpenClaw Security Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nPresents a man-in-the-middle network red-teaming framework for evaluating autonomous OpenClaw agents under live traffic manipulation.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2026-03-19\nAuthors: Haochen Zhao, Shaoyang Cui\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.18762v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-18762/"}
{"id": "arxiv:2603.19191", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff31371580af6a0c43a6c26f076d5060ee06f974d3a6c1ade67348dc0af4859d", "submitted_at": "2026-03-19T17:47:47Z", "updated_at": "2026-03-19T17:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19191v1"}, "authors": ["Zehao Li", "Zhenyu Wu", "Yibo Zhao", "Bowen Yang", "Jingjing Xie", "Zhaoyang Liu", "Zhoumianze Liu", "Kaiming Jin", "Jianze Liang", "Zonglin Li", "Feng Wu", "Bowen Zhou", "Zun Wang", "Zichen Ding"], "canonical_source_url": "https://arxiv.org/abs/2603.19191", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-19191/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff31371580af6a0c43a6c26f076d5060ee06f974d3a6c1ade67348dc0af4859d", "submitted_at": "2026-03-19T17:47:47Z", "updated_at": "2026-03-19T17:47:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19191v1"}}, "environments": ["android", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2603.19191"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19191"}], "name": "OS-Themis: A Scalable Critic Framework for Generalist GUI Rewards", "publication": null, "published_at": "2026-03-19", "record_sha256": "839261aa6745d9be18d6e0ede4b754af9c9e0dad89ec707ff0b5b5733844b92c", "relations": [], "resource_id": "arxiv:2603.19191", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.19191v1", "topics": ["reward-models", "reinforcement-learning", "evaluation"], "updated_at": "2026-03-19"}, "text": "OS-Themis: A Scalable Critic Framework for Generalist GUI Rewards\nResource type: paper\nReview status: curated\nScope: core\n\nPresents OS-Themis, a multi-agent critic that divides GUI trajectories into verifiable milestones and reviews their evidence before assigning an outcome. OGRBench evaluates GUI rewards across platforms, with AndroidWorld experiments examining training and trajectory filtering.\nPlatforms: android, cross-platform\nTopics: reward-models, reinforcement-learning, evaluation\nPublished: 2026-03-19\nAuthors: Zehao Li, Zhenyu Wu, Yibo Zhao, Bowen Yang, Jingjing Xie, Zhaoyang Liu, Zhoumianze Liu, Kaiming Jin, Jianze Liang, Zonglin Li, Feng Wu, Bowen Zhou, Zun Wang, Zichen Ding\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.19191v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-19191/"}
{"id": "arxiv:2603.19685", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5538933a0e3d11df6ff4d767849979703d83412599fbbb0f8b432c4d9850633b", "submitted_at": "2026-03-20T06:32:16Z", "updated_at": "2026-03-20T06:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19685v1"}, "authors": ["Taiyi Wang", "Sian Gooding", "Florian Hartmann", "Oriana Riva", "Edward Grefenstette"], "canonical_source_url": "https://arxiv.org/abs/2603.19685", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-19685/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5538933a0e3d11df6ff4d767849979703d83412599fbbb0f8b432c4d9850633b", "submitted_at": "2026-03-20T06:32:16Z", "updated_at": "2026-03-20T06:32:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19685v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2603.19685"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19685"}], "name": "A Subgoal-driven Framework for Improving Long-Horizon LLM Agents", "publication": null, "published_at": "2026-03-20", "record_sha256": "165ec6ab9e62467563de4b171211a4d15113f1d5062ea4f43e14ceceaa6a3c37", "relations": [], "resource_id": "arxiv:2603.19685", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.19685v1", "topics": ["planning", "reinforcement-learning", "reward-models", "long-horizon", "evaluation"], "updated_at": "2026-03-20"}, "text": "A Subgoal-driven Framework for Improving Long-Horizon LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThe framework separates inference-time subgoal planning from MiRA, a reinforcement-learning method that supplies milestone rewards. WebArena-Lite experiments evaluate these mechanisms for long-horizon web navigation.\nPlatforms: web\nTopics: planning, reinforcement-learning, reward-models, long-horizon, evaluation\nPublished: 2026-03-20\nAuthors: Taiyi Wang, Sian Gooding, Florian Hartmann, Oriana Riva, Edward Grefenstette\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2603.19685v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-19685/"}
{"id": "arxiv:2603.19993", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfa6cd9aead44150fdc997654ded7253b2890e936b94e44be2fb33642b24d302", "submitted_at": "2026-03-20T14:43:53Z", "updated_at": "2026-03-20T14:43:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19993v1"}, "authors": ["Rozain Shakeel", "Abdul Rahman Mohammad Ali", "Muneeb Mushtaq", "Tausifa Jan Saleem", "Tajamul Ashraf"], "canonical_source_url": "https://arxiv.org/abs/2603.19993", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-19993/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cfa6cd9aead44150fdc997654ded7253b2890e936b94e44be2fb33642b24d302", "submitted_at": "2026-03-20T14:43:53Z", "updated_at": "2026-03-20T14:43:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.19993v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2603.19993"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.19993"}], "name": "MedSPOT: A Workflow-Aware Sequential Grounding Benchmark for Clinical GUI", "publication": null, "published_at": "2026-03-20", "record_sha256": "70701cf823b59f8f08934b76618c0d5494e92db6d1dbf61636283e4ed5c977a2", "relations": [], "resource_id": "arxiv:2603.19993", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.19993v1", "topics": ["grounding", "evaluation"], "updated_at": "2026-03-20"}, "text": "MedSPOT: A Workflow-Aware Sequential Grounding Benchmark for Clinical GUI\nResource type: paper\nReview status: curated\nScope: core\n\nBenchmarks sequential target localization in clinical software workflows, emphasizing how early mistakes affect subsequent interaction steps.\nPlatforms: desktop\nTopics: grounding, evaluation\nPublished: 2026-03-20\nAuthors: Rozain Shakeel, Abdul Rahman Mohammad Ali, Muneeb Mushtaq, Tausifa Jan Saleem, Tajamul Ashraf\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.19993v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-19993/"}
{"id": "arxiv:2603.20340", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e8182e64c7a57a2dd6eb381762bc9a60b08e13ba5a337d0e9125ab6a5518b6c", "submitted_at": "2026-03-20T09:25:50Z", "updated_at": "2026-04-15T07:34:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.20340v3"}, "authors": ["Zijian Lu", "Yiping Zuo", "Yupeng Nie", "Xin He", "Weibei Fan", "Lianyong Qi", "Shi Jin"], "canonical_source_url": "https://arxiv.org/abs/2603.20340", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-20340/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e8182e64c7a57a2dd6eb381762bc9a60b08e13ba5a337d0e9125ab6a5518b6c", "submitted_at": "2026-03-20T09:25:50Z", "updated_at": "2026-04-15T07:34:34Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.20340v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.20340"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.20340"}], "name": "ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents", "publication": null, "published_at": "2026-03-20", "record_sha256": "8731744acfe32c573110c18547cb6921e548a4dccf50383ab91c785680b7a55f", "relations": [], "resource_id": "arxiv:2603.20340", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.20340v3", "topics": ["web-navigation", "planning", "robustness"], "updated_at": "2026-04-15"}, "text": "ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTurns draft web-agent skills into structured executable artifacts that support deterministic checks, fault localization and local repair. VisualWebArena and MiniWoB experiments examine execution and reuse within the same benchmark family.\nPlatforms: web\nTopics: web-navigation, planning, robustness\nPublished: 2026-03-20\nAuthors: Zijian Lu, Yiping Zuo, Yupeng Nie, Xin He, Weibei Fan, Lianyong Qi, Shi Jin\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract limits its portability evidence to transfer within a benchmark family, not generalization across arbitrary benchmarks.\n\nPrimary source: https://arxiv.org/abs/2603.20340v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-20340/"}
{"id": "arxiv:2603.20351", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f25f4999fc9d20433f96e22651311cccf78b5e2ad7cbc5578fe1765ac8be6a00", "submitted_at": "2026-03-20T12:19:41Z", "updated_at": "2026-03-20T12:19:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.20351v1"}, "authors": ["Yizhe Zhao", "Yongjian Fu", "Zihao Feng", "Hao Pan", "Yongheng Deng", "Yaoxue Zhang", "Ju Ren"], "canonical_source_url": "https://arxiv.org/abs/2603.20351", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-20351/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f25f4999fc9d20433f96e22651311cccf78b5e2ad7cbc5578fe1765ac8be6a00", "submitted_at": "2026-03-20T12:19:41Z", "updated_at": "2026-03-20T12:19:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.20351v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:21:52Z", "role": "primary", "sha256": "c194d15b7e422aae48b4e05216a1f1d84471b6666718b9d3dd4c9d250c356e77", "supports": ["name", "kind", "scope", "environments", "topics", "summary", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2210.03347%2C2403.04984%2C2405.07090%2C2406.06282%2C2409.00088%2C2409.10811%2C2411.10943%2C2503.09241%2C2503.20279%2C2503.23350%2C2505.05440%2C2505.10593%2C2505.16422%2C2505.21279%2C2505.24878%2C2506.02153%2C2506.02839%2C2506.05079%2C2506.07976%2C2508.02046%2C2511.07885%2C2511.11248%2C2512.21720%2C2603.20351%2C2604.19299&max_results=25"}], "identifiers": {"arxiv": "2603.20351"}, "kind": "paper", "last_verified_at": "2026-10-18T04:21:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.20351"}], "name": "MANA: Towards Efficient Mobile Ad Detection via Multimodal Agentic UI Navigation", "publication": null, "published_at": "2026-03-20", "record_sha256": "645a30cc4fb20c4a42b4f366c0dae1f2526d4bc36a2089b179b3d586825a62d5", "relations": [], "resource_id": "arxiv:2603.20351", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.20351v1", "topics": ["mobile-navigation", "screen-understanding", "efficiency", "safety-security"], "updated_at": "2026-03-20"}, "text": "MANA: Towards Efficient Mobile Ad Detection via Multimodal Agentic UI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nMANA combines visual, temporal and exploration signals to guide navigation toward mobile advertisements. It evaluates ad detection and interaction cost on commercial smartphones, including obfuscated-ad cases.\nPlatforms: unspecified\nTopics: mobile-navigation, screen-understanding, efficiency, safety-security\nPublished: 2026-03-20\nAuthors: Yizhe Zhao, Yongjian Fu, Zihao Feng, Hao Pan, Yongheng Deng, Yaoxue Zhang, Ju Ren\nRecorded arXiv version: v1\n\nReviewed against the complete official arXiv abstract. Experimental results and linked artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.20351v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-20351/"}
{"id": "arxiv:2603.22529", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e4ed7391969359c2594e8b96b8a3a559ef98d2c84a10c1a2c39b3bd962b58c5", "submitted_at": "2026-03-23T19:47:58Z", "updated_at": "2026-03-23T19:47:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.22529v1"}, "authors": ["Shoubin Yu", "Lei Shu", "Antoine Yang", "Yao Fu", "Srinivas Sunkara", "Maria Wang", "Jindong Chen", "Mohit Bansal", "Boqing Gong"], "canonical_source_url": "https://arxiv.org/abs/2603.22529", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-22529/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e4ed7391969359c2594e8b96b8a3a559ef98d2c84a10c1a2c39b3bd962b58c5", "submitted_at": "2026-03-23T19:47:58Z", "updated_at": "2026-03-23T19:47:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.22529v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.22529"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.22529"}], "name": "Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos", "publication": null, "published_at": "2026-03-23", "record_sha256": "a18409a5d54679e7d59e121004af0fd8c7474c1d730c4ed8236d93a9d992cf92", "relations": [], "resource_id": "arxiv:2603.22529", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.22529v1", "topics": ["evaluation", "video", "web-navigation", "planning"], "updated_at": "2026-03-23"}, "text": "Ego2Web: A Web Agent Benchmark Grounded in Egocentric Videos\nResource type: paper\nReview status: curated\nScope: core\n\nEgo2Web pairs first-person videos with online tasks that require agents to connect the user’s physical context to web actions. It also evaluates an automatic task judge against human judgments and analyzes failures in video understanding and web execution.\nPlatforms: web\nTopics: evaluation, video, web-navigation, planning\nPublished: 2026-03-23\nAuthors: Shoubin Yu, Lei Shu, Antoine Yang, Yao Fu, Srinivas Sunkara, Maria Wang, Jindong Chen, Mohit Bansal, Boqing Gong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.22529v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-22529/"}
{"id": "arxiv:2603.23007", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37960f42162d2bff2fca5e71e350f1ef7c4d86e3f438fd777a88dc2280d3aabe", "submitted_at": "2026-03-24T09:51:43Z", "updated_at": "2026-03-24T09:51:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23007v1"}, "authors": ["Yutao Luo", "Haotian Zhu", "Shuchao Pang", "Zhigang Lu", "Tian Dong", "Yongbin Zhou", "Minhui Xue"], "canonical_source_url": "https://arxiv.org/abs/2603.23007", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-23007/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37960f42162d2bff2fca5e71e350f1ef7c4d86e3f438fd777a88dc2280d3aabe", "submitted_at": "2026-03-24T09:51:43Z", "updated_at": "2026-03-24T09:51:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23007v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.23007"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.23007"}], "name": "AgentRAE: Remote Action Execution through Notification-based Visual Backdoors against Screenshots-based Mobile GUI Agents", "publication": null, "published_at": "2026-03-24", "record_sha256": "e4d878d2d08cf3f85e114aeeb029889fcd5d4c0a496f584eb055c0ec023fed0b", "relations": [], "resource_id": "arxiv:2603.23007", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.23007v1", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-03-24"}, "text": "AgentRAE: Remote Action Execution through Notification-based Visual Backdoors against Screenshots-based Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAgentRAE studies notification-based visual backdoors that redirect mobile GUI-agent actions. Its training separates sensitivity to icon differences from trigger/action association, with evaluations of clean behavior and targeted redirection.\nPlatforms: unspecified\nTopics: safety-security, robustness, evaluation\nPublished: 2026-03-24\nAuthors: Yutao Luo, Haotian Zhu, Shuchao Pang, Zhigang Lu, Tian Dong, Yongbin Zhou, Minhui Xue\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.23007v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-23007/"}
{"id": "arxiv:2603.23791", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb8164fa076c55cddf73508e6b68ce24a6594917e65862504ba9f08b4e97ee48", "submitted_at": "2026-03-24T23:49:15Z", "updated_at": "2026-03-24T23:49:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23791v1"}, "authors": ["Qianlong Lan", "Anuj Kaul"], "canonical_source_url": "https://arxiv.org/abs/2603.23791", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-23791/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb8164fa076c55cddf73508e6b68ce24a6594917e65862504ba9f08b4e97ee48", "submitted_at": "2026-03-24T23:49:15Z", "updated_at": "2026-03-24T23:49:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.23791v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2603.23791"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.23791"}], "name": "The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense", "publication": null, "published_at": "2026-03-24", "record_sha256": "f6bf5acad61371a80a4d8ce53d3bbac31e05d240fa81ce77f1cc3e03c1634178", "relations": [], "resource_id": "arxiv:2603.23791", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.23791v1", "topics": ["safety-security", "web-navigation"], "updated_at": "2026-03-24"}, "text": "The Cognitive Firewall:Securing Browser Based AI Agents Against Indirect Prompt Injection Via Hybrid Edge Cloud Defense\nResource type: paper\nReview status: curated\nScope: core\n\nProposes a split-compute browser defense pairing a local visual sentinel, a cloud planner, and deterministic policy enforcement against indirect prompt injection.\nPlatforms: web\nTopics: safety-security, web-navigation\nPublished: 2026-03-24\nAuthors: Qianlong Lan, Anuj Kaul\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.23791v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-23791/"}
{"id": "arxiv:2603.24157", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fb6a294616e1be46c5da05f60cda7d43c05d8a1ce71116c5e30cb94507cd56f", "submitted_at": "2026-03-25T10:25:48Z", "updated_at": "2026-03-25T10:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24157v1"}, "authors": ["Akash Ghosh", "Tajamul Ashraf", "Rishu Kumar Singh", "Numan Saeed", "Sriparna Saha", "Xiuying Chen", "Salman Khan"], "canonical_source_url": "https://arxiv.org/abs/2603.24157", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-24157/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8fb6a294616e1be46c5da05f60cda7d43c05d8a1ce71116c5e30cb94507cd56f", "submitted_at": "2026-03-25T10:25:48Z", "updated_at": "2026-03-25T10:25:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24157v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2603.24157"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24157"}], "name": "CarePilot: A Multi-Agent Framework for Long-Horizon Computer Task Automation in Healthcare", "publication": null, "published_at": "2026-03-25", "record_sha256": "fed0cfa10de6b3eed1ee9ff67cd3bdd949f82ca03be7268ed48e380e0a7d16c9", "relations": [], "resource_id": "arxiv:2603.24157", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.24157v1", "topics": ["automation", "planning", "memory", "evaluation", "long-horizon", "workflow"], "updated_at": "2026-03-25"}, "text": "CarePilot: A Multi-Agent Framework for Long-Horizon Computer Task Automation in Healthcare\nResource type: paper\nReview status: curated\nScope: core\n\nCareFlow benchmarks long healthcare-software workflows, while CarePilot coordinates action proposals, execution checks and corrective feedback using short- and long-term memory. Its actor consumes visual interfaces and system state to select semantic actions.\nPlatforms: unspecified\nTopics: automation, planning, memory, evaluation, long-horizon, workflow\nPublished: 2026-03-25\nAuthors: Akash Ghosh, Tajamul Ashraf, Rishu Kumar Singh, Numan Saeed, Sriparna Saha, Xiuying Chen, Salman Khan\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2603.24157v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-24157/"}
{"id": "arxiv:2603.24440", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15237fda2f3799314009635017487749dbedc9b1a0ffdd6b56a8d28eaf01481d", "submitted_at": "2026-03-25T15:52:56Z", "updated_at": "2026-03-25T15:52:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24440v1"}, "authors": ["Xiangru Jian", "Shravan Nayak", "Kevin Qinghong Lin", "Aarash Feizi", "Kaixin Li", "Patrice Bechard", "Spandana Gella", "Sai Rajeswar"], "canonical_source_url": "https://arxiv.org/abs/2603.24440", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-24440/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "15237fda2f3799314009635017487749dbedc9b1a0ffdd6b56a8d28eaf01481d", "submitted_at": "2026-03-25T15:52:56Z", "updated_at": "2026-03-25T15:52:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24440v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.24440"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24440"}], "name": "CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents", "publication": null, "published_at": "2026-03-25", "record_sha256": "2c03d2f32b0ba2d7311dd81334c2a6b79e4ad2fab8f78596c872f6eb7ac65fe6", "relations": [], "resource_id": "arxiv:2603.24440", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.24440v1", "topics": ["data-generation", "video", "evaluation", "grounding", "workflow"], "updated_at": "2026-03-25"}, "text": "CUA-Suite: Massive Human-annotated Video Demonstrations for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCUA-Suite combines continuous expert desktop demonstrations in VideoCUA with the UI-Vision evaluation benchmark and GroundCUA grounding data. Screen recordings, cursor traces and layered annotations support study of temporal behavior and professional software workflows.\nPlatforms: desktop\nTopics: data-generation, video, evaluation, grounding, workflow\nPublished: 2026-03-25\nAuthors: Xiangru Jian, Shravan Nayak, Kevin Qinghong Lin, Aarash Feizi, Kaixin Li, Patrice Bechard, Spandana Gella, Sai Rajeswar\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.24440v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-24440/"}
{"id": "arxiv:2603.24533", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa03c8716edf782ce1a86e404107aa2569be63484dfedb472d3710dc82f405d0", "submitted_at": "2026-03-25T17:10:29Z", "updated_at": "2026-03-25T17:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24533v1"}, "authors": ["Zichuan Lin", "Feiyu Liu", "Yijun Yang", "Jiafei Lyu", "Yiming Gao", "Yicheng Liu", "Zhicong Lu", "Yangbin Yu", "Mingyu Yang", "Junyou Li", "Deheng Ye", "Jie Jiang"], "canonical_source_url": "https://arxiv.org/abs/2603.24533", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-24533/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa03c8716edf782ce1a86e404107aa2569be63484dfedb472d3710dc82f405d0", "submitted_at": "2026-03-25T17:10:29Z", "updated_at": "2026-03-25T17:10:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.24533v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.24533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.24533"}], "name": "UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience", "publication": null, "published_at": "2026-03-25", "record_sha256": "cc180add6df9f0792a3a2af6e588fb80706b5c70f6c5f4c9f50b2cd1bea141c5", "relations": [], "resource_id": "arxiv:2603.24533", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.24533v1", "topics": ["imitation-learning", "data-generation", "mobile-navigation"], "updated_at": "2026-03-25"}, "text": "UI-Voyager: A Self-Evolving GUI Agent Learning via Failed Experience\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a two-stage mobile GUI learner that first co-evolves trajectories and policies through rejection fine-tuning. Group Relative Self-Distillation then uses successful rollout branches to provide step-level supervision for failed trajectories, evaluated on AndroidWorld.\nPlatforms: android\nTopics: imitation-learning, data-generation, mobile-navigation\nPublished: 2026-03-25\nAuthors: Zichuan Lin, Feiyu Liu, Yijun Yang, Jiafei Lyu, Yiming Gao, Yicheng Liu, Zhicong Lu, Yangbin Yu, Mingyu Yang, Junyou Li, Deheng Ye, Jie Jiang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.24533v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-24533/"}
{"id": "arxiv:2603.25226", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8414eb53c66f745fb074d9f2699f76ffea81cb55f98aacf9ce896401937f80c5", "submitted_at": "2026-03-26T09:27:29Z", "updated_at": "2026-03-26T09:27:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25226v1"}, "authors": ["Fanheng Kong", "Jingyuan Zhang", "Yang Yue", "Chenxi Sun", "Yang Tian", "Shi Feng", "Xiaocui Yang", "Daling Wang", "Yu Tian", "Jun Du", "Wenchong Zeng", "Han Li", "Kun Gai"], "canonical_source_url": "https://arxiv.org/abs/2603.25226", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-25226/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8414eb53c66f745fb074d9f2699f76ffea81cb55f98aacf9ce896401937f80c5", "submitted_at": "2026-03-26T09:27:29Z", "updated_at": "2026-03-26T09:27:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25226v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.25226"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.25226"}], "name": "WebTestBench: Evaluating Computer-Use Agents towards End-to-End Automated Web Testing", "publication": null, "published_at": "2026-03-26", "record_sha256": "b690b5728733df531154d0c17dc175bd90ceb8dfbbd4984deacbdee360f4b0a5", "relations": [], "resource_id": "arxiv:2603.25226", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.25226v1", "topics": ["evaluation", "web-navigation", "automation"], "updated_at": "2026-03-26"}, "text": "WebTestBench: Evaluating Computer-Use Agents towards End-to-End Automated Web Testing\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebTestBench for end-to-end website testing, separating checklist generation from defect detection. Its WebTester baseline evaluates test coverage, detection and sustained interaction across web application categories.\nPlatforms: web\nTopics: evaluation, web-navigation, automation\nPublished: 2026-03-26\nAuthors: Fanheng Kong, Jingyuan Zhang, Yang Yue, Chenxi Sun, Yang Tian, Shi Feng, Xiaocui Yang, Daling Wang, Yu Tian, Jun Du, Wenchong Zeng, Han Li, Kun Gai\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.25226v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-25226/"}
{"id": "arxiv:2603.25864", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b3dd3efca196fa3dd168431102dbb5f0f763e7d9e54b3d709ad2362a9c6f154", "submitted_at": "2026-03-26T19:37:53Z", "updated_at": "2026-03-26T19:37:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25864v1"}, "authors": ["Saelyne Yang", "Jaesang Yu", "Yi-Hao Peng", "Kevin Qinghong Lin", "Jae Won Cho", "Yale Song", "Juho Kim"], "canonical_source_url": "https://arxiv.org/abs/2603.25864", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-25864/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b3dd3efca196fa3dd168431102dbb5f0f763e7d9e54b3d709ad2362a9c6f154", "submitted_at": "2026-03-26T19:37:53Z", "updated_at": "2026-03-26T19:37:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.25864v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2603.25864"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.25864"}], "name": "GUIDE: A Benchmark for Understanding and Assisting Users in Open-Ended GUI Tasks", "publication": null, "published_at": "2026-03-26", "record_sha256": "91876ed0ce7cd2c43d2142ab108fc53afe6fdee003446e0c44c4cb9cccd02450", "relations": [], "resource_id": "arxiv:2603.25864", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.25864v1", "topics": ["evaluation", "human-in-the-loop", "video", "screen-understanding"], "updated_at": "2026-03-26"}, "text": "GUIDE: A Benchmark for Understanding and Assisting Users in Open-Ended GUI Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nGUIDE evaluates whether models can recognize a user’s behavior, infer intent and decide when to offer help during open-ended software tasks. It uses narrated screen recordings of novice users and studies the effect of additional user context on these assistance decisions.\nPlatforms: unspecified\nTopics: evaluation, human-in-the-loop, video, screen-understanding\nPublished: 2026-03-26\nAuthors: Saelyne Yang, Jaesang Yu, Yi-Hao Peng, Kevin Qinghong Lin, Jae Won Cho, Yale Song, Juho Kim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.25864v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-25864/"}
{"id": "arxiv:2603.26041", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "738d5048bb7a137d848b6f6068a7ce92694469f1e7bce0702dce422c230a3d0b", "submitted_at": "2026-03-27T03:21:19Z", "updated_at": "2026-08-09T16:37:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.26041v4"}, "authors": ["Daiqiang Li", "Zihao Pan", "Zeyu Zhang", "Xuyang Liu", "Shijia Xu", "Ronghao Chen", "Huacan Wang", "Honggang Chen", "Linfeng Zhang", "Zhangquan Chen", "Haiyun Jiang"], "canonical_source_url": "https://arxiv.org/abs/2603.26041", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-26041/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "738d5048bb7a137d848b6f6068a7ce92694469f1e7bce0702dce422c230a3d0b", "submitted_at": "2026-03-27T03:21:19Z", "updated_at": "2026-08-09T16:37:37Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2603.26041v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.26041"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26041"}], "name": "Where and How to Prune: An Empirical Study of Visual Token Pruning for GUI Agent Navigation", "publication": null, "published_at": "2026-03-27", "record_sha256": "fccd3f52f214bc20f3596a1ffc715b40bb2db8b5cc69c502daf3b14c45b1b924", "relations": [], "resource_id": "arxiv:2603.26041", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.26041v4", "topics": ["efficiency", "interaction-history", "screen-understanding"], "updated_at": "2026-08-09"}, "text": "Where and How to Prune: An Empirical Study of Visual Token Pruning for GUI Agent Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nStudies visual token pruning for GUI navigation, separating reusable screenshot encoder embeddings from later language-model computation. The analysis examines foreground-background token allocation and spatial coverage when pruning historical screenshots.\nPlatforms: unspecified\nTopics: efficiency, interaction-history, screen-understanding\nPublished: 2026-03-27\nAuthors: Daiqiang Li, Zihao Pan, Zeyu Zhang, Xuyang Liu, Shijia Xu, Ronghao Chen, Huacan Wang, Honggang Chen, Linfeng Zhang, Zhangquan Chen, Haiyun Jiang\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.26041v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-26041/"}
{"id": "arxiv:2603.26211", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0838afd6777af3e3995fa3bf9645102f98962cccd81c1db438a007b4cd33c442", "submitted_at": "2026-03-27T09:34:07Z", "updated_at": "2026-03-27T09:34:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.26211v1"}, "authors": ["Shrinidhi Kumbhar", "Haofu Liao", "Srikar Appalaraju", "Kunwar Yashraj Singh"], "canonical_source_url": "https://arxiv.org/abs/2603.26211", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-26211/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0838afd6777af3e3995fa3bf9645102f98962cccd81c1db438a007b4cd33c442", "submitted_at": "2026-03-27T09:34:07Z", "updated_at": "2026-03-27T09:34:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.26211v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.26211"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26211"}], "name": "Towards GUI Agents: Vision-Language Diffusion Models for GUI Grounding", "publication": null, "published_at": "2026-03-27", "record_sha256": "e2965e4f1fa2f61613cfd22413d2553b65733b243c214f8cebf45b3a20827cb3", "relations": [], "resource_id": "arxiv:2603.26211", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.26211v1", "topics": ["grounding", "action-prediction", "efficiency"], "updated_at": "2026-03-27"}, "text": "Towards GUI Agents: Vision-Language Diffusion Models for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nAdapts a discrete vision-language diffusion model to single-turn GUI action and bounding-box prediction, using a hybrid masking schedule for coordinate geometry. Experiments across web, desktop and mobile datasets examine grounding accuracy and inference tradeoffs.\nPlatforms: web, desktop\nTopics: grounding, action-prediction, efficiency\nPublished: 2026-03-27\nAuthors: Shrinidhi Kumbhar, Haofu Liao, Srikar Appalaraju, Kunwar Yashraj Singh\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.26211v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-26211/"}
{"id": "arxiv:2603.26266", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b388086aeb8ac47e386099e5cd68cb89bb03978ee7d5897950181035d0abb723", "submitted_at": "2026-03-27T10:33:08Z", "updated_at": "2026-06-30T04:47:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.26266v3"}, "authors": ["Rui Xie", "Zhi Gao", "Chenrui Shi", "Zirui Shang", "Lu Chen", "Qing Li"], "canonical_source_url": "https://arxiv.org/abs/2603.26266", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-26266/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b388086aeb8ac47e386099e5cd68cb89bb03978ee7d5897950181035d0abb723", "submitted_at": "2026-03-27T10:33:08Z", "updated_at": "2026-06-30T04:47:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2603.26266v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.26266"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.26266"}], "name": "GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play Annotation", "publication": null, "published_at": "2026-03-27", "record_sha256": "e293f64f788b1bd59dfb112a8455693d8ad52fbcebf0a4b3e28d93e53f9a4331", "relations": [], "resource_id": "arxiv:2603.26266", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.26266v3", "topics": ["video", "planning", "grounding", "training-free"], "updated_at": "2026-06-30"}, "text": "GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play Annotation\nResource type: paper\nReview status: curated\nScope: core\n\nGUIDE retrieves relevant software tutorials through subtitle-based video search, then infers planning and grounding knowledge from consecutive annotated keyframes. It inserts that knowledge into existing GUI-agent modules without parameter updates and evaluates the plug-in on OSWorld.\nPlatforms: unspecified\nTopics: video, planning, grounding, training-free\nPublished: 2026-03-27\nAuthors: Rui Xie, Zhi Gao, Chenrui Shi, Zirui Shang, Lu Chen, Qing Li\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.26266v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-26266/"}
{"id": "arxiv:2603.28069", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe857c9f7e7819a9c2f99448f77bbeae15ae4e2ac8ad308000f3d217b0177fc", "submitted_at": "2026-03-30T06:15:06Z", "updated_at": "2026-03-30T06:15:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.28069v1"}, "authors": ["Christopher Clark", "Yue Yang", "Jae Sung Park", "Zixian Ma", "Jieyu Zhang", "Rohun Tripathi", "Mohammadreza Salehi", "Sangho Lee", "Taira Anderson", "Winson Han", "Ranjay Krishna"], "canonical_source_url": "https://arxiv.org/abs/2603.28069", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-28069/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fe857c9f7e7819a9c2f99448f77bbeae15ae4e2ac8ad308000f3d217b0177fc", "submitted_at": "2026-03-30T06:15:06Z", "updated_at": "2026-03-30T06:15:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.28069v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2603.28069"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.28069"}], "name": "MolmoPoint: Better Pointing for VLMs with Grounding Tokens", "publication": null, "published_at": "2026-03-30", "record_sha256": "aaf85228cb9897bc43be2859b808aa634de7d72e5f285c9a43a694da9c03fd8e", "relations": [], "resource_id": "arxiv:2603.28069", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2603.28069v1", "topics": ["grounding", "screen-understanding", "video", "efficiency"], "updated_at": "2026-03-30"}, "text": "MolmoPoint: Better Pointing for VLMs with Grounding Tokens\nResource type: paper\nReview status: curated\nScope: adjacent\n\nMolmoPoint replaces textual coordinate generation with pointing tokens that select visual regions and refine them through subpatches. The general image/video method includes GUI-pointing evaluation on ScreenSpotPro alongside other visual grounding tasks.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, video, efficiency\nPublished: 2026-03-30\nAuthors: Christopher Clark, Yue Yang, Jae Sung Park, Zixian Ma, Jieyu Zhang, Rohun Tripathi, Mohammadreza Salehi, Sangho Lee, Taira Anderson, Winson Han, Ranjay Krishna\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent general VLM pointing method with an explicit GUI-grounding experiment; no autonomous GUI-task execution is established.\n\nPrimary source: https://arxiv.org/abs/2603.28069v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-28069/"}
{"id": "arxiv:2603.29020", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4f03e65387e3aad99b8c4cef166a7a8b4a9e2954ad932b8daf566e30f156595", "submitted_at": "2026-03-30T21:27:28Z", "updated_at": "2026-03-30T21:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29020v1"}, "authors": ["Deepak Akkil", "Mowafak Allaham", "Amal Raj", "Tamer Abuelsaad", "Ravi Kokku"], "canonical_source_url": "https://arxiv.org/abs/2603.29020", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-29020/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4f03e65387e3aad99b8c4cef166a7a8b4a9e2954ad932b8daf566e30f156595", "submitted_at": "2026-03-30T21:27:28Z", "updated_at": "2026-03-30T21:27:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29020v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.29020"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.29020"}], "name": "Emergence WebVoyager: Toward Consistent and Transparent Evaluation of (Web) Agents in The Wild", "publication": null, "published_at": "2026-03-30", "record_sha256": "5ae26fcd283c147d8a0195959520a43dd4e76146a94dd6cebb294652d0c7b67f", "relations": [], "resource_id": "arxiv:2603.29020", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.29020v1", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-03-30"}, "text": "Emergence WebVoyager: Toward Consistent and Transparent Evaluation of (Web) Agents in The Wild\nResource type: paper\nReview status: curated\nScope: core\n\nAudits sources of ambiguity and operational variation in WebVoyager evaluation and introduces standardized guidance for task setup, failure handling, annotation and reporting. Emergence WebVoyager applies this protocol to a web-agent evaluation.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2026-03-30\nAuthors: Deepak Akkil, Mowafak Allaham, Amal Raj, Tamer Abuelsaad, Ravi Kokku\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.29020v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-29020/"}
{"id": "arxiv:2603.29318", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c71b713efd6ea47e7c78f5dd51d06c5cfdba750529bfbcdfc8e1c69897cbcc9", "submitted_at": "2026-03-31T06:37:12Z", "updated_at": "2026-03-31T06:37:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29318v1"}, "authors": ["Hongyi Nie", "Xunyuan Liu", "Yudong Bai", "Yaqing Wang", "Yang Liu", "Quanming Yao", "Zhen Wang"], "canonical_source_url": "https://arxiv.org/abs/2603.29318", "citation_url": "https://awesomecua.github.io/resources/arxiv-2603-29318/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1c71b713efd6ea47e7c78f5dd51d06c5cfdba750529bfbcdfc8e1c69897cbcc9", "submitted_at": "2026-03-31T06:37:12Z", "updated_at": "2026-03-31T06:37:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2603.29318v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2603.29318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2603.29318"}], "name": "PSPA-Bench: A Personalized Benchmark for Smartphone GUI Agent", "publication": null, "published_at": "2026-03-31", "record_sha256": "fb930e8a3767c5e004563063e24690677036b0937bb2579c567d99f65dcae52e", "relations": [], "resource_id": "arxiv:2603.29318", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2603.29318v1", "topics": ["evaluation", "mobile-navigation"], "updated_at": "2026-03-31"}, "text": "PSPA-Bench: A Personalized Benchmark for Smartphone GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nPSPA-Bench evaluates personalized smartphone GUI behavior using user-specific instructions and a process-aware assessment of interaction traces. Its comparison of GUI agents diagnoses weaknesses in adapting workflows and preferences rather than measuring generic task completion alone.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation\nPublished: 2026-03-31\nAuthors: Hongyi Nie, Xunyuan Liu, Yudong Bai, Yaqing Wang, Yang Liu, Quanming Yao, Zhen Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2603.29318v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2603-29318/"}
{"id": "arxiv:2604.00886", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce7f70ff1f8c7d4058d8fdf00cb6cf1d61c0367a50ba76fc96d356bb07452c60", "submitted_at": "2026-04-01T13:33:27Z", "updated_at": "2026-04-01T13:33:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.00886v1"}, "authors": ["Nan Wang", "Zhiwei Jin", "Chen Chen", "Haonan Lu"], "canonical_source_url": "https://arxiv.org/abs/2604.00886", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-00886/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce7f70ff1f8c7d4058d8fdf00cb6cf1d61c0367a50ba76fc96d356bb07452c60", "submitted_at": "2026-04-01T13:33:27Z", "updated_at": "2026-04-01T13:33:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.00886v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:16Z", "role": "primary", "sha256": "9289bc047b59e93f75703b72f5e92d9277e115238427519c343df550220e351b", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2602.19870%2C2602.22769%2C2602.23166%2C2602.23320%2C2603.00198%2C2603.00842%2C2603.00991%2C2603.01050%2C2603.01145%2C2603.02236%2C2603.02795%2C2603.03278%2C2603.04549%2C2603.06713%2C2603.07106%2C2603.07670%2C2603.07853%2C2603.08706%2C2603.09716%2C2603.10002%2C2603.10664%2C2603.11139%2C2603.11320%2C2603.12483%2C2603.13026%2C2603.13247%2C2603.15714%2C2603.16065%2C2603.16289%2C2603.16839%2C2603.17216%2C2603.17378%2C2603.19313%2C2603.19685%2C2603.21489%2C2603.22527%2C2603.23638%2C2603.24157%2C2603.24639%2C2603.27646%2C2603.28069%2C2603.28407%2C2603.29418%2C2603.29847%2C2604.00528%2C2604.00886%2C2604.01194%2C2604.01438%2C2604.01560%2C2604.02985&max_results=50"}], "identifiers": {"arxiv": "2604.00886"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.00886"}], "name": "PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding", "publication": null, "published_at": "2026-04-01", "record_sha256": "aeea72180720d0f9a6d2f24f5c290ad63ed4f4250cce135b6c7b86ad1db8a7c5", "relations": [], "resource_id": "arxiv:2604.00886", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2604.00886v1", "topics": ["efficiency", "screen-understanding", "training-free", "evaluation"], "updated_at": "2026-04-01"}, "text": "PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding\nResource type: paper\nReview status: curated\nScope: adjacent\n\nPixelPrune removes pixel-redundant patches before the vision encoder, reducing both encoding and downstream language-model work. The training-free method evaluates lossless and controlled-lossy compression on document and GUI tasks.\nPlatforms: unspecified\nTopics: efficiency, screen-understanding, training-free, evaluation\nPublished: 2026-04-01\nAuthors: Nan Wang, Zhiwei Jin, Chen Chen, Haonan Lu\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent cross-domain VLM compression method with actual GUI-benchmark evaluation; the abstract does not establish a full computer-use policy.\n\nPrimary source: https://arxiv.org/abs/2604.00886v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-00886/"}
{"id": "arxiv:2604.00986", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f137028b7608598fb1885a0bf1330e00c9404851571f856665aae1af1d9d63c", "submitted_at": "2026-04-01T14:50:50Z", "updated_at": "2026-04-02T13:51:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.00986v2"}, "authors": ["Zhengyang Tang", "Ke Ji", "Xidong Wang", "Zihan Ye", "Xinyuan Wang", "Yiduo Guo", "Ziniu Li", "Chenxin Li", "Jingyuan Hu", "Shunian Chen", "Tongxu Luo", "Jiaxi Bi", "Zeyu Qin", "Shaobo Wang", "Xin Lai", "Pengyuan Lyu", "Junyi Li", "Can Xu", "Chengquan Zhang", "Han Hu", "Ming Yan", "Benyou Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.00986", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-00986/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3f137028b7608598fb1885a0bf1330e00c9404851571f856665aae1af1d9d63c", "submitted_at": "2026-04-01T14:50:50Z", "updated_at": "2026-04-02T13:51:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.00986v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2604.00986"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.00986"}], "name": "Do Phone-Use Agents Respect Your Privacy?", "publication": null, "published_at": "2026-04-01", "record_sha256": "1eed288345736694aa4bb814e04d0012c868baedacbdf23d30f165dfb78f4760", "relations": [], "resource_id": "arxiv:2604.00986", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.00986v2", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-04-02"}, "text": "Do Phone-Use Agents Respect Your Privacy?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MyPhoneBench to verify whether phone-use agents minimize personal data disclosure when filling forms during benign mobile tasks.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2026-04-01\nAuthors: Zhengyang Tang, Ke Ji, Xidong Wang, Zihan Ye, Xinyuan Wang, Yiduo Guo, Ziniu Li, Chenxin Li, Jingyuan Hu, Shunian Chen, Tongxu Luo, Jiaxi Bi, Zeyu Qin, Shaobo Wang, Xin Lai, Pengyuan Lyu, Junyi Li, Can Xu, Chengquan Zhang, Han Hu, Ming Yan, Benyou Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.00986v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-00986/"}
{"id": "arxiv:2604.01676", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47e2cb6818c8e26889da590d26073ccc84474bb2e3391f9b184e5d7f8a31ef1d", "submitted_at": "2026-04-02T06:27:52Z", "updated_at": "2026-04-04T01:55:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.01676v2"}, "authors": ["Zirui Zhao", "Jun Hao Liew", "Yan Yang", "Wenzhuo Yang", "Ziyang Luo", "Doyen Sahoo", "Silvio Savarese", "Junnan Li"], "canonical_source_url": "https://arxiv.org/abs/2604.01676", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-01676/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47e2cb6818c8e26889da590d26073ccc84474bb2e3391f9b184e5d7f8a31ef1d", "submitted_at": "2026-04-02T06:27:52Z", "updated_at": "2026-04-04T01:55:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.01676v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2604.01676"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.01676"}], "name": "GPA: Learning GUI Process Automation from Demonstrations", "publication": null, "published_at": "2026-04-02", "record_sha256": "ca65b04d0d7b0741dfc67f0607d87100e1a02e7fc27e614f77e60876c18d054e", "relations": [], "resource_id": "arxiv:2604.01676", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.01676v2", "topics": ["automation", "workflow", "robustness"], "updated_at": "2026-04-04"}, "text": "GPA: Learning GUI Process Automation from Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nPresents demonstration-based GUI process replay with Sequential Monte Carlo localization and readiness checks. GPA executes the learned workflow locally and can serve as the GUI execution component of an agent using MCP or CLI tools.\nPlatforms: unspecified\nTopics: automation, workflow, robustness\nPublished: 2026-04-02\nAuthors: Zirui Zhao, Jun Hao Liew, Yan Yang, Wenzhuo Yang, Ziyang Luo, Doyen Sahoo, Silvio Savarese, Junnan Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.01676v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-01676/"}
{"id": "arxiv:2604.02345", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac86c2ff2bf18feb3edfbe76285d33718d0c6d9564aff5a9d62f8778cdc5147", "submitted_at": "2026-02-11T17:02:38Z", "updated_at": "2026-02-11T17:02:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.02345v1"}, "authors": ["Mengzhou Wu", "Yuzhe Guo", "Yuan Cao", "Haochuan Lu", "Songhe Zhu", "Pingzhe Qu", "Xin Chen", "Kang Qin", "Zhongpu Wang", "Xiaode Zhang", "Xinyi Wang", "Wei Dai", "Gang Cao", "Yuetang Deng", "Zhi Gong", "Dezhi Ran", "Linyi Li", "Wei Yang", "Tao Xie"], "canonical_source_url": "https://arxiv.org/abs/2604.02345", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-02345/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bac86c2ff2bf18feb3edfbe76285d33718d0c6d9564aff5a9d62f8778cdc5147", "submitted_at": "2026-02-11T17:02:38Z", "updated_at": "2026-02-11T17:02:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.02345v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.02345"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.02345"}], "name": "UI-Oceanus: Scaling GUI Agents with Synthetic Environmental Dynamics", "publication": null, "published_at": "2026-02-11", "record_sha256": "e04dcec5e0e0011b881c52ce0758f83432f6349f7ba5594469bd350f30eea00e", "relations": [], "resource_id": "arxiv:2604.02345", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.02345v1", "topics": ["world-models", "data-generation", "state-tracking", "evaluation"], "updated_at": "2026-02-11"}, "text": "UI-Oceanus: Scaling GUI Agents with Synthetic Environmental Dynamics\nResource type: paper\nReview status: curated\nScope: core\n\nUI-Oceanus converts autonomous GUI exploration and execution feedback into supervision for predicting future interface states. It compares forward and inverse objectives during continual pretraining and examines downstream offline and online navigation as synthetic data scales.\nPlatforms: unspecified\nTopics: world-models, data-generation, state-tracking, evaluation\nPublished: 2026-02-11\nAuthors: Mengzhou Wu, Yuzhe Guo, Yuan Cao, Haochuan Lu, Songhe Zhu, Pingzhe Qu, Xin Chen, Kang Qin, Zhongpu Wang, Xiaode Zhang, Xinyi Wang, Wei Dai, Gang Cao, Yuetang Deng, Zhi Gong, Dezhi Ran, Linyi Li, Wei Yang, Tao Xie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.02345v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-02345/"}
{"id": "arxiv:2604.02623", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a37f4bdc495e10e4ae301ebff6faa28cdfa6de152d4619c051083b5dce41e6ba", "submitted_at": "2026-04-03T01:25:12Z", "updated_at": "2026-04-07T14:45:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.02623v2"}, "authors": ["Wei Zou", "Mingwen Dong", "Miguel Romero Calvo", "Shuaichen Chang", "Jiang Guo", "Dongkyu Lee", "Xing Niu", "Xiaofei Ma", "Yanjun Qi", "Jiarong Jiang"], "canonical_source_url": "https://arxiv.org/abs/2604.02623", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-02623/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a37f4bdc495e10e4ae301ebff6faa28cdfa6de152d4619c051083b5dce41e6ba", "submitted_at": "2026-04-03T01:25:12Z", "updated_at": "2026-04-07T14:45:15Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.02623v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.02623"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.02623"}], "name": "Poison Once, Exploit Forever: Environment-Injected Memory Poisoning Attacks on Web Agents", "publication": null, "published_at": "2026-04-03", "record_sha256": "6c577aa28b48f3980bcfdf821fbd82119f407ed7005d8f7cbb8368319cff4d61", "relations": [], "resource_id": "arxiv:2604.02623", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.02623v2", "topics": ["safety-security", "memory", "web-navigation"], "updated_at": "2026-04-07"}, "text": "Poison Once, Exploit Forever: Environment-Injected Memory Poisoning Attacks on Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies persistent web-agent memory contamination introduced through environmental observations instead of direct memory access. Experiments on WebArena and VisualWebArena examine cross-session effects and how interaction failures alter susceptibility.\nPlatforms: web\nTopics: safety-security, memory, web-navigation\nPublished: 2026-04-03\nAuthors: Wei Zou, Mingwen Dong, Miguel Romero Calvo, Shuaichen Chang, Jiang Guo, Dongkyu Lee, Xing Niu, Xiaofei Ma, Yanjun Qi, Jiarong Jiang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.02623v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-02623/"}
{"id": "arxiv:2604.04399", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "856e9920aea7d821061b3258b31aa32fd37b5579d628f5507f45ba83f237b77b", "submitted_at": "2026-04-06T03:58:43Z", "updated_at": "2026-04-06T03:58:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.04399v1"}, "authors": ["Yuwen Zhai", "Runze Li", "Liang Wang", "Nian Shi", "Liwu Xu", "Wei Zhang", "Ran Lin", "Bo Xu", "Benlei Cui"], "canonical_source_url": "https://arxiv.org/abs/2604.04399", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-04399/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "856e9920aea7d821061b3258b31aa32fd37b5579d628f5507f45ba83f237b77b", "submitted_at": "2026-04-06T03:58:43Z", "updated_at": "2026-04-06T03:58:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.04399v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.04399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.04399"}], "name": "GUIDE: Interpretable GUI Agent Evaluation via Hierarchical Diagnosis", "publication": null, "published_at": "2026-04-06", "record_sha256": "b986017ebe90bbe7370af6a6398c0259da76fc864cbdc76b605befa50d61f57d", "relations": [], "resource_id": "arxiv:2604.04399", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.04399v1", "topics": ["evaluation", "long-horizon", "state-tracking"], "updated_at": "2026-04-06"}, "text": "GUIDE: Interpretable GUI Agent Evaluation via Hierarchical Diagnosis\nResource type: paper\nReview status: curated\nScope: core\n\nGUIDE evaluates GUI trajectories by segmenting them into subtasks, diagnosing each segment, and aggregating structured findings into a task verdict. It tests diagnostic judgments on e-commerce, web-agent and Android-control traces rather than relying on one holistic trajectory score.\nPlatforms: web, android\nTopics: evaluation, long-horizon, state-tracking\nPublished: 2026-04-06\nAuthors: Yuwen Zhai, Runze Li, Liang Wang, Nian Shi, Liwu Xu, Wei Zhang, Ran Lin, Bo Xu, Benlei Cui\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.04399v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-04399/"}
{"id": "arxiv:2604.04918", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1411ee71ad8f65e4941af4ef538f52da6e9bf9c57fb537b6ecf37c8478609693", "submitted_at": "2026-04-06T17:57:16Z", "updated_at": "2026-09-12T06:05:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.04918v2"}, "authors": ["Chaoran Chen", "Zhiping Zhang", "Zeya Chen", "Eryue Xu", "Yinuo Yang", "Ibrahim Khalilov", "Simret A Gebreegziabher", "Yanfang Ye", "Ziang Xiao", "Yaxing Yao", "Tianshi Li", "Toby Jia-Jun Li"], "canonical_source_url": "https://arxiv.org/abs/2604.04918", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-04918/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1411ee71ad8f65e4941af4ef538f52da6e9bf9c57fb537b6ecf37c8478609693", "submitted_at": "2026-04-06T17:57:16Z", "updated_at": "2026-09-12T06:05:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.04918v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.04918"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.04918"}], "name": "Comparing Human Oversight Strategies for Computer-Use Agents", "publication": null, "published_at": "2026-04-06", "record_sha256": "08a80713f03f5c1dfae41b843cd1cddb8c8fa3fa9f9c315f3fd61684e8516e23", "relations": [], "resource_id": "arxiv:2604.04918", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.04918v2", "topics": ["human-in-the-loop", "safety-security", "evaluation"], "updated_at": "2026-09-12"}, "text": "Comparing Human Oversight Strategies for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCompares four human oversight strategies during live web-agent sessions, examining exposure to problematic actions and opportunities to intervene. The analysis identifies timing and attention problems in control handoffs and upfront plan approval.\nPlatforms: web\nTopics: human-in-the-loop, safety-security, evaluation\nPublished: 2026-04-06\nAuthors: Chaoran Chen, Zhiping Zhang, Zeya Chen, Eryue Xu, Yinuo Yang, Ibrahim Khalilov, Simret A Gebreegziabher, Yanfang Ye, Ziang Xiao, Yaxing Yao, Tianshi Li, Toby Jia-Jun Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.04918v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-04918/"}
{"id": "arxiv:2604.05157", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "238aac264fb21022f7c546ae3aa3f4ea104c381f5d318f6d962f22fb00095216", "submitted_at": "2026-04-06T20:39:30Z", "updated_at": "2026-05-28T08:13:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.05157v3"}, "authors": ["Rongqian Chen", "Yu Li", "Zeyu Fang", "Sizhe Tang", "Weidong Cao", "Tian Lan"], "canonical_source_url": "https://arxiv.org/abs/2604.05157", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-05157/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "238aac264fb21022f7c546ae3aa3f4ea104c381f5d318f6d962f22fb00095216", "submitted_at": "2026-04-06T20:39:30Z", "updated_at": "2026-05-28T08:13:55Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.05157v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.05157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.05157"}], "name": "IntentScore: Intent-Conditioned Action Evaluation for Computer-Use Agents", "publication": null, "published_at": "2026-04-06", "record_sha256": "0244122f2bd145dac87a288ed170edf0976a52aae80a2f26287f03492408121b", "relations": [], "resource_id": "arxiv:2604.05157", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.05157v3", "topics": ["reward-models", "planning", "evaluation"], "updated_at": "2026-05-28"}, "text": "IntentScore: Intent-Conditioned Action Evaluation for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains IntentScore on offline desktop GUI interactions to assess candidate actions using both their state relevance and planning intent. The reward model is evaluated as an action reranker for Agent S3 on OSWorld.\nPlatforms: desktop\nTopics: reward-models, planning, evaluation\nPublished: 2026-04-06\nAuthors: Rongqian Chen, Yu Li, Zeyu Fang, Sizhe Tang, Weidong Cao, Tian Lan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.05157v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-05157/"}
{"id": "arxiv:2604.05477", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c9ef656cbee786bec0d37ac4915bfd04adb36410eda8e647fab0d8605f90753", "submitted_at": "2026-04-07T06:16:23Z", "updated_at": "2026-04-07T06:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.05477v1"}, "authors": ["Yuzhe Zhang", "Xianwei Xue", "Xingyong Wu", "Mengke Chen", "Chen Liu", "Xinran He", "Run Shao", "Feiran Liu", "Huanmin Xu", "Qiutong Pan", "Haiwei Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.05477", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-05477/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7c9ef656cbee786bec0d37ac4915bfd04adb36410eda8e647fab0d8605f90753", "submitted_at": "2026-04-07T06:16:23Z", "updated_at": "2026-04-07T06:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.05477v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.05477"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.05477"}], "name": "Don't Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-Correction", "publication": null, "published_at": "2026-04-07", "record_sha256": "dc8cb3b904788f64b46244e5bd9696b5ff9f9cfb3a4d47cef062424fa7a48817", "relations": [], "resource_id": "arxiv:2604.05477", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.05477v1", "topics": ["robustness", "reinforcement-learning", "state-tracking"], "updated_at": "2026-04-07"}, "text": "Don't Act Blindly: Robust GUI Automation via Action-Effect Verification and Self-Correction\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces VeriGUI, which checks expected action effects and reasons about recovery before subsequent GUI actions. Training combines synthetic failure trajectories with verification-weighted GRPO, evaluated on an AndroidControl-based robustness benchmark.\nPlatforms: android\nTopics: robustness, reinforcement-learning, state-tracking\nPublished: 2026-04-07\nAuthors: Yuzhe Zhang, Xianwei Xue, Xingyong Wu, Mengke Chen, Chen Liu, Xinran He, Run Shao, Feiran Liu, Huanmin Xu, Qiutong Pan, Haiwei Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.05477v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-05477/"}
{"id": "arxiv:2604.06126", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9cb6daad8832261ebe861fc1cb74429b819305e14ecfeb9698616013e6ac8", "submitted_at": "2026-04-07T17:38:15Z", "updated_at": "2026-04-07T17:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06126v1"}, "authors": ["Pranjal Aggarwal", "Graham Neubig", "Sean Welleck"], "canonical_source_url": "https://arxiv.org/abs/2604.06126", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-06126/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9cb6daad8832261ebe861fc1cb74429b819305e14ecfeb9698616013e6ac8", "submitted_at": "2026-04-07T17:38:15Z", "updated_at": "2026-04-07T17:38:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06126v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.06126"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06126"}], "name": "Gym-Anything: Turn any Software into an Agent Environment", "publication": null, "published_at": "2026-04-07", "record_sha256": "55cb63c0bc78509ccd58cea8251cd7dfe369e5f794acd4c140a433a201a603a8", "relations": [], "resource_id": "arxiv:2604.06126", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.06126v1", "topics": ["data-generation", "sandbox", "long-horizon", "evaluation"], "updated_at": "2026-04-07"}, "text": "Gym-Anything: Turn any Software into an Agent Environment\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Gym-Anything, where a setup agent configures software environments and an independent auditor checks the resulting evidence. CUA-World supplies long-horizon tasks across professional applications, with experiments on trajectory distillation and test-time auditing.\nPlatforms: unspecified\nTopics: data-generation, sandbox, long-horizon, evaluation\nPublished: 2026-04-07\nAuthors: Pranjal Aggarwal, Graham Neubig, Sean Welleck\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.06126v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-06126/"}
{"id": "arxiv:2604.06240", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b452684b305c4f20ded686289b448fbe4f08bb3b34f515a43e2ce5d6dc4ef506", "submitted_at": "2026-04-05T06:26:05Z", "updated_at": "2026-04-05T06:26:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06240v1"}, "authors": ["Corby Rosset", "Pratyusha Sharma", "Andrew Zhao", "Miguel Gonzalez-Fernandez", "Ahmed Awadallah"], "canonical_source_url": "https://arxiv.org/abs/2604.06240", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-06240/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b452684b305c4f20ded686289b448fbe4f08bb3b34f515a43e2ce5d6dc4ef506", "submitted_at": "2026-04-05T06:26:05Z", "updated_at": "2026-04-05T06:26:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.06240v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.06240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06240"}], "name": "The Art of Building Verifiers for Computer Use Agents", "publication": null, "published_at": "2026-04-05", "record_sha256": "d56e72dde160b19dc42d6afce590160c481a644afd5217f46e700107fde629e6", "relations": [], "resource_id": "arxiv:2604.06240", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.06240v1", "topics": ["reward-models", "evaluation", "long-horizon"], "updated_at": "2026-04-05"}, "text": "The Art of Building Verifiers for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a web-trajectory verifier built around structured rubrics, separate process and outcome judgments, and explicit treatment of uncontrollable failures. CUAVerifierBench supplies human labels for evaluating these choices and screenshot-based context management.\nPlatforms: web\nTopics: reward-models, evaluation, long-horizon\nPublished: 2026-04-05\nAuthors: Corby Rosset, Pratyusha Sharma, Andrew Zhao, Miguel Gonzalez-Fernandez, Ahmed Awadallah\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.06240v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-06240/"}
{"id": "arxiv:2604.06367", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "400d8a9f7aed4f150dd5198d9a1561d2df0454069495594ca165ad4254a1be59", "submitted_at": "2026-04-07T18:43:21Z", "updated_at": "2026-06-16T22:37:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.06367v2"}, "authors": ["Guruprasad Viswanathan Ramesh", "Asmit Nayak", "Basieem Siddique", "Kassem Fawaz"], "canonical_source_url": "https://arxiv.org/abs/2604.06367", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-06367/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "400d8a9f7aed4f150dd5198d9a1561d2df0454069495594ca165ad4254a1be59", "submitted_at": "2026-04-07T18:43:21Z", "updated_at": "2026-06-16T22:37:53Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.06367v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.06367"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06367"}], "name": "WebSP-Eval: Evaluating Web Agents on Website Security and Privacy Tasks", "publication": null, "published_at": "2026-04-07", "record_sha256": "036b4c74bf1c6b49be2d2cd0d48b3560deb71d273eb271950473815ce7ecfda4", "relations": [], "resource_id": "arxiv:2604.06367", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.06367v2", "topics": ["safety-security", "evaluation", "web-navigation", "state-tracking"], "updated_at": "2026-06-16"}, "text": "WebSP-Eval: Evaluating Web Agents on Website Security and Privacy Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nWebSP-Eval evaluates agents carrying out website security and privacy tasks such as preference changes and session management. A browser extension controls account and initial state across runs, enabling analysis of failures on stateful UI controls.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation, state-tracking\nPublished: 2026-04-07\nAuthors: Guruprasad Viswanathan Ramesh, Asmit Nayak, Basieem Siddique, Kassem Fawaz\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.06367v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-06367/"}
{"id": "arxiv:2604.06995", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b99643e43c5e20af088d6fa62152e03256c947af76a6b3c0abfc6f74a0f7d648", "submitted_at": "2026-04-08T12:12:09Z", "updated_at": "2026-10-02T06:30:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.06995v3"}, "authors": ["Songze Li", "Xiaoke Guo", "Tianqi Liu", "Biao Yi", "Zhaoyan Gong", "Zhiqiang Liu", "Huajun Chen", "Wen Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.06995", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-06995/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b99643e43c5e20af088d6fa62152e03256c947af76a6b3c0abfc6f74a0f7d648", "submitted_at": "2026-04-08T12:12:09Z", "updated_at": "2026-10-02T06:30:10Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.06995v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2604.06995"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.06995"}], "name": "What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning", "publication": null, "published_at": "2026-04-08", "record_sha256": "fadecef84a2f3defb0d036be455c6cd90532ff44f91b077b1930c7cf79cb23c9", "relations": [], "resource_id": "arxiv:2604.06995", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.06995v3", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": "2026-10-02"}, "text": "What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UILoop, which inserts explicit UI-element localization, function and usage reasoning between screen observation and action. UI Comprehension-Bench evaluates these element-level capabilities with three metrics.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, evaluation\nPublished: 2026-04-08\nAuthors: Songze Li, Xiaoke Guo, Tianqi Liu, Biao Yi, Zhaoyan Gong, Zhiqiang Liu, Huajun Chen, Wen Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.06995v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-06995/"}
{"id": "arxiv:2604.07277", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4d0dbeac950760f4f676ab7c651a881fd275f4adb69d46a94a11c4f10a03980", "submitted_at": "2026-04-08T16:40:26Z", "updated_at": "2026-05-08T03:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07277v2"}, "authors": ["Guo Gan", "Yuxuan Ding", "Cong Chen", "Yuwei Ren", "Yin Huang", "Hong Zhou"], "canonical_source_url": "https://arxiv.org/abs/2604.07277", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-07277/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f4d0dbeac950760f4f676ab7c651a881fd275f4adb69d46a94a11c4f10a03980", "submitted_at": "2026-04-08T16:40:26Z", "updated_at": "2026-05-08T03:06:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07277v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2604.07277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.07277"}], "name": "Android Coach: Improve Online Agentic Training Efficiency with Single State Multiple Actions", "publication": null, "published_at": "2026-04-08", "record_sha256": "26208294d74f068cb748ef66c66f2e066e193f3f3081a69b0cdd8f1e178bd120", "relations": [], "resource_id": "arxiv:2604.07277", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.07277v2", "topics": ["reinforcement-learning", "reward-models", "efficiency"], "updated_at": "2026-05-08"}, "text": "Android Coach: Improve Online Agentic Training Efficiency with Single State Multiple Actions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Android Coach, which uses a learned critic to evaluate multiple candidate actions at each observed emulator state. Process rewards and a group-wise advantage estimator support online training, evaluated on AndroidLab and AndroidWorld.\nPlatforms: android\nTopics: reinforcement-learning, reward-models, efficiency\nPublished: 2026-04-08\nAuthors: Guo Gan, Yuxuan Ding, Cong Chen, Yuwei Ren, Yin Huang, Hong Zhou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.07277v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-07277/"}
{"id": "arxiv:2604.07831", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5427eedff0e97f9c73101403754c38e8928ad71abab7b7eff5f136a30ee2205", "submitted_at": "2026-04-09T05:32:34Z", "updated_at": "2026-07-08T02:39:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07831v2"}, "authors": ["Wenkui Yang", "Chao Jin", "Haisu Zhu", "Weilin Luo", "Derek Yuen", "Kun Shao", "Junxian Duan", "Huaibo Huang", "Jie Cao", "Ran He"], "canonical_source_url": "https://arxiv.org/abs/2604.07831", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-07831/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5427eedff0e97f9c73101403754c38e8928ad71abab7b7eff5f136a30ee2205", "submitted_at": "2026-04-09T05:32:34Z", "updated_at": "2026-07-08T02:39:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.07831v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.07831"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.07831"}], "name": "Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection", "publication": null, "published_at": "2026-04-09", "record_sha256": "dd61494eb92093521689a3d91eb3c6315686de0769ff779f420c6f1d8ad748b8", "relations": [], "resource_id": "arxiv:2604.07831", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.07831v2", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-07-08"}, "text": "Are GUI Agents Focused Enough? Automated Distraction via Semantic-level UI Element Injection\nResource type: paper\nReview status: curated\nScope: core\n\nThis study tests black-box GUI-grounding failures caused by strategically added, innocuous-looking UI elements. Its iterative overlay search is compared with random injection and repeated independent grounding trials to distinguish persistent misdirection from incidental clutter.\nPlatforms: unspecified\nTopics: safety-security, grounding, robustness, evaluation\nPublished: 2026-04-09\nAuthors: Wenkui Yang, Chao Jin, Haisu Zhu, Weilin Luo, Derek Yuen, Kun Shao, Junxian Duan, Huaibo Huang, Jie Cao, Ran He\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.07831v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-07831/"}
{"id": "arxiv:2604.08516", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc0d512a11b88a9036d06d0fcc26a8ab0d0bc54fad986d4b6225c8073897bc2", "submitted_at": "2026-04-09T17:54:02Z", "updated_at": "2026-04-09T17:54:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.08516v1"}, "authors": ["Tanmay Gupta", "Piper Wolters", "Zixian Ma", "Peter Sushko", "Rock Yuren Pang", "Diego Llanes", "Yue Yang", "Taira Anderson", "Boyuan Zheng", "Zhongzheng Ren", "Harsh Trivedi", "Taylor Blanton", "Caleb Ouellette", "Winson Han", "Ali Farhadi", "Ranjay Krishna"], "canonical_source_url": "https://arxiv.org/abs/2604.08516", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-08516/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dbc0d512a11b88a9036d06d0fcc26a8ab0d0bc54fad986d4b6225c8073897bc2", "submitted_at": "2026-04-09T17:54:02Z", "updated_at": "2026-04-09T17:54:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.08516v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.08516"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.08516"}], "name": "MolmoWeb: Open Visual Web Agent and Open Data for the Open Web", "publication": null, "published_at": "2026-04-09", "record_sha256": "82270cb3545f6c81a4ce380feff3c8906af41174808d82db0179cde3971c4083", "relations": [], "resource_id": "arxiv:2604.08516", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.08516v1", "topics": ["web-navigation", "data-generation", "imitation-learning"], "updated_at": "2026-04-09"}, "text": "MolmoWeb: Open Visual Web Agent and Open Data for the Open Web\nResource type: paper\nReview status: curated\nScope: core\n\nPresents MolmoWebMix, combining synthetic browser trajectories, human demonstrations and GUI perception data, and MolmoWeb screenshot-to-action policies trained from it. Browser benchmarks also test parallel rollouts with best-of-N selection.\nPlatforms: web\nTopics: web-navigation, data-generation, imitation-learning\nPublished: 2026-04-09\nAuthors: Tanmay Gupta, Piper Wolters, Zixian Ma, Peter Sushko, Rock Yuren Pang, Diego Llanes, Yue Yang, Taira Anderson, Boyuan Zheng, Zhongzheng Ren, Harsh Trivedi, Taylor Blanton, Caleb Ouellette, Winson Han, Ali Farhadi, Ranjay Krishna\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract states that artifacts will be released; this annotation does not independently verify release status or licenses.\n\nPrimary source: https://arxiv.org/abs/2604.08516v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-08516/"}
{"id": "arxiv:2604.09155", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e53647528b9ee9d98f93ee665d0d886f9b400a129873e240704041df291f5749", "submitted_at": "2026-04-10T09:41:21Z", "updated_at": "2026-04-10T09:41:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09155v1"}, "authors": ["Yushi Feng", "Junye Du", "Qifan Wang", "Zizhan Ma", "Qian Niu", "Yutaka Matsuo", "Long Feng", "Lequan Yu"], "canonical_source_url": "https://arxiv.org/abs/2604.09155", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-09155/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e53647528b9ee9d98f93ee665d0d886f9b400a129873e240704041df291f5749", "submitted_at": "2026-04-10T09:41:21Z", "updated_at": "2026-04-10T09:41:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09155v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.09155"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09155"}], "name": "CORA: Conformal Risk-Controlled Agents for Safeguarded Mobile GUI Automation", "publication": null, "published_at": "2026-04-10", "record_sha256": "81570060a6c32d78c925e7b3006dbbec62ac2fc8d12dc75959096168dec9f16d", "relations": [], "resource_id": "arxiv:2604.09155", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.09155v1", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-04-10"}, "text": "CORA: Conformal Risk-Controlled Agents for Safeguarded Mobile GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CORA, which calibrates a risk model to choose between executing and abstaining from mobile GUI actions, then diagnoses rejected actions to suggest interventions. Phone-Harm provides step-level safety labels for evaluating this safeguarding process.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2026-04-10\nAuthors: Yushi Feng, Junye Du, Qifan Wang, Zizhan Ma, Qian Niu, Yutaka Matsuo, Long Feng, Lequan Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The annotation describes risk calibration without asserting an unconditional safety guarantee.\n\nPrimary source: https://arxiv.org/abs/2604.09155v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-09155/"}
{"id": "arxiv:2604.09581", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "445e520a60e0e3cc0c3c519e21bfa8eebfa33d4fe4320c437e67b94a02d84504", "submitted_at": "2026-02-25T18:59:42Z", "updated_at": "2026-04-14T20:54:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.09581v2"}, "authors": ["Wee Joe Tan", "Zi Rui Lucas Lim", "Shashank Durgad", "Karim Obegi", "Aiden Yiliu Li"], "canonical_source_url": "https://arxiv.org/abs/2604.09581", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-09581/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "445e520a60e0e3cc0c3c519e21bfa8eebfa33d4fe4320c437e67b94a02d84504", "submitted_at": "2026-02-25T18:59:42Z", "updated_at": "2026-04-14T20:54:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.09581v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:43Z", "role": "primary", "sha256": "c432f419481a105a8ce71174ec4aea23c4f0d36ddf89cfe1d0974fa349f6cb59", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=400&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.09581"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:43Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09581"}], "name": "Avenir-UX: Automated UX Evaluation via Simulated Human Web Interaction with GUI Grounding", "publication": null, "published_at": "2026-02-25", "record_sha256": "655e84b8fcc11bceb9202db33f87ed38ae8cddc621d227cf37598433908a16fc", "relations": [], "resource_id": "arxiv:2604.09581", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.09581v2", "topics": ["automation", "evaluation", "ui-design", "grounding"], "updated_at": "2026-04-14"}, "text": "Avenir-UX: Automated UX Evaluation via Simulated Human Web Interaction with GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nAvenir-UX combines grounded web interaction with simulated user profiles and structured usability measures to produce evaluations of website journeys. The paper describes its architecture and use of SUS, step-wise ease questions and think-aloud traces for automated UX reporting.\nPlatforms: web\nTopics: automation, evaluation, ui-design, grounding\nPublished: 2026-02-25\nAuthors: Wee Joe Tan, Zi Rui Lucas Lim, Shashank Durgad, Karim Obegi, Aiden Yiliu Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.09581v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-09581/"}
{"id": "arxiv:2604.09815", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd2bc0ee1955dbc8e98a9306a3d6260251cbab89d36ac117e4dc49b6fa99766e", "submitted_at": "2026-04-10T18:46:05Z", "updated_at": "2026-04-10T18:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09815v1"}, "authors": ["Tiantian He", "Yihang Chen", "Keyue Jiang", "Ka Yiu Lee", "Kaiwen Zhou", "Kun Shao", "Shuai Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.09815", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-09815/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dd2bc0ee1955dbc8e98a9306a3d6260251cbab89d36ac117e4dc49b6fa99766e", "submitted_at": "2026-04-10T18:46:05Z", "updated_at": "2026-04-10T18:46:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09815v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.09815"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09815"}], "name": "EE-MCP: Self-Evolving MCP-GUI Agents via Automated Environment Generation and Experience Learning", "publication": null, "published_at": "2026-04-10", "record_sha256": "078b26541e5bf3d6b3cd12333e4131b9ff75f75d6d2c0bc0b0f6cf86cab27fd7", "relations": [], "resource_id": "arxiv:2604.09815", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.09815v1", "topics": ["tool-integration", "data-generation", "imitation-learning", "memory"], "updated_at": "2026-04-10"}, "text": "EE-MCP: Self-Evolving MCP-GUI Agents via Automated Environment Generation and Experience Learning\nResource type: paper\nReview status: curated\nScope: core\n\nEE-MCP studies application-dependent mixtures of GUI actions and MCP calls using generated environments, trajectory collection and filtered training. An experience bank extracts rules from trajectory comparisons for inference-time improvement, distinct from model distillation.\nPlatforms: desktop\nTopics: tool-integration, data-generation, imitation-learning, memory\nPublished: 2026-04-10\nAuthors: Tiantian He, Yihang Chen, Keyue Jiang, Ka Yiu Lee, Kaiwen Zhou, Kun Shao, Shuai Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.09815v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-09815/"}
{"id": "arxiv:2604.09937", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55cedd4ad9760618fe9b3354e384f69b69b23b65f75837a3898484c9c6e55586", "submitted_at": "2026-04-10T22:33:39Z", "updated_at": "2026-04-10T22:33:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09937v1"}, "authors": ["Suhana Bedi", "Ryan Welch", "Ethan Steinberg", "Michael Wornow", "Taeil Matthew Kim", "Haroun Ahmed", "Peter Sterling", "Bravim Purohit", "Qurat Akram", "Angelic Acosta", "Esther Nubla", "Pritika Sharma", "Michael A. Pfeffer", "Sanmi Koyejo", "Nigam H. Shah"], "canonical_source_url": "https://arxiv.org/abs/2604.09937", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-09937/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "55cedd4ad9760618fe9b3354e384f69b69b23b65f75837a3898484c9c6e55586", "submitted_at": "2026-04-10T22:33:39Z", "updated_at": "2026-04-10T22:33:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.09937v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.09937"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.09937"}], "name": "HealthAdminBench: Evaluating Computer-Use Agents on Healthcare Administration Tasks", "publication": null, "published_at": "2026-04-10", "record_sha256": "7c0fe188ca592cb891d244cd8b93695aa5ed26c6b398d2543738ce91d46a1f6e", "relations": [], "resource_id": "arxiv:2604.09937", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.09937v1", "topics": ["evaluation", "workflow", "industry"], "updated_at": "2026-04-10"}, "text": "HealthAdminBench: Evaluating Computer-Use Agents on Healthcare Administration Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces HealthAdminBench with GUI environments for an electronic health record, payer portals and fax handling. Expert-defined administrative tasks are decomposed into verifiable subtasks to compare local progress with end-to-end workflow completion.\nPlatforms: unspecified\nTopics: evaluation, workflow, industry\nPublished: 2026-04-10\nAuthors: Suhana Bedi, Ryan Welch, Ethan Steinberg, Michael Wornow, Taeil Matthew Kim, Haroun Ahmed, Peter Sterling, Bravim Purohit, Qurat Akram, Angelic Acosta, Esther Nubla, Pritika Sharma, Michael A. Pfeffer, Sanmi Koyejo, Nigam H. Shah\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.09937v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-09937/"}
{"id": "arxiv:2604.10988", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1db8d789fb8ba2d5e61ca24536670fd83491c305f532adfecf7193821f7eea52", "submitted_at": "2026-04-13T04:45:27Z", "updated_at": "2026-04-13T04:45:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.10988v1"}, "authors": ["Peng Yuan", "Yuyang Yin", "Yuxuan Cai", "Zheng Wei"], "canonical_source_url": "https://arxiv.org/abs/2604.10988", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-10988/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1db8d789fb8ba2d5e61ca24536670fd83491c305f532adfecf7193821f7eea52", "submitted_at": "2026-04-13T04:45:27Z", "updated_at": "2026-04-13T04:45:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.10988v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.10988"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.10988"}], "name": "WebForge: Breaking the Realism-Reproducibility-Scalability Trilemma in Browser Agent Benchmark", "publication": null, "published_at": "2026-04-13", "record_sha256": "beae6c6ae8a4aa48993174aa43c987a77c1071b9039d49fee46492e05482bc1e", "relations": [], "resource_id": "arxiv:2604.10988", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.10988v1", "topics": ["data-generation", "evaluation", "web-navigation"], "updated_at": "2026-04-13"}, "text": "WebForge: Breaking the Realism-Reproducibility-Scalability Trilemma in Browser Agent Benchmark\nResource type: paper\nReview status: curated\nScope: core\n\nPresents WebForge, a pipeline that plans, generates, refines and validates self-contained interactive websites. WebForge-Bench uses these environments to evaluate browser agents across domains and controlled difficulty levels.\nPlatforms: web\nTopics: data-generation, evaluation, web-navigation\nPublished: 2026-04-13\nAuthors: Peng Yuan, Yuyang Yin, Yuxuan Cai, Zheng Wei\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.10988v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-10988/"}
{"id": "arxiv:2604.11784", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df56b38203bc5ec8d7e871eb2ad9fca9891a049df62b7d72796fe400eb8bce5e", "submitted_at": "2026-04-13T17:52:04Z", "updated_at": "2026-04-13T17:52:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.11784v1"}, "authors": ["Fei Tang", "Zhiqiong Lu", "Boxuan Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2604.11784", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-11784/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "df56b38203bc5ec8d7e871eb2ad9fca9891a049df62b7d72796fe400eb8bce5e", "submitted_at": "2026-04-13T17:52:04Z", "updated_at": "2026-04-13T17:52:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.11784v1"}}, "environments": ["android", "harmonyos", "ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.11784"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.11784"}], "name": "ClawGUI: A Unified Framework for Training, Evaluating, and Deploying GUI Agents", "publication": null, "published_at": "2026-04-13", "record_sha256": "900cf86fedf96aa3d6e992af4054cb283689a0e180545d9fd5a675591723f351", "relations": [], "resource_id": "arxiv:2604.11784", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.11784v1", "topics": ["reinforcement-learning", "evaluation", "tool-integration", "mobile-navigation"], "updated_at": "2026-04-13"}, "text": "ClawGUI: A Unified Framework for Training, Evaluating, and Deploying GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents ClawGUI, a shared harness for GUI-agent reinforcement learning, benchmark evaluation and deployment on mobile devices. It combines parallel virtual and physical environments, process rewards, and hybrid CLI-GUI execution with persistent memory.\nPlatforms: android, harmonyos, ios\nTopics: reinforcement-learning, evaluation, tool-integration, mobile-navigation\nPublished: 2026-04-13\nAuthors: Fei Tang, Zhiqiong Lu, Boxuan Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.11784v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-11784/"}
{"id": "arxiv:2604.12284", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64d4323050a5ad0f7265231337eb22a4c0823cee41587099c8bdc5db6b320dea", "submitted_at": "2026-04-14T04:50:35Z", "updated_at": "2026-04-14T04:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.12284v1"}, "authors": ["Yulin Chen", "Tri Cao", "Haoran Li", "Yue Liu", "Yibo Li", "Yufei He", "Le Minh Khoi", "Yangqiu Song", "Shuicheng Yan", "Bryan Hooi"], "canonical_source_url": "https://arxiv.org/abs/2604.12284", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-12284/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64d4323050a5ad0f7265231337eb22a4c0823cee41587099c8bdc5db6b320dea", "submitted_at": "2026-04-14T04:50:35Z", "updated_at": "2026-04-14T04:50:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.12284v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.12284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.12284"}], "name": "WebAgentGuard: A Reasoning-Driven Guard Model for Detecting Prompt Injection Attacks in Web Agents", "publication": null, "published_at": "2026-04-14", "record_sha256": "5d6480e271a133a2854a11e245df1014e75fe18c8706a785272f5604a6b233e7", "relations": [], "resource_id": "arxiv:2604.12284", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.12284v1", "topics": ["safety-security", "data-generation", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": "2026-04-14"}, "text": "WebAgentGuard: A Reasoning-Driven Guard Model for Detecting Prompt Injection Attacks in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebAgentGuard separates prompt-injection detection from the operating web agent through a parallel multimodal guard. The guard is trained on synthetic visual/textual examples using reasoning-focused supervised fine-tuning followed by reinforcement learning.\nPlatforms: web\nTopics: safety-security, data-generation, imitation-learning, reinforcement-learning, evaluation\nPublished: 2026-04-14\nAuthors: Yulin Chen, Tri Cao, Haoran Li, Yue Liu, Yibo Li, Yufei He, Le Minh Khoi, Yangqiu Song, Shuicheng Yan, Bryan Hooi\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.12284v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-12284/"}
{"id": "arxiv:2604.13019", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa89e3a09c0ce8b09bdb932f3f00bd2dccf8af4da11a727e675ddf8b2c578146", "submitted_at": "2026-04-14T17:55:46Z", "updated_at": "2026-05-28T06:47:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.13019v3"}, "authors": ["Himangi Mittal", "Gaurav Mittal", "Nelson Daniel Troncoso", "Yu Hu"], "canonical_source_url": "https://arxiv.org/abs/2604.13019", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-13019/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa89e3a09c0ce8b09bdb932f3f00bd2dccf8af4da11a727e675ddf8b2c578146", "submitted_at": "2026-04-14T17:55:46Z", "updated_at": "2026-05-28T06:47:16Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.13019v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.13019"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13019"}], "name": "PrecisionCUA: Iterative Visual Refinement for Pixel-Precise Cursor Grounding in Code Editors", "publication": null, "published_at": "2026-04-14", "record_sha256": "bc8e64d3c9fb31bb80fe574692973967a3607ca31317aba1a2c3091e7123bd6b", "relations": [], "resource_id": "arxiv:2604.13019", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.13019v3", "topics": ["grounding", "robustness", "screen-understanding"], "updated_at": "2026-05-28"}, "text": "PrecisionCUA: Iterative Visual Refinement for Pixel-Precise Cursor Grounding in Code Editors\nResource type: paper\nReview status: curated\nScope: core\n\nStudies iterative cursor grounding in dense code-editor interfaces, using visual feedback from earlier attempts to correct displacement. PrecisionCUA evaluates this closed-loop process across multiple model families on coding-interface tasks.\nPlatforms: unspecified\nTopics: grounding, robustness, screen-understanding\nPublished: 2026-04-14\nAuthors: Himangi Mittal, Gaurav Mittal, Nelson Daniel Troncoso, Yu Hu\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.13019v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-13019/"}
{"id": "arxiv:2604.13318", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf709a3d7096bc07218206edb1accfdda64036a39ea770e637b3e073dd9a76c8", "submitted_at": "2026-04-14T21:48:15Z", "updated_at": "2026-08-31T07:02:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.13318v2"}, "authors": ["Zhaoyang Wang", "Qianhui Wu", "Xuchao Zhang", "Chaoyun Zhang", "Wenlin Yao", "Fazle Elahi Faisal", "Baolin Peng", "Si Qin", "Suman Nath", "Qingwei Lin", "Chetan Bansal", "Dongmei Zhang", "Saravan Rajmohan", "Jianfeng Gao", "Huaxiu Yao"], "canonical_source_url": "https://arxiv.org/abs/2604.13318", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-13318/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf709a3d7096bc07218206edb1accfdda64036a39ea770e637b3e073dd9a76c8", "submitted_at": "2026-04-14T21:48:15Z", "updated_at": "2026-08-31T07:02:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.13318v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.13318"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13318"}], "name": "WebXSkill: Skill Learning for Autonomous Web Agents", "publication": null, "published_at": "2026-04-14", "record_sha256": "7a40b3e0282244c26ba2feb49f7cf39a0a48a8bd957bdb3c6713392e78ae0569", "relations": [], "resource_id": "arxiv:2604.13318", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.13318v2", "topics": ["memory", "planning", "web-navigation", "automation"], "updated_at": "2026-08-31"}, "text": "WebXSkill: Skill Learning for Autonomous Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebXSkill pairs parameterized action programs with step-level language guidance, mining reusable skills from synthetic browser-agent traces. A URL-based graph supports retrieval, and deployment alternates between direct program execution and guided native planning.\nPlatforms: web\nTopics: memory, planning, web-navigation, automation\nPublished: 2026-04-14\nAuthors: Zhaoyang Wang, Qianhui Wu, Xuchao Zhang, Chaoyun Zhang, Wenlin Yao, Fazle Elahi Faisal, Baolin Peng, Si Qin, Suman Nath, Qingwei Lin, Chetan Bansal, Dongmei Zhang, Saravan Rajmohan, Jianfeng Gao, Huaxiu Yao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.13318v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-13318/"}
{"id": "arxiv:2604.13822", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6e81b448c9c22e2fd25dce519db3fdc0585b0265b94d448751d7a37ed910dfa", "submitted_at": "2026-04-15T12:55:46Z", "updated_at": "2026-04-15T12:55:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13822v1"}, "authors": ["Zhengxi Lu", "Fei Tang", "Guangyi Liu", "Kaitao Song", "Xu Tan", "Jin Ma", "Wenqi Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2604.13822", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-13822/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6e81b448c9c22e2fd25dce519db3fdc0585b0265b94d448751d7a37ed910dfa", "submitted_at": "2026-04-15T12:55:46Z", "updated_at": "2026-04-15T12:55:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.13822v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.13822"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.13822"}], "name": "UI-Copilot: Advancing Long-Horizon GUI Automation via Tool-Integrated Policy Optimization", "publication": null, "published_at": "2026-04-15", "record_sha256": "f4c2a731cc51b534fdbfde60a58b71abfe7539c2218022f06dfdd26bce4307ce", "relations": [], "resource_id": "arxiv:2604.13822", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.13822v1", "topics": ["tool-integration", "memory", "planning", "long-horizon", "mobile-navigation"], "updated_at": "2026-04-15"}, "text": "UI-Copilot: Advancing Long-Horizon GUI Automation via Tool-Integrated Policy Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nUI-Copilot separates GUI execution from on-demand memory retrieval and numerical assistance. Its training procedure treats tool selection and multi-turn task execution as distinct optimization targets while keeping persistent observations outside transient context.\nPlatforms: android\nTopics: tool-integration, memory, planning, long-horizon, mobile-navigation\nPublished: 2026-04-15\nAuthors: Zhengxi Lu, Fei Tang, Guangyi Liu, Kaitao Song, Xu Tan, Jin Ma, Wenqi Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.13822v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-13822/"}
{"id": "arxiv:2604.14113", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68709f498ebd5d475593ca8ac5bd543a695086d4dffadc113fad908495c234d9", "submitted_at": "2026-04-15T17:32:28Z", "updated_at": "2026-04-15T17:32:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14113v1"}, "authors": ["Fei Tang", "Bofan Chen", "Zhengxi Lu", "Tongbo Chen", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2604.14113", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-14113/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68709f498ebd5d475593ca8ac5bd543a695086d4dffadc113fad908495c234d9", "submitted_at": "2026-04-15T17:32:28Z", "updated_at": "2026-04-15T17:32:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14113v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.14113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14113"}], "name": "UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI Grounding", "publication": null, "published_at": "2026-04-15", "record_sha256": "852b6aaf4caf53f087a8bfbedd31b9d2f0b93b5ad3a1a26914cca06f05ff4db2", "relations": [], "resource_id": "arxiv:2604.14113", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.14113v1", "topics": ["grounding", "training-free", "efficiency"], "updated_at": "2026-04-15"}, "text": "UI-Zoomer: Uncertainty-Driven Adaptive Zoom-In for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces UI-Zoomer, a training-free grounding method that uses candidate agreement and token confidence to decide when to crop and retry. Prediction variance determines a separate crop radius for each uncertain interface target.\nPlatforms: unspecified\nTopics: grounding, training-free, efficiency\nPublished: 2026-04-15\nAuthors: Fei Tang, Bofan Chen, Zhengxi Lu, Tongbo Chen, Songqin Nong, Tao Jiang, Wenhao Xu, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.14113v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-14113/"}
{"id": "arxiv:2604.14262", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc36ff0c9c4bd39de360824967e53c18fe57562cb8da1462d3e63c0dd348e71", "submitted_at": "2026-04-15T16:39:22Z", "updated_at": "2026-07-01T10:12:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.14262v2"}, "authors": ["Yangyue Wang", "Harshvardhan Sikka", "Yash Mathur", "Tony Zhou", "Jinu Nyachhyon", "Pranav Guruprasad"], "canonical_source_url": "https://arxiv.org/abs/2604.14262", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-14262/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5dc36ff0c9c4bd39de360824967e53c18fe57562cb8da1462d3e63c0dd348e71", "submitted_at": "2026-04-15T16:39:22Z", "updated_at": "2026-07-01T10:12:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.14262v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.14262"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14262"}], "name": "GUI-Perturbed: Domain Randomization Reveals Systematic Brittleness in GUI Grounding Models", "publication": null, "published_at": "2026-04-15", "record_sha256": "ebe4630abf9571a9d8127ccb303b81bd236e88f16b04046018413c5667784955", "relations": [], "resource_id": "arxiv:2604.14262", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.14262v2", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-07-01"}, "text": "GUI-Perturbed: Domain Randomization Reveals Systematic Brittleness in GUI Grounding Models\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-Perturbed, which varies screenshot appearance and instruction wording independently to diagnose grounding failures. Experiments on three related model architectures examine relational instructions, browser zoom and augmentation-based fine-tuning.\nPlatforms: web\nTopics: grounding, robustness, evaluation\nPublished: 2026-04-15\nAuthors: Yangyue Wang, Harshvardhan Sikka, Yash Mathur, Tony Zhou, Jinu Nyachhyon, Pranav Guruprasad\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.14262v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-14262/"}
{"id": "arxiv:2604.14668", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd3b5e5b9a607023c82150673255eaf2f90c22db86d3c049b2a2fd536aaef17f", "submitted_at": "2026-04-16T06:22:48Z", "updated_at": "2026-04-16T06:22:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14668v1"}, "authors": ["Pan Hao", "Rishi Selvakumaran", "Jacob Sun", "Qianwen Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.14668", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-14668/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd3b5e5b9a607023c82150673255eaf2f90c22db86d3c049b2a2fd536aaef17f", "submitted_at": "2026-04-16T06:22:48Z", "updated_at": "2026-04-16T06:22:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14668v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.14668"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14668"}], "name": "Beyond Chat and Clicks: GUI Agents for In-Situ Assistance via Live Interface Transformation", "publication": null, "published_at": "2026-04-16", "record_sha256": "6ebc99d864dda93c2067db9b179a445c9ff55221469d7b606393b61e8cb8e6a6", "relations": [], "resource_id": "arxiv:2604.14668", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.14668v1", "topics": ["human-in-the-loop", "ui-design", "grounding", "evaluation"], "updated_at": "2026-04-16"}, "text": "Beyond Chat and Clicks: GUI Agents for In-Situ Assistance via Live Interface Transformation\nResource type: paper\nReview status: curated\nScope: core\n\nDOMSteer translates help requests and live page context into reversible DOM changes such as control highlighting, contextual tooltips and layout reorganization. The work evaluates in-place GUI assistance with interface tests, use cases and a comparative user study.\nPlatforms: web\nTopics: human-in-the-loop, ui-design, grounding, evaluation\nPublished: 2026-04-16\nAuthors: Pan Hao, Rishi Selvakumaran, Jacob Sun, Qianwen Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.14668v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-14668/"}
{"id": "arxiv:2604.14956", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e2b5d7149adcbe634df0bfdc9c34ba436e51e6891df5c4836daea9795061164", "submitted_at": "2026-04-16T12:53:03Z", "updated_at": "2026-04-16T12:53:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14956v1"}, "authors": ["Wenhao Wang", "Haoting Shi", "Mengying Yuan", "Yiquan Lin", "Panrong Tong", "Hanzhang Zhou", "Guangyi Liu", "Pengxiang Zhao", "Yue Wang", "Siheng Chen"], "canonical_source_url": "https://arxiv.org/abs/2604.14956", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-14956/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1e2b5d7149adcbe634df0bfdc9c34ba436e51e6891df5c4836daea9795061164", "submitted_at": "2026-04-16T12:53:03Z", "updated_at": "2026-04-16T12:53:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.14956v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.14956"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.14956"}], "name": "FedGUI: Benchmarking Federated GUI Agents across Heterogeneous Platforms, Devices, and Operating Systems", "publication": null, "published_at": "2026-04-16", "record_sha256": "f866d78beac607b5fc075291af73b80f30b83c4feb259f9edb9bcb9a91cf5258", "relations": [], "resource_id": "arxiv:2604.14956", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.14956v1", "topics": ["evaluation", "robustness"], "updated_at": "2026-04-16"}, "text": "FedGUI: Benchmarking Federated GUI Agents across Heterogeneous Platforms, Devices, and Operating Systems\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces FedGUI, a federated GUI-learning benchmark built from six datasets spanning mobile, web and desktop interfaces. Its experiments separate platform, device, operating-system and data-source heterogeneity when evaluating collaborative training.\nPlatforms: web, desktop, cross-platform\nTopics: evaluation, robustness\nPublished: 2026-04-16\nAuthors: Wenhao Wang, Haoting Shi, Mengying Yuan, Yiquan Lin, Panrong Tong, Hanzhang Zhou, Guangyi Liu, Pengxiang Zhao, Yue Wang, Siheng Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.14956v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-14956/"}
{"id": "arxiv:2604.15093", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12c2e5e2ae9bdc871bb63d195194c5387129accf5e8b26d12b21650b92fe82b6", "submitted_at": "2026-04-16T14:53:08Z", "updated_at": "2026-04-16T14:53:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.15093v1"}, "authors": ["Kanzhi Cheng", "Zehao Li", "Zheng Ma", "Nuo Chen", "Jialin Cao", "Qiushi Sun", "Zichen Ding", "Fangzhi Xu", "Hang Yan", "Jiajun Chen", "Anh Tuan Luu", "Jianbing Zhang", "Lewei Lu", "Dahua Lin"], "canonical_source_url": "https://arxiv.org/abs/2604.15093", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-15093/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12c2e5e2ae9bdc871bb63d195194c5387129accf5e8b26d12b21650b92fe82b6", "submitted_at": "2026-04-16T14:53:08Z", "updated_at": "2026-04-16T14:53:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.15093v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.15093"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.15093"}], "name": "OpenMobile: Building Open Mobile Agents with Task and Trajectory Synthesis", "publication": null, "published_at": "2026-04-16", "record_sha256": "50cfeb6fa986b31863ae867dd5a2b4b79e4f7f0961bb1fb4466bec3b5cb82ef8", "relations": [], "resource_id": "arxiv:2604.15093", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.15093v1", "topics": ["data-generation", "imitation-learning", "mobile-navigation"], "updated_at": "2026-04-16"}, "text": "OpenMobile: Building Open Mobile Agents with Task and Trajectory Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nPresents OpenMobile, which builds exploration-based environment memory to synthesize task instructions and switches between learner and expert policies during trajectory collection. The resulting data includes recovery behavior and is tested through mobile-agent training and benchmark-overlap analysis.\nPlatforms: android\nTopics: data-generation, imitation-learning, mobile-navigation\nPublished: 2026-04-16\nAuthors: Kanzhi Cheng, Zehao Li, Zheng Ma, Nuo Chen, Jialin Cao, Qiushi Sun, Zichen Ding, Fangzhi Xu, Hang Yan, Jiajun Chen, Anh Tuan Luu, Jianbing Zhang, Lewei Lu, Dahua Lin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.15093v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-15093/"}
{"id": "arxiv:2604.16515", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e47028e752dd023e71414664a193beba74e1c4f1afe765fdf992d560dbea27", "submitted_at": "2026-04-15T10:46:38Z", "updated_at": "2026-04-15T10:46:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16515v1"}, "authors": ["Jiachen Qian", "Zhaolu Kang"], "canonical_source_url": "https://arxiv.org/abs/2604.16515", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-16515/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e47028e752dd023e71414664a193beba74e1c4f1afe765fdf992d560dbea27", "submitted_at": "2026-04-15T10:46:38Z", "updated_at": "2026-04-15T10:46:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16515v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2604.16515"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16515"}], "name": "Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial Perturbations", "publication": null, "published_at": "2026-04-15", "record_sha256": "fe71a00184c6465679fe159f5589953f1ddc1d696143aff745794bf687234dad", "relations": [], "resource_id": "arxiv:2604.16515", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.16515v1", "topics": ["safety-security", "robustness"], "updated_at": "2026-04-15"}, "text": "Penny Wise, Pixel Foolish: Bypassing Price Constraints in Multimodal Agents via Visual Adversarial Perturbations\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PriceBlind, a visual perturbation method for controlled screenshot-based evaluation of price-constrained multimodal agents. The reported transfer experiments use a simplified single-turn coordinate-selection protocol.\nPlatforms: unspecified\nTopics: safety-security, robustness\nPublished: 2026-04-15\nAuthors: Jiachen Qian, Zhaolu Kang\nRecorded arXiv version: v1\n\nScope is screenshot-based interface understanding and safety. The retained abstract does not establish a specific operating system or live transaction execution; no attack results were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.16515v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-16515/"}
{"id": "arxiv:2604.16520", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b1788e6fbb4c14a6d993ce329bf7bf05e3a1e6ffbfc7c14a9fa9d36fbf7ef16", "submitted_at": "2026-04-15T18:22:50Z", "updated_at": "2026-04-15T18:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16520v1"}, "authors": ["Haomin Zhuang", "Hanwen Xing", "Xiangliang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.16520", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-16520/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b1788e6fbb4c14a6d993ce329bf7bf05e3a1e6ffbfc7c14a9fa9d36fbf7ef16", "submitted_at": "2026-04-15T18:22:50Z", "updated_at": "2026-04-15T18:22:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.16520v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2604.16520"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16520"}], "name": "AgentClick: A Skill-Based Human-in-the-Loop Review Layer for Terminal AI Agents", "publication": null, "published_at": "2026-04-15", "record_sha256": "4fdd68c6f25f92c38cb37fa72549b935ec52a7ed6ec0390eb8a387dcf38cd38c", "relations": [], "resource_id": "arxiv:2604.16520", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2604.16520v1", "topics": ["human-in-the-loop", "observability", "workflow", "memory", "tool-integration"], "updated_at": "2026-04-15"}, "text": "AgentClick: A Skill-Based Human-in-the-Loop Review Layer for Terminal AI Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nAgentClick gives terminal-based agents a browser review interface for plans, drafts, execution traces and editable memory. A local server and skill plugin let users inspect proposed actions and provide structured feedback.\nPlatforms: web\nTopics: human-in-the-loop, observability, workflow, memory, tool-integration\nPublished: 2026-04-15\nAuthors: Haomin Zhuang, Hanwen Xing, Xiangliang Zhang\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent supervision interface: the browser UI is for human review of terminal agents; the abstract does not claim that those agents operate graphical interfaces.\n\nPrimary source: https://arxiv.org/abs/2604.16520v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-16520/"}
{"id": "arxiv:2604.16548", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb75ea30086ecf069d8adda1513a6e906cfcfcedcd4cab01d0986a850aa8009a", "submitted_at": "2026-04-17T06:28:22Z", "updated_at": "2026-09-22T07:33:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.16548v3"}, "authors": ["Zehao Lin", "Xixuan Hao", "Renyu Fu", "Shaobo Cui", "Kai Chen", "Chunyu Li", "Zhiyu Li", "Feiyu Xiong"], "canonical_source_url": "https://arxiv.org/abs/2604.16548", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-16548/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb75ea30086ecf069d8adda1513a6e906cfcfcedcd4cab01d0986a850aa8009a", "submitted_at": "2026-04-17T06:28:22Z", "updated_at": "2026-09-22T07:33:24Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.16548v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:08.054642+00:00", "role": "primary", "sha256": "f795535d55a4def0f2ce7dfc41a9d1d22ad38eddf48fb4f1feed98b6b5b66721", "supports": ["summary", "scope", "notes"], "url": "https://arxiv.org/html/2604.16548v3"}], "identifiers": {"arxiv": "2604.16548"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.16548"}], "name": "A Survey on Long-Term Memory Security in LLM Agents: Attacks, Defenses, and Governance Across the Memory Lifecycle", "publication": null, "published_at": "2026-04-17", "record_sha256": "5b0878985b0b9403b9546ff25d2ded7c8118e6f5984261c024ce670f2b42b8f1", "relations": [], "resource_id": "arxiv:2604.16548", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2604.16548v3", "topics": ["memory", "safety-security", "survey"], "updated_at": "2026-09-22"}, "text": "A Survey on Long-Term Memory Security in LLM Agents: Attacks, Defenses, and Governance Across the Memory Lifecycle\nResource type: survey\nReview status: curated\nScope: adjacent\n\nOrganizes security risks and defenses for persistent agent memory into six lifecycle phases and proposes five verifiable governance primitives.\nPlatforms: unspecified\nTopics: memory, safety-security, survey\nPublished: 2026-04-17\nAuthors: Zehao Lin, Xixuan Hao, Renyu Fu, Shaobo Cui, Kai Chen, Chunyu Li, Zhiyu Li, Feiyu Xiong\nRecorded arXiv version: v3\n\nAdjacent memory-security survey. Full-text Sections 3.1 and 3.3 discuss poisoned web observations persisted into later agent actions; the work is broader than GUI agents and does not establish a platform implementation.\n\nPrimary source: https://arxiv.org/abs/2604.16548v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-16548/"}
{"id": "arxiv:2604.17284", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b102c9363b1be77273c7fab6f7f5cde74157c646671c93b9503770fe60dd70c", "submitted_at": "2026-04-19T06:55:59Z", "updated_at": "2026-04-19T06:55:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17284v1"}, "authors": ["Chao Jin", "Wenkui Yang", "Hao Sun", "Yuqi Liao", "Qianyi Jiang", "Kai Zhou", "Jie Cao", "Ran He", "Huaibo Huang"], "canonical_source_url": "https://arxiv.org/abs/2604.17284", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-17284/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b102c9363b1be77273c7fab6f7f5cde74157c646671c93b9503770fe60dd70c", "submitted_at": "2026-04-19T06:55:59Z", "updated_at": "2026-04-19T06:55:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17284v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.17284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17284"}], "name": "HalluClear: Diagnosing, Evaluating and Mitigating Hallucinations in GUI Agents", "publication": null, "published_at": "2026-04-19", "record_sha256": "818f8fe1ee0878fa5da8031cfcae3dcff6082f78f0a9a0629771133924761329", "relations": [], "resource_id": "arxiv:2604.17284", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.17284v1", "topics": ["robustness", "grounding", "evaluation"], "updated_at": "2026-04-19"}, "text": "HalluClear: Diagnosing, Evaluating and Mitigating Hallucinations in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces HalluClear with a GUI hallucination taxonomy, expert-calibrated evaluation workflow and structured-reasoning post-training scheme. It evaluates how these components affect hallucinations, grounding and action fidelity across GUI agents.\nPlatforms: unspecified\nTopics: robustness, grounding, evaluation\nPublished: 2026-04-19\nAuthors: Chao Jin, Wenkui Yang, Hao Sun, Yuqi Liao, Qianyi Jiang, Kai Zhou, Jie Cao, Ran He, Huaibo Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.17284v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-17284/"}
{"id": "arxiv:2604.17817", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "176f4d68e2f5084bee1938ee1399978052cdce2d2250ef06c16ddbf49c977da1", "submitted_at": "2026-04-20T05:15:14Z", "updated_at": "2026-04-20T05:15:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17817v1"}, "authors": ["Shiquan Zhang", "Tianyi Zhang", "Le Fang", "Simon D'Alfonso", "Hong Jia", "Vassilis Kostakos"], "canonical_source_url": "https://arxiv.org/abs/2604.17817", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-17817/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "176f4d68e2f5084bee1938ee1399978052cdce2d2250ef06c16ddbf49c977da1", "submitted_at": "2026-04-20T05:15:14Z", "updated_at": "2026-04-20T05:15:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17817v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}], "identifiers": {"arxiv": "2604.17817"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17817"}], "name": "Do LLMs Need to See Everything? A Benchmark and Study of Failures in LLM-driven Smartphone Automation using Screentext vs. Screenshots", "publication": null, "published_at": "2026-04-20", "record_sha256": "d476da7b43fcf60342f2d8b10c02fa6801e7f5ccb4a14ac92930d8bc606688a0", "relations": [], "resource_id": "arxiv:2604.17817", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.17817v1", "topics": ["evaluation", "mobile-navigation", "screen-understanding", "robustness"], "updated_at": "2026-04-20"}, "text": "Do LLMs Need to See Everything? A Benchmark and Study of Failures in LLM-driven Smartphone Automation using Screentext vs. Screenshots\nResource type: paper\nReview status: curated\nScope: core\n\nDailyDroid compares text-only and text-plus-screenshot agents on everyday Android tasks across multiple apps and difficulty levels. Its failure analysis examines accessibility, observation modality and agent/application behavior.\nPlatforms: android\nTopics: evaluation, mobile-navigation, screen-understanding, robustness\nPublished: 2026-04-20\nAuthors: Shiquan Zhang, Tianyi Zhang, Le Fang, Simon D'Alfonso, Hong Jia, Vassilis Kostakos\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.17817v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-17817/"}
{"id": "arxiv:2604.17821", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "123577fbb4431f22bc18d2d32187744268f9d63b3149e9e1775fb44673771feb", "submitted_at": "2026-04-20T05:19:49Z", "updated_at": "2026-04-21T03:15:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.17821v2"}, "authors": ["Lingfeng Zhang", "Yongan Sun", "Jinpeng Hu", "Hui Ma", "Yang Ying", "Kuien Liu", "Zenglin Shi", "Meng Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.17821", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-17821/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "123577fbb4431f22bc18d2d32187744268f9d63b3149e9e1775fb44673771feb", "submitted_at": "2026-04-20T05:19:49Z", "updated_at": "2026-04-21T03:15:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.17821v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.17821"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17821"}], "name": "WebUncertainty: Dual-Level Uncertainty Driven Planning and Reasoning For Autonomous Web Agent", "publication": null, "published_at": "2026-04-20", "record_sha256": "f4fbbbd45f5a8ad881eb03a64ca96d4737d773fca79cbd74a9ee2793d6f684ec", "relations": [], "resource_id": "arxiv:2604.17821", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.17821v2", "topics": ["planning", "web-navigation", "long-horizon", "robustness"], "updated_at": "2026-04-21"}, "text": "WebUncertainty: Dual-Level Uncertainty Driven Planning and Reasoning For Autonomous Web Agent\nResource type: paper\nReview status: curated\nScope: core\n\nWebUncertainty adapts planning mode to task uncertainty and guides action-level tree search with confidence-derived uncertainty estimates. Its evaluations examine autonomous execution of dynamic, long-horizon web tasks on WebArena and WebVoyager.\nPlatforms: web\nTopics: planning, web-navigation, long-horizon, robustness\nPublished: 2026-04-20\nAuthors: Lingfeng Zhang, Yongan Sun, Jinpeng Hu, Hui Ma, Yang Ying, Kuien Liu, Zenglin Shi, Meng Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.17821v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-17821/"}
{"id": "arxiv:2604.17849", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a508db9eba8a7db5b52273bff32a16189f5a8eed19d6563e0f84313188a0aad1", "submitted_at": "2026-04-20T05:59:04Z", "updated_at": "2026-04-20T05:59:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17849v1"}, "authors": ["Gonzalo Gonzalez-Pumariega", "Saaket Agashe", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.17849", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-17849/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a508db9eba8a7db5b52273bff32a16189f5a8eed19d6563e0f84313188a0aad1", "submitted_at": "2026-04-20T05:59:04Z", "updated_at": "2026-04-20T05:59:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.17849v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.17849"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.17849"}], "name": "On the Reliability of Computer Use Agents", "publication": null, "published_at": "2026-04-20", "record_sha256": "4e345ed75c845e668b4a20ac3260eff31b21bef4f8315f0f0e1bd5ce5f354fe0", "relations": [], "resource_id": "arxiv:2604.17849", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.17849v1", "topics": ["evaluation", "robustness"], "updated_at": "2026-04-20"}, "text": "On the Reliability of Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study examines repeated OSWorld executions to separate execution stochasticity, task-specification ambiguity and behavioral variation. Paired task-level tests show why one successful run is insufficient evidence of reliable computer use.\nPlatforms: desktop\nTopics: evaluation, robustness\nPublished: 2026-04-20\nAuthors: Gonzalo Gonzalez-Pumariega, Saaket Agashe, Jiachen Yang, Ang Li, Xin Eric Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.17849v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-17849/"}
{"id": "arxiv:2604.18240", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "91a9cdb6c8cb124d2e6033f0decc85e0c73bd972984d973ff687bbdc2be21176", "submitted_at": "2026-04-20T13:23:38Z", "updated_at": "2026-04-20T13:23:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18240v1"}, "authors": ["Wentao Shi", "Yu Wang", "Yuyang Zhao", "Yuxin Chen", "Fuli Feng", "Xueyuan Hao", "Xi Su", "Qi Gu", "Hui Su", "Xunliang Cai", "Xiangnan He"], "canonical_source_url": "https://arxiv.org/abs/2604.18240", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-18240/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "91a9cdb6c8cb124d2e6033f0decc85e0c73bd972984d973ff687bbdc2be21176", "submitted_at": "2026-04-20T13:23:38Z", "updated_at": "2026-04-20T13:23:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18240v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2604.18240"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.18240"}], "name": "AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation", "publication": null, "published_at": "2026-04-20", "record_sha256": "855b87502be62ae6c462c154e5cf2c07c862e8a4cd0a889c3ddcc8fa44921357", "relations": [], "resource_id": "arxiv:2604.18240", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.18240v1", "topics": ["evaluation", "reward-models", "state-tracking"], "updated_at": "2026-04-20"}, "text": "AJ-Bench: Benchmarking Agent-as-a-Judge for Environment-Aware Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nAJ-Bench evaluates judge agents that actively gather evidence from search, data systems and graphical interfaces. Its annotated tasks and trajectories test information acquisition, state verification and process verification.\nPlatforms: unspecified\nTopics: evaluation, reward-models, state-tracking\nPublished: 2026-04-20\nAuthors: Wentao Shi, Yu Wang, Yuyang Zhao, Yuxin Chen, Fuli Feng, Xueyuan Hao, Xi Su, Qi Gu, Hui Su, Xunliang Cai, Xiangnan He\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2604.18240v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-18240/"}
{"id": "arxiv:2604.18860", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8eb499729268ef287a9b9bf73275f526d47dabf2d7de399beaead2fe54cfd74e", "submitted_at": "2026-04-20T21:36:16Z", "updated_at": "2026-04-20T21:36:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18860v1"}, "authors": ["Wenpeng Xu"], "canonical_source_url": "https://arxiv.org/abs/2604.18860", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-18860/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8eb499729268ef287a9b9bf73275f526d47dabf2d7de399beaead2fe54cfd74e", "submitted_at": "2026-04-20T21:36:16Z", "updated_at": "2026-04-20T21:36:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.18860v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.18860"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.18860"}], "name": "Temporal UI State Inconsistency in Desktop GUI Agents: Formalizing and Defending Against TOCTOU Attacks on Computer-Use Agents", "publication": null, "published_at": "2026-04-20", "record_sha256": "cba4e9b361946feaab32fc604e0715c5d69677576b45c4dae9e4213776616eb9", "relations": [], "resource_id": "arxiv:2604.18860", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.18860v1", "topics": ["safety-security", "robustness", "evaluation", "state-tracking"], "updated_at": "2026-04-20"}, "text": "Temporal UI State Inconsistency in Desktop GUI Agents: Formalizing and Defending Against TOCTOU Attacks on Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study examines UI changes between a desktop agent’s screenshot observation and action dispatch. Its pre-execution state checks address overlay and focus manipulation, while the reported DOM-injection blind spot limits the defense’s coverage.\nPlatforms: desktop, web\nTopics: safety-security, robustness, evaluation, state-tracking\nPublished: 2026-04-20\nAuthors: Wenpeng Xu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.18860v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-18860/"}
{"id": "arxiv:2604.19742", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "74ec7a508a815781cc7e721aa6c5e351f086f6cab46da48e22c2b58f52af8ce0", "submitted_at": "2026-04-21T17:59:16Z", "updated_at": "2026-09-29T10:11:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.19742v2"}, "authors": ["Zhiyuan Peng", "Wei Tao", "Xin Yin", "Chenhao Ying", "Yuan Luo", "Yiwen Guo"], "canonical_source_url": "https://arxiv.org/abs/2604.19742", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-19742/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "74ec7a508a815781cc7e721aa6c5e351f086f6cab46da48e22c2b58f52af8ce0", "submitted_at": "2026-04-21T17:59:16Z", "updated_at": "2026-09-29T10:11:48Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.19742v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2604.19742"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.19742"}], "name": "PlayCoder: Making LLM-Generated GUI Code Playable", "publication": null, "published_at": "2026-04-21", "record_sha256": "106903476799a5dc9612c9b37bf4d87f38eaef16f021c8517e556d8868acd855", "relations": [], "resource_id": "arxiv:2604.19742", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.19742v2", "topics": ["evaluation", "automation", "ui-design", "workflow"], "updated_at": "2026-09-29"}, "text": "PlayCoder: Making LLM-Generated GUI Code Playable\nResource type: paper\nReview status: curated\nScope: core\n\nPlayEval measures whether generated GUI applications remain logically playable, using PlayTester to execute task-oriented GUI playthroughs. PlayCoder closes the loop between repository-aware generation, interactive testing and targeted code repair.\nPlatforms: unspecified\nTopics: evaluation, automation, ui-design, workflow\nPublished: 2026-04-21\nAuthors: Zhiyuan Peng, Wei Tao, Xin Yin, Chenhao Ying, Yuan Luo, Yiwen Guo\nRecorded arXiv version: v2\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2604.19742v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-19742/"}
{"id": "arxiv:2604.21268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e15661746ab05f2894e23d28b4675e13919bbcd71407fbc33260d84f330ea5d", "submitted_at": "2026-04-23T04:23:31Z", "updated_at": "2026-04-23T04:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.21268v1"}, "authors": ["Wenkai Wang", "Xiyun Li", "Hongcan Guo", "Wenhao Yu", "Tianqing Fang", "Haitao Mi", "Dong Yu", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.21268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-21268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e15661746ab05f2894e23d28b4675e13919bbcd71407fbc33260d84f330ea5d", "submitted_at": "2026-04-23T04:23:31Z", "updated_at": "2026-04-23T04:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.21268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.21268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.21268"}], "name": "Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding", "publication": null, "published_at": "2026-04-23", "record_sha256": "8019089614f272b40fff4482b86e63a1beb3fb111bc584b53b7ec63b71074be5", "relations": [], "resource_id": "arxiv:2604.21268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.21268v1", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": "2026-04-23"}, "text": "Measure Twice, Click Once: Co-evolving Proposer and Visual Critic via Reinforcement Learning for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nTrains a GUI grounding proposer together with a visual critic that evaluates candidate locations rendered on screenshots. An adaptive reinforcement-learning schedule balances their development, using proposal diversity and critic feedback during training.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, self-reflection\nPublished: 2026-04-23\nAuthors: Wenkai Wang, Xiyun Li, Hongcan Guo, Wenhao Yu, Tianqing Fang, Haitao Mi, Dong Yu, Shengyu Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.21268v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-21268/"}
{"id": "arxiv:2604.21375", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e675476121ab3ef51a8627ccfcc86d76eddde73987ea882398f72e3cb31c5d9", "submitted_at": "2026-04-23T07:42:37Z", "updated_at": "2026-09-24T23:58:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.21375v3"}, "authors": ["Qijun Han", "Haoqin Tu", "Zijun Wang", "Haoyue Dai", "Yiyang Zhou", "Nancy Lau", "Alvaro A. Cardenas", "Yuhui Xu", "Ran Xu", "Caiming Xiong", "Zeyu Zheng", "Huaxiu Yao", "Yuyin Zhou", "Cihang Xie"], "canonical_source_url": "https://arxiv.org/abs/2604.21375", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-21375/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e675476121ab3ef51a8627ccfcc86d76eddde73987ea882398f72e3cb31c5d9", "submitted_at": "2026-04-23T07:42:37Z", "updated_at": "2026-09-24T23:58:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.21375v3"}}, "environments": ["desktop", "linux", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2604.21375"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.21375"}], "name": "VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation", "publication": null, "published_at": "2026-04-23", "record_sha256": "13e49e4bb114cc33acee2624d1457ece3b4a0d5a099171b8290fc2fe236a3bd2", "relations": [], "resource_id": "arxiv:2604.21375", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.21375v3", "topics": ["robustness", "self-reflection", "tool-integration", "evaluation"], "updated_at": "2026-09-24"}, "text": "VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation\nResource type: paper\nReview status: curated\nScope: core\n\nVLAA-GUI combines mandatory checks for visually supported completion, multi-tier loop recovery and on-demand workflow search. Optional coding and grounding agents support execution, with component ablations on Linux and Windows GUI tasks.\nPlatforms: desktop, linux, windows\nTopics: robustness, self-reflection, tool-integration, evaluation\nPublished: 2026-04-23\nAuthors: Qijun Han, Haoqin Tu, Zijun Wang, Haoyue Dai, Yiyang Zhou, Nancy Lau, Alvaro A. Cardenas, Yuhui Xu, Ran Xu, Caiming Xiong, Zeyu Zheng, Huaxiu Yao, Yuyin Zhou, Cihang Xie\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.21375v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-21375/"}
{"id": "arxiv:2604.23941", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fe6a322737dd58339673bb1c956345ff585d79df6185ce1f9eaa452b37a1e01", "submitted_at": "2026-04-27T01:29:02Z", "updated_at": "2026-04-27T01:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.23941v1"}, "authors": ["Hongxin Li", "Yuntao Chen", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.23941", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-23941/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fe6a322737dd58339673bb1c956345ff585d79df6185ce1f9eaa452b37a1e01", "submitted_at": "2026-04-27T01:29:02Z", "updated_at": "2026-04-27T01:29:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.23941v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T07:54:03Z", "role": "primary", "sha256": "13d5969845070cd2877dc51ae5300c2d623e5daef00c8a1552ffa46af3f721a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2401.05507%2C2406.07275%2C2409.16427%2C2410.00079%2C2503.07701%2C2503.22738%2C2505.15859%2C2505.17735%2C2505.23596%2C2507.05257%2C2507.06134%2C2507.21953%2C2507.23336%2C2508.06960%2C2508.08997%2C2509.02473%2C2509.23694%2C2510.09116%2C2510.14438%2C2511.18423%2C2512.12818%2C2512.13564%2C2601.10758%2C2601.12449%2C2601.21841%2C2603.01152%2C2603.19993%2C2603.20576%2C2604.04503%2C2604.17562%2C2604.23941%2C2605.16265%2C2606.06448%2C2606.06462%2C2606.29193%2C2607.01647%2C2607.13157&max_results=37"}], "identifiers": {"arxiv": "2604.23941"}, "kind": "paper", "last_verified_at": "2026-10-16T07:54:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.23941"}], "name": "GoClick: Lightweight Element Grounding Model for Autonomous GUI Interaction", "publication": null, "published_at": "2026-04-27", "record_sha256": "2a802f207650b87b77463aea03c6fe280057505910ee4a908bf7cb7f23d2f8f8", "relations": [], "resource_id": "arxiv:2604.23941", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.23941v1", "topics": ["grounding", "efficiency"], "updated_at": "2026-04-27"}, "text": "GoClick: Lightweight Element Grounding Model for Autonomous GUI Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nStudies a compact encoder-decoder grounding model and progressive training-data refinement for resource-constrained GUI interaction.\nPlatforms: cross-platform\nTopics: grounding, efficiency\nPublished: 2026-04-27\nAuthors: Hongxin Li, Yuntao Chen, Zhaoxiang Zhang\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.23941v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-23941/"}
{"id": "arxiv:2604.24348", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "772420683208561df8493c9ed5d68b341e9e098f5729945b972f3555e976d799", "submitted_at": "2026-04-27T11:44:26Z", "updated_at": "2026-04-27T11:44:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24348v1"}, "authors": ["Zheng Wu", "Yi Hua", "Zhaoyuan Huang", "Chenhao Xue", "Yijie Lu", "Pengzhou Cheng", "Zongru Wu", "Lingzhong Dong", "Gongshen Liu", "Xinghao Jiang", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.24348", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-24348/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "772420683208561df8493c9ed5d68b341e9e098f5729945b972f3555e976d799", "submitted_at": "2026-04-27T11:44:26Z", "updated_at": "2026-04-27T11:44:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24348v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2604.24348"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24348"}], "name": "OS-SPEAR: A Toolkit for the Safety, Performance,Efficiency, and Robustness Analysis of OS Agents", "publication": null, "published_at": "2026-04-27", "record_sha256": "07fe37b9837d8720225bd898aedea17a05f9ccd5f5c59f1af158ba5ff55be09a", "relations": [], "resource_id": "arxiv:2604.24348", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.24348v1", "topics": ["evaluation", "safety-security", "efficiency", "robustness"], "updated_at": "2026-04-27"}, "text": "OS-SPEAR: A Toolkit for the Safety, Performance,Efficiency, and Robustness Analysis of OS Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProvides an evaluation toolkit for joint safety, efficiency, and multimodal robustness analysis of operating-system GUI agents.\nPlatforms: unspecified\nTopics: evaluation, safety-security, efficiency, robustness\nPublished: 2026-04-27\nAuthors: Zheng Wu, Yi Hua, Zhaoyuan Huang, Chenhao Xue, Yijie Lu, Pengzhou Cheng, Zongru Wu, Lingzhong Dong, Gongshen Liu, Xinghao Jiang, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.24348v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-24348/"}
{"id": "arxiv:2604.24441", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65325912365effe8adbc39301a378d7526c60d1185691f3a89cb8acd4aa53c3f", "submitted_at": "2026-04-27T13:06:27Z", "updated_at": "2026-04-27T13:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24441v1"}, "authors": ["Hongxin Li", "Xiping Wang", "Jingran Su", "Zheng Ju", "Yuntao Chen", "Qing Li", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.24441", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-24441/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "65325912365effe8adbc39301a378d7526c60d1185691f3a89cb8acd4aa53c3f", "submitted_at": "2026-04-27T13:06:27Z", "updated_at": "2026-04-27T13:06:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24441v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2604.24441"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24441"}], "name": "AutoGUI-v2: A Comprehensive Multi-Modal GUI Functionality Understanding Benchmark", "publication": null, "published_at": "2026-04-27", "record_sha256": "f38d0c7d1a2545bdffc24be108516d6245892973512c4d0bea9438e759c0b08f", "relations": [], "resource_id": "arxiv:2604.24441", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.24441v1", "topics": ["screen-understanding", "grounding", "state-tracking", "evaluation", "data-generation"], "updated_at": "2026-04-27"}, "text": "AutoGUI-v2: A Comprehensive Multi-Modal GUI Functionality Understanding Benchmark\nResource type: paper\nReview status: curated\nScope: core\n\nAutoGUI-v2 evaluates interface functionality, grounding and predicted interaction outcomes using hierarchical regions derived through VLM-human annotation. Its tasks span multiple operating systems and expose differences between functional localization, captioning and transition-logic understanding.\nPlatforms: cross-platform\nTopics: screen-understanding, grounding, state-tracking, evaluation, data-generation\nPublished: 2026-04-27\nAuthors: Hongxin Li, Xiping Wang, Jingran Su, Zheng Ju, Yuntao Chen, Qing Li, Zhaoxiang Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.24441v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-24441/"}
{"id": "arxiv:2604.24964", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db70a5736d04d26f10afe3d9a0cfb08955f65f69480a0345ff13b30b2e83df23", "submitted_at": "2026-04-27T20:05:41Z", "updated_at": "2026-04-27T20:05:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24964v1"}, "authors": ["Lawrence Keunho Jang", "Jing Yu Koh", "Daniel Fried", "Ruslan Salakhutdinov"], "canonical_source_url": "https://arxiv.org/abs/2604.24964", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-24964/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "db70a5736d04d26f10afe3d9a0cfb08955f65f69480a0345ff13b30b2e83df23", "submitted_at": "2026-04-27T20:05:41Z", "updated_at": "2026-04-27T20:05:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.24964v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.24964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.24964"}], "name": "Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks", "publication": null, "published_at": "2026-04-27", "record_sha256": "a0a7f7db1d58451d7aa9b5e634100ed51a0a8bd38ae0742023ffd0f81ea3a9c1", "relations": [], "resource_id": "arxiv:2604.24964", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.24964v1", "topics": ["evaluation", "long-horizon", "efficiency"], "updated_at": "2026-04-27"}, "text": "Odysseys: Benchmarking Web Agents on Realistic Long Horizon Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Odysseys, a live-web benchmark of long, multi-site tasks derived from browsing sessions. Graded task rubrics and a rubric-score-per-step measure evaluate both partial progress and trajectory efficiency.\nPlatforms: web\nTopics: evaluation, long-horizon, efficiency\nPublished: 2026-04-27\nAuthors: Lawrence Keunho Jang, Jing Yu Koh, Daniel Fried, Ruslan Salakhutdinov\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.24964v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-24964/"}
{"id": "arxiv:2604.25380", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b49f54319922268ba880a1c1e8a27ffa7597e228b18be04caa1a6952a6456a", "submitted_at": "2026-04-28T08:43:58Z", "updated_at": "2026-05-08T04:29:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.25380v2"}, "authors": ["Enqi Liu", "Liyuan Pan", "Zhi Gao", "Yan Yang", "Chenrui Shi", "Yang Liu", "Jingrong Wu", "Qing Li"], "canonical_source_url": "https://arxiv.org/abs/2604.25380", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-25380/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54b49f54319922268ba880a1c1e8a27ffa7597e228b18be04caa1a6952a6456a", "submitted_at": "2026-04-28T08:43:58Z", "updated_at": "2026-05-08T04:29:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2604.25380v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.25380"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.25380"}], "name": "Benchmarking and Improving GUI Agents in High-Dynamic Environments", "publication": null, "published_at": "2026-04-28", "record_sha256": "3bb88b570435e62691eb187ba548527f77d70ba7ce91cf11395fc0410900c195", "relations": [], "resource_id": "arxiv:2604.25380", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.25380v2", "topics": ["video", "state-tracking", "evaluation"], "updated_at": "2026-05-08"}, "text": "Benchmarking and Improving GUI Agents in High-Dynamic Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces DynamicGUIBench for applications whose interfaces change between actions, together with DynamicUI, an agent that processes screen recordings. The agent selects informative frames, filters inconsistent action reasoning and uses trajectory reflection for subsequent decisions.\nPlatforms: unspecified\nTopics: video, state-tracking, evaluation\nPublished: 2026-04-28\nAuthors: Enqi Liu, Liyuan Pan, Zhi Gao, Yan Yang, Chenrui Shi, Yang Liu, Jingrong Wu, Qing Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.25380v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-25380/"}
{"id": "arxiv:2604.25562", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2610c2658221d7da0bd888bf7b143ca6c4175791058570f6562cfa0c8c3a946", "submitted_at": "2026-04-28T12:32:21Z", "updated_at": "2026-04-28T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.25562v1"}, "authors": ["Mengyao Du", "Han Fang", "Haokai Ma", "Jiahao Chen", "Kai Xu", "Quanjun Yin", "Ee-Chien Chang"], "canonical_source_url": "https://arxiv.org/abs/2604.25562", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-25562/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d2610c2658221d7da0bd888bf7b143ca6c4175791058570f6562cfa0c8c3a946", "submitted_at": "2026-04-28T12:32:21Z", "updated_at": "2026-04-28T12:32:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.25562v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.25562"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.25562"}], "name": "SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents", "publication": null, "published_at": "2026-04-28", "record_sha256": "68e3bb225ccc8bd5ee3566c7c009bcaa021ae800ab4144f0be7958255fb624f3", "relations": [], "resource_id": "arxiv:2604.25562", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.25562v1", "topics": ["safety-security", "screen-understanding", "evaluation", "efficiency"], "updated_at": "2026-04-28"}, "text": "SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSnapGuard detects screenshot-based prompt injection through complementary visual-stability and recovered textual cues. Its evaluation compares detection quality and inference cost across attack and benign-page settings, rather than establishing universal runtime prevention.\nPlatforms: web\nTopics: safety-security, screen-understanding, evaluation, efficiency\nPublished: 2026-04-28\nAuthors: Mengyao Du, Han Fang, Haokai Ma, Jiahao Chen, Kai Xu, Quanjun Yin, Ee-Chien Chang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.25562v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-25562/"}
{"id": "arxiv:2604.26020", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb6d24f7dc38921da883c4850b37f1c7eb48a03de920356d4ec2e095dbb44b04", "submitted_at": "2026-04-28T18:04:11Z", "updated_at": "2026-04-28T18:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.26020v1"}, "authors": ["Alice Gao", "Weixi Tong", "Rishab Vempati", "Katharina Reinecke", "R. Benjamin Shapiro", "Tianyi Zhang", "Jason Wu"], "canonical_source_url": "https://arxiv.org/abs/2604.26020", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-26020/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb6d24f7dc38921da883c4850b37f1c7eb48a03de920356d4ec2e095dbb44b04", "submitted_at": "2026-04-28T18:04:11Z", "updated_at": "2026-04-28T18:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.26020v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.26020"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.26020"}], "name": "Training Computer Use Agents to Assess the Usability of Graphical User Interfaces", "publication": null, "published_at": "2026-04-28", "record_sha256": "20c053661727f04591c88ab1dbe22e0c8cd86b136b54723896fdbdfd50039222", "relations": [], "resource_id": "arxiv:2604.26020", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.26020v1", "topics": ["ui-design", "evaluation", "automation"], "updated_at": "2026-04-28"}, "text": "Training Computer Use Agents to Assess the Usability of Graphical User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nuxCUA learns to assess interface usability by choosing interaction flows, executing them and predicting a numerical usability score. Training uses interactive UIs with usability labels and human preferences, with evaluations on synthetic and real interfaces.\nPlatforms: unspecified\nTopics: ui-design, evaluation, automation\nPublished: 2026-04-28\nAuthors: Alice Gao, Weixi Tong, Rishab Vempati, Katharina Reinecke, R. Benjamin Shapiro, Tianyi Zhang, Jason Wu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.26020v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-26020/"}
{"id": "arxiv:2604.26148", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b572e37eefab40a6a21fa4e4589324380ad9f66873a4da5746fc0f46aaeb1f6", "submitted_at": "2026-04-28T22:15:06Z", "updated_at": "2026-08-11T15:48:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.26148v3"}, "authors": ["Chen Liang", "Xirui Jiang", "Naihao Deng", "Eytan Adar", "Anhong Guo"], "canonical_source_url": "https://arxiv.org/abs/2604.26148", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-26148/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5b572e37eefab40a6a21fa4e4589324380ad9f66873a4da5746fc0f46aaeb1f6", "submitted_at": "2026-04-28T22:15:06Z", "updated_at": "2026-08-11T15:48:53Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.26148v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2604.26148"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.26148"}], "name": "Beyond Screenshots: Evaluating VLMs' Understanding of UI Animations", "publication": null, "published_at": "2026-04-28", "record_sha256": "d63c7c59775f2bd26842875079cfc8c21fd8f4f445361994dc26f32e3f7d18df", "relations": [], "resource_id": "arxiv:2604.26148", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.26148v3", "topics": ["evaluation", "screen-understanding", "video", "ui-design"], "updated_at": "2026-08-11"}, "text": "Beyond Screenshots: Evaluating VLMs' Understanding of UI Animations\nResource type: paper\nReview status: curated\nScope: core\n\nAniMINT evaluates how VLMs perceive, explain and interpret UI animations using densely annotated videos. The study distinguishes basic motion recognition from higher-level animation meaning and probes motion, context and perceptual cues.\nPlatforms: unspecified\nTopics: evaluation, screen-understanding, video, ui-design\nPublished: 2026-04-28\nAuthors: Chen Liang, Xirui Jiang, Naihao Deng, Eytan Adar, Anhong Guo\nRecorded arXiv version: v3\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2604.26148v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-26148/"}
{"id": "arxiv:2604.27151", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68183be37dbd78314aca35f7f1420e5f2f8da8d8c0042a258b45ae03c90606cf", "submitted_at": "2026-04-29T19:59:36Z", "updated_at": "2026-04-29T19:59:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27151v1"}, "authors": ["Jinbiao Wei", "Kangqi Ni", "Yilun Zhao", "Guo Gan", "Arman Cohan"], "canonical_source_url": "https://arxiv.org/abs/2604.27151", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-27151/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "68183be37dbd78314aca35f7f1420e5f2f8da8d8c0042a258b45ae03c90606cf", "submitted_at": "2026-04-29T19:59:36Z", "updated_at": "2026-04-29T19:59:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27151v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.27151"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27151"}], "name": "Step-level Optimization for Efficient Computer-use Agents", "publication": null, "published_at": "2026-04-29", "record_sha256": "ba06ccde0a95bc8ea8a1f0daa93d5ecd3b1ded2dd96564ed6f7ea252f27cd9a0", "relations": [], "resource_id": "arxiv:2604.27151", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.27151v1", "topics": ["efficiency", "robustness", "state-tracking", "interaction-history", "long-horizon"], "updated_at": "2026-04-29"}, "text": "Step-level Optimization for Efficient Computer-use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProposes a step-level model cascade that runs a smaller computer-use policy until learned monitors flag stalled progress or semantic drift. Separate history-based stuck detection and milestone checks determine when to escalate for recovery or verification without retraining the larger model.\nPlatforms: unspecified\nTopics: efficiency, robustness, state-tracking, interaction-history, long-horizon\nPublished: 2026-04-29\nAuthors: Jinbiao Wei, Kangqi Ni, Yilun Zhao, Guo Gan, Arman Cohan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.27151v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-27151/"}
{"id": "arxiv:2604.27253", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "276d8a62192c0ad576a1d986bb27d3978963b93466a822858e3313957dd83a81", "submitted_at": "2026-04-29T22:57:35Z", "updated_at": "2026-04-29T22:57:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27253v1"}, "authors": ["Fazle Elahi Faisal", "Qianhui Wu", "Baolin Peng", "Jianfeng Gao"], "canonical_source_url": "https://arxiv.org/abs/2604.27253", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-27253/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "276d8a62192c0ad576a1d986bb27d3978963b93466a822858e3313957dd83a81", "submitted_at": "2026-04-29T22:57:35Z", "updated_at": "2026-04-29T22:57:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27253v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.27253"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27253"}], "name": "AutoSurfer -- Teaching Web Agents through Comprehensive Surfing, Learning, and Modeling", "publication": null, "published_at": "2026-04-29", "record_sha256": "bec2ced8048d13272f35c5243a7285181cabc973b7ecf956e65fe1d2155e1373", "relations": [], "resource_id": "arxiv:2604.27253", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.27253v1", "topics": ["web-navigation", "data-generation", "imitation-learning", "evaluation"], "updated_at": "2026-04-29"}, "text": "AutoSurfer -- Teaching Web Agents through Comprehensive Surfing, Learning, and Modeling\nResource type: paper\nReview status: curated\nScope: core\n\nAutoSurfer explores website actions breadth-first, retaining page and trajectory knowledge to guide task synthesis and trajectory refinement. It evaluates the resulting training data by fine-tuning a web agent on WebArena and examining synthesized-task diversity.\nPlatforms: web\nTopics: web-navigation, data-generation, imitation-learning, evaluation\nPublished: 2026-04-29\nAuthors: Fazle Elahi Faisal, Qianhui Wu, Baolin Peng, Jianfeng Gao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2604.27253v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-27253/"}
{"id": "arxiv:2604.27776", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2120860b9c735466a1538ad26c7cda0bc436ab0c6204febdefa08ad66f0036b7", "submitted_at": "2026-04-30T12:13:27Z", "updated_at": "2026-04-30T12:13:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27776v1"}, "authors": ["Jinchao Li", "Yunxin Li", "Chenrui Zhao", "Zhenran Xu", "Baotian Hu", "Min Zhang"], "canonical_source_url": "https://arxiv.org/abs/2604.27776", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-27776/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2120860b9c735466a1538ad26c7cda0bc436ab0c6204febdefa08ad66f0036b7", "submitted_at": "2026-04-30T12:13:27Z", "updated_at": "2026-04-30T12:13:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2604.27776v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2604.27776"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27776"}], "name": "WindowsWorld: A Process-Centric Benchmark of Autonomous GUI Agents in Professional Cross-Application Environments", "publication": null, "published_at": "2026-04-30", "record_sha256": "0fee642849e1e323adb4b8061bd91109fa310379f1569fb549a17c6da2b69907", "relations": [], "resource_id": "arxiv:2604.27776", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.27776v1", "topics": ["evaluation", "cross-app", "workflow"], "updated_at": "2026-04-30"}, "text": "WindowsWorld: A Process-Centric Benchmark of Autonomous GUI Agents in Professional Cross-Application Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WindowsWorld for professional workflows spanning multiple desktop applications, with generated tasks refined through human review. Intermediate subgoals support process-oriented evaluation of cross-application reasoning and execution efficiency.\nPlatforms: desktop\nTopics: evaluation, cross-app, workflow\nPublished: 2026-04-30\nAuthors: Jinchao Li, Yunxin Li, Chenrui Zhao, Zhenran Xu, Baotian Hu, Min Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Desktop use is explicit; no additional operating-system claim is inferred solely from the benchmark name.\n\nPrimary source: https://arxiv.org/abs/2604.27776v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-27776/"}
{"id": "arxiv:2604.27996", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e7b07a811df7e74a28f91a5b6c59c69b7ba066bd6c64452a299e678a3187733", "submitted_at": "2026-04-30T15:22:28Z", "updated_at": "2026-06-26T20:31:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.27996v3"}, "authors": ["Jackson Vonderhorst", "Kuangshi Ai", "Haichao Miao", "Shusen Liu", "Chaoli Wang"], "canonical_source_url": "https://arxiv.org/abs/2604.27996", "citation_url": "https://awesomecua.github.io/resources/arxiv-2604-27996/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8e7b07a811df7e74a28f91a5b6c59c69b7ba066bd6c64452a299e678a3187733", "submitted_at": "2026-04-30T15:22:28Z", "updated_at": "2026-06-26T20:31:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2604.27996v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2604.27996"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2604.27996"}], "name": "Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization", "publication": null, "published_at": "2026-04-30", "record_sha256": "c42e4bfb381e9bfd953028f26132730f426a171716f4f778859821d06ac6194b", "relations": [], "resource_id": "arxiv:2604.27996", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2604.27996v3", "topics": ["science", "evaluation", "tool-integration", "memory"], "updated_at": "2026-06-26"}, "text": "Exploring LLM Agent Designs and Interaction Modalities for Scientific Visualization\nResource type: paper\nReview status: curated\nScope: core\n\nThis study compares specialized, computer-use and coding agents on scientific visualization workflows. It examines GUI, CLI, code and tool-call interaction modes, including how persistent memory changes performance across repeated trials.\nPlatforms: unspecified\nTopics: science, evaluation, tool-integration, memory\nPublished: 2026-04-30\nAuthors: Jackson Vonderhorst, Kuangshi Ai, Haichao Miao, Shusen Liu, Chaoli Wang\nRecorded arXiv version: v3\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2604.27996v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2604-27996/"}
{"id": "arxiv:2605.00551", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87ef518590e29328ce752791df974c9e2215eac656fcf43fed97dd3e847f3cf7", "submitted_at": "2026-05-01T10:16:00Z", "updated_at": "2026-05-01T10:16:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.00551v1"}, "authors": ["Michito Takeshita", "Takuro Kawada", "Takumi Ohashi", "Shunsuke Kitada", "Hitoshi Iyatomi"], "canonical_source_url": "https://arxiv.org/abs/2605.00551", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-00551/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87ef518590e29328ce752791df974c9e2215eac656fcf43fed97dd3e847f3cf7", "submitted_at": "2026-05-01T10:16:00Z", "updated_at": "2026-05-01T10:16:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.00551v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.00551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.00551"}], "name": "A11y-Compressor: A Framework for Enhancing the Efficiency of GUI Agent Observations through Visual Context Reconstruction and Redundancy Reduction", "publication": null, "published_at": "2026-05-01", "record_sha256": "63f4c2802859e81745df565cd0cd83ceafac0671215a5b3f1277e7c7961741fa", "relations": [], "resource_id": "arxiv:2605.00551", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.00551v1", "topics": ["screen-understanding", "efficiency", "grounding", "evaluation"], "updated_at": "2026-05-01"}, "text": "A11y-Compressor: A Framework for Enhancing the Efficiency of GUI Agent Observations through Visual Context Reconstruction and Redundancy Reduction\nResource type: paper\nReview status: curated\nScope: core\n\nA11y-Compressor transforms linearized accessibility trees through modal detection, redundancy reduction and semantic structuring. Its OSWorld evaluation measures both observation-token reduction and task success.\nPlatforms: unspecified\nTopics: screen-understanding, efficiency, grounding, evaluation\nPublished: 2026-05-01\nAuthors: Michito Takeshita, Takuro Kawada, Takumi Ohashi, Shunsuke Kitada, Hitoshi Iyatomi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.00551v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-00551/"}
{"id": "arxiv:2605.01208", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8738f2584f096c1d54555adcf8bdbbb283f0d3178eb1f5e7b7c8a61b2965c15e", "submitted_at": "2026-05-02T02:54:36Z", "updated_at": "2026-05-02T02:54:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.01208v1"}, "authors": ["Haowen Hu", "Pengzhou Cheng", "Zheng Wu", "Lingzhong Dong", "Gongshen Liu", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2605.01208", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-01208/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8738f2584f096c1d54555adcf8bdbbb283f0d3178eb1f5e7b7c8a61b2965c15e", "submitted_at": "2026-05-02T02:54:36Z", "updated_at": "2026-05-02T02:54:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.01208v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.01208"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.01208"}], "name": "Faithful Mobile GUI Agents with Guided Advantage Estimator", "publication": null, "published_at": "2026-05-02", "record_sha256": "c089f9ed3082aac35604208f7497b511884a4fe2c9c7d703f836445394ff7935", "relations": [], "resource_id": "arxiv:2605.01208", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.01208v1", "topics": ["reinforcement-learning", "robustness", "mobile-navigation"], "updated_at": "2026-05-02"}, "text": "Faithful Mobile GUI Agents with Guided Advantage Estimator\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Faithful-Agent, which first teaches abstention under perturbed evidence and then refines GUI policies with an anchor-based, variance-adaptive advantage estimator. A thought-action consistency reward encourages decisions grounded in displayed evidence and instructions.\nPlatforms: unspecified\nTopics: reinforcement-learning, robustness, mobile-navigation\nPublished: 2026-05-02\nAuthors: Haowen Hu, Pengzhou Cheng, Zheng Wu, Lingzhong Dong, Gongshen Liu, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.01208v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-01208/"}
{"id": "arxiv:2605.02630", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dc8c2f5a5a97bc372a1bf5ab3fe0537e53a681f808d45877f55cdef6556858c", "submitted_at": "2026-05-04T14:18:46Z", "updated_at": "2026-05-04T14:18:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02630v1"}, "authors": ["Ruilin Yao", "Shegnwu Xiong", "Tianyu Zou", "Shili Xiong", "Yi Rong"], "canonical_source_url": "https://arxiv.org/abs/2605.02630", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-02630/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2dc8c2f5a5a97bc372a1bf5ab3fe0537e53a681f808d45877f55cdef6556858c", "submitted_at": "2026-05-04T14:18:46Z", "updated_at": "2026-05-04T14:18:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.02630v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.02630"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.02630"}], "name": "AutoFocus: Uncertainty-Aware Active Visual Search for GUI Grounding", "publication": null, "published_at": "2026-05-04", "record_sha256": "3aaa40cd1fb65d97560851ab3cf688dc7a3645adb15595204f5eb83e0101fb46", "relations": [], "resource_id": "arxiv:2605.02630", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.02630v1", "topics": ["grounding", "training-free", "screen-understanding"], "updated_at": "2026-05-04"}, "text": "AutoFocus: Uncertainty-Aware Active Visual Search for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nPresents AutoFocus, a training-free grounding method that turns sampled coordinates and token uncertainty into spatial region proposals. Shape-aware crops and visual comparison select a target while retaining surrounding interface context.\nPlatforms: unspecified\nTopics: grounding, training-free, screen-understanding\nPublished: 2026-05-04\nAuthors: Ruilin Yao, Shegnwu Xiong, Tianyu Zou, Shili Xiong, Yi Rong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.02630v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-02630/"}
{"id": "arxiv:2605.05509", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ff44befcfdb97b67d1680c97784f9e9665d8d9879c093984f42d06c9516a342", "submitted_at": "2026-05-06T23:19:38Z", "updated_at": "2026-05-06T23:19:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.05509v1"}, "authors": ["Sohom Datta", "Alex Nahapetyan", "William Enck", "Alexandros Kapravelos"], "canonical_source_url": "https://arxiv.org/abs/2605.05509", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-05509/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ff44befcfdb97b67d1680c97784f9e9665d8d9879c093984f42d06c9516a342", "submitted_at": "2026-05-06T23:19:38Z", "updated_at": "2026-05-06T23:19:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.05509v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2605.05509"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.05509"}], "name": "WAAA! Web Adversaries Against Agentic Browsers", "publication": null, "published_at": "2026-05-06", "record_sha256": "329017bbe6e7603fe03b1e5758e20d4508301c9dc0b4551a1f365e905e079453", "relations": [], "resource_id": "arxiv:2605.05509", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.05509v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-05-06"}, "text": "WAAA! Web Adversaries Against Agentic Browsers\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates agentic browsers against classical web attacks and visual social-engineering lures beyond indirect prompt injection.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2026-05-06\nAuthors: Sohom Datta, Alex Nahapetyan, William Enck, Alexandros Kapravelos\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.05509v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-05509/"}
{"id": "arxiv:2605.05765", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ece93c00c5556dfe50ddd0ba3f52b725f34d1bee18ea3565951bb78343d8ed0", "submitted_at": "2026-05-07T06:58:34Z", "updated_at": "2026-05-21T06:55:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.05765v2"}, "authors": ["Xiaoming Ren", "Ru Zhen", "Chao Li", "Yang Song", "Qiuxia Hou", "Yanhao Zhang", "Peng Liu", "Qi Qi", "Quanlong Zheng", "Qi Wu", "Zhenyi Liao", "Binqiang Pan", "Haobo Ji", "Haonan Lu"], "canonical_source_url": "https://arxiv.org/abs/2605.05765", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-05765/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ece93c00c5556dfe50ddd0ba3f52b725f34d1bee18ea3565951bb78343d8ed0", "submitted_at": "2026-05-07T06:58:34Z", "updated_at": "2026-05-21T06:55:40Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.05765v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:10.101459+00:00", "role": "primary", "sha256": "80375cd810dcda2e9ccca2c6fbee44730de8449972036c548bde312566b20673", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2605.05765v2"}], "identifiers": {"arxiv": "2605.05765"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:51.952206+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.05765"}], "name": "X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction", "publication": null, "published_at": "2026-05-07", "record_sha256": "6522f1f14db9f523f43babcea3731df776a5edc1958881712d2aa0d37031848b", "relations": [], "resource_id": "arxiv:2605.05765", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.05765v2", "topics": ["mobile-navigation", "memory", "grounding", "workflow"], "updated_at": "2026-05-21"}, "text": "X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes an Android agent architecture combining multimodal perception, working and personal memory, hybrid XML/visual grounding, and reusable skills captured from user interactions.\nPlatforms: android\nTopics: mobile-navigation, memory, grounding, workflow\nPublished: 2026-05-07\nAuthors: Xiaoming Ren, Ru Zhen, Chao Li, Yang Song, Qiuxia Hou, Yanhao Zhang, Peng Liu, Qi Qi, Quanlong Zheng, Qi Wu, Zhenyi Liao, Binqiang Pan, Haobo Ji, Haonan Lu\nRecorded arXiv version: v2\n\nThe report presents demonstrations. Its behavior-cloning component captures runtime skills; it does not establish policy-weight imitation training, and literal trace replay is described as future work. No general benchmark result or release completeness is inferred. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (2 Framework; 3 Omni Perception; 4 Omni Memory; 5.1 Omni Action in App Ecosystem; 5.2 Trajectory Cloned Execution); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2605.05765v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-05765/"}
{"id": "arxiv:2605.06761", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e1fb158dad88d32f8754ac27f12346ba8fbe5f9a5e736fba3e81390071895a3", "submitted_at": "2026-05-07T17:17:10Z", "updated_at": "2026-05-07T17:17:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.06761v1"}, "authors": ["Oğuzhan Fatih Kar", "Roman Bachmann", "Yuanzheng Gong", "Anders Boesen Lindbo Larsen", "Afshin Dehghan"], "canonical_source_url": "https://arxiv.org/abs/2605.06761", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-06761/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2e1fb158dad88d32f8754ac27f12346ba8fbe5f9a5e736fba3e81390071895a3", "submitted_at": "2026-05-07T17:17:10Z", "updated_at": "2026-05-07T17:17:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.06761v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.06761"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.06761"}], "name": "Weblica: Scalable and Reproducible Training Environments for Visual Web Agents", "publication": null, "published_at": "2026-05-07", "record_sha256": "d373b7675b3b84f5cdcf931ca856b522f8f8c64b27ffcc579c98e96cd7ba32e6", "relations": [], "resource_id": "arxiv:2605.06761", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.06761v1", "topics": ["sandbox", "data-generation", "reinforcement-learning"], "updated_at": "2026-05-07"}, "text": "Weblica: Scalable and Reproducible Training Environments for Visual Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Weblica, which combines HTTP-level capture and replay with website-grounded environment synthesis. The resulting interactive web environments support reinforcement learning and tests of navigation transfer and inference-time scaling.\nPlatforms: web\nTopics: sandbox, data-generation, reinforcement-learning\nPublished: 2026-05-07\nAuthors: Oğuzhan Fatih Kar, Roman Bachmann, Yuanzheng Gong, Anders Boesen Lindbo Larsen, Afshin Dehghan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.06761v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-06761/"}
{"id": "arxiv:2605.07134", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6bf74d53b22530951788bd329167f5ba9474827c712590ae9c6d316fac4d62bc", "submitted_at": "2026-05-08T02:11:23Z", "updated_at": "2026-05-08T02:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.07134v1"}, "authors": ["Donguk Kwon", "Dongha Lee"], "canonical_source_url": "https://arxiv.org/abs/2605.07134", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-07134/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6bf74d53b22530951788bd329167f5ba9474827c712590ae9c6d316fac4d62bc", "submitted_at": "2026-05-08T02:11:23Z", "updated_at": "2026-05-08T02:11:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.07134v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.07134"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.07134"}], "name": "Region4Web: Rethinking Observation Space Granularity for Web Agents", "publication": null, "published_at": "2026-05-08", "record_sha256": "72043f97296b29ddf5cfa8a552ca448bb0f45c69f5666baa05ed469a2e284251", "relations": [], "resource_id": "arxiv:2605.07134", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.07134v1", "topics": ["screen-understanding", "web-navigation", "efficiency"], "updated_at": "2026-05-08"}, "text": "Region4Web: Rethinking Observation Space Granularity for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nReorganizes webpage accessibility trees into functional regions through hierarchical decomposition and semantic abstraction. PageDigest supplies a persistent region-level observation to web agents, evaluated on WebArena.\nPlatforms: web\nTopics: screen-understanding, web-navigation, efficiency\nPublished: 2026-05-08\nAuthors: Donguk Kwon, Dongha Lee\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.07134v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-07134/"}
{"id": "arxiv:2605.09317", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "729ff6b37da62bb8a9a911865554520e3b29219d9735e087c3186705428d7b05", "submitted_at": "2026-05-10T04:31:23Z", "updated_at": "2026-05-10T04:31:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09317v1"}, "authors": ["Guibin Zhang", "Yaohui Ling", "Fanci Meng", "Kun Wang", "Shuicheng Yan"], "canonical_source_url": "https://arxiv.org/abs/2605.09317", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-09317/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "729ff6b37da62bb8a9a911865554520e3b29219d9735e087c3186705428d7b05", "submitted_at": "2026-05-10T04:31:23Z", "updated_at": "2026-05-10T04:31:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09317v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.09317"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.09317"}], "name": "Mem-W: Latent Memory-Native GUI Agents", "publication": null, "published_at": "2026-05-10", "record_sha256": "3cb954e0b5d1a4dfd80bcf0ba2a6c74e1181d7cdea3d64b454141c1e4b448656", "relations": [], "resource_id": "arxiv:2605.09317", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.09317v1", "topics": ["memory", "long-horizon", "imitation-learning"], "updated_at": "2026-05-10"}, "text": "Mem-W: Latent Memory-Native GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Mem-W, which compresses previous trajectories and current-session segments into latent memory tokens used directly by the GUI policy. Self-distillation and outcome supervision train this memory representation for web and mobile navigation.\nPlatforms: web\nTopics: memory, long-horizon, imitation-learning\nPublished: 2026-05-10\nAuthors: Guibin Zhang, Yaohui Ling, Fanci Meng, Kun Wang, Shuicheng Yan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.09317v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-09317/"}
{"id": "arxiv:2605.09497", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57d784c70b50d2c25055c6d139c0a5acfc97da3a4effbd63536b72d6df38a69", "submitted_at": "2026-05-10T12:11:52Z", "updated_at": "2026-05-10T12:11:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09497v1"}, "authors": ["Yilin Zhang", "Yingkai Hua", "Chunyu Wei", "Xin Wang", "Yueguo Chen"], "canonical_source_url": "https://arxiv.org/abs/2605.09497", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-09497/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b57d784c70b50d2c25055c6d139c0a5acfc97da3a4effbd63536b72d6df38a69", "submitted_at": "2026-05-10T12:11:52Z", "updated_at": "2026-05-10T12:11:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.09497v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.09497"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.09497"}], "name": "Don't Click That: Teaching Web Agents to Resist Deceptive Interfaces", "publication": null, "published_at": "2026-05-10", "record_sha256": "519f011ac9dffffa2acc0d21961cd4be355f372c361b83b21399b316f95a6c7f", "relations": [], "resource_id": "arxiv:2605.09497", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.09497v1", "topics": ["safety-security", "reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-05-10"}, "text": "Don't Click That: Teaching Web Agents to Resist Deceptive Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nDUDE combines asymmetric hybrid-reward learning with summaries of deceptive-interface failures to guide safer web actions. Its RUC benchmark evaluates deception susceptibility alongside ordinary task performance across interface scenarios.\nPlatforms: web\nTopics: safety-security, reinforcement-learning, evaluation, robustness\nPublished: 2026-05-10\nAuthors: Yilin Zhang, Yingkai Hua, Chunyu Wei, Xin Wang, Yueguo Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.09497v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-09497/"}
{"id": "arxiv:2605.10347", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4374ec1b7c44804bb78ea56bd0aa3214083419f37780d79c5fcfb68468727bbf", "submitted_at": "2026-05-11T10:49:31Z", "updated_at": "2026-05-22T05:43:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.10347v2"}, "authors": ["Weikai Xu", "Kun Huang", "Yunren Feng", "Jiaxing Li", "Yuhan Chen", "Yuxuan Liu", "Zhizheng Jiang", "Heng Qu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Xiaolin Hu", "Bo An"], "canonical_source_url": "https://arxiv.org/abs/2605.10347", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-10347/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4374ec1b7c44804bb78ea56bd0aa3214083419f37780d79c5fcfb68468727bbf", "submitted_at": "2026-05-11T10:49:31Z", "updated_at": "2026-05-22T05:43:30Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.10347v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.10347"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.10347"}], "name": "How Mobile World Model Guides GUI Agents?", "publication": null, "published_at": "2026-05-11", "record_sha256": "2d797608c1f015105cc8f645a39441ab76385157af77f1570d39450c08f8d881", "relations": [], "resource_id": "arxiv:2605.10347", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.10347v2", "topics": ["world-models", "data-generation", "evaluation"], "updated_at": "2026-05-22"}, "text": "How Mobile World Model Guides GUI Agents?\nResource type: paper\nReview status: curated\nScope: core\n\nCompares mobile world models that predict state changes as delta text, full text, images or renderable code. Downstream Android benchmarks examine synthetic training trajectories and test-time guidance, including limits of post-action reflection.\nPlatforms: android\nTopics: world-models, data-generation, evaluation\nPublished: 2026-05-11\nAuthors: Weikai Xu, Kun Huang, Yunren Feng, Jiaxing Li, Yuhan Chen, Yuxuan Liu, Zhizheng Jiang, Heng Qu, Pengzhi Gao, Wei Liu, Jian Luan, Xiaolin Hu, Bo An\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.10347v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-10347/"}
{"id": "arxiv:2605.10966", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0bd75e6bf9d5a8185a2204cdd99d42f9d96c102f7297abcadb9f8425e9246349", "submitted_at": "2026-05-08T10:57:19Z", "updated_at": "2026-05-08T10:57:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.10966v1"}, "authors": ["Chiyeong Heo", "Jaechang Kim", "Junhyuk Kwon", "Hoyoung Kim", "Dongmin Park", "Jonghyun Lee", "Jungseul Ok"], "canonical_source_url": "https://arxiv.org/abs/2605.10966", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-10966/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0bd75e6bf9d5a8185a2204cdd99d42f9d96c102f7297abcadb9f8425e9246349", "submitted_at": "2026-05-08T10:57:19Z", "updated_at": "2026-05-08T10:57:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.10966v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:19Z", "role": "primary", "sha256": "05ec5d668ac75d3d7c5b71ea7d6e76a3fbf5ebeb742e0843db2329e5e41aa208", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2604.03598%2C2604.04759%2C2604.05547%2C2604.06170%2C2604.07340%2C2604.10547%2C2604.10938%2C2604.11978%2C2604.12282%2C2604.13018%2C2604.13346%2C2604.13954%2C2604.15579%2C2604.15671%2C2604.15715%2C2604.16520%2C2604.18240%2C2604.18934%2C2604.19572%2C2604.19667%2C2604.19742%2C2604.20841%2C2604.20987%2C2604.22436%2C2604.23781%2C2604.24479%2C2604.26148%2C2604.27621%2C2604.27996%2C2604.28093%2C2604.28139%2C2604.28181%2C2605.00425%2C2605.01462%2C2605.02240%2C2605.03042%2C2605.03242%2C2605.04811%2C2605.05191%2C2605.05811%2C2605.06110%2C2605.07073%2C2605.07358%2C2605.08374%2C2605.08985%2C2605.10912%2C2605.10966%2C2605.11363%2C2605.11732%2C2605.11817&max_results=50"}], "identifiers": {"arxiv": "2605.10966"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.10966"}], "name": "MMTB: Evaluating Terminal Agents on Multimedia-File Tasks", "publication": null, "published_at": "2026-05-08", "record_sha256": "7ee4fdf850ba62139409b8a32bd6dc67c0fb247698ea5143ca7fcc0f0562a63c", "relations": [], "resource_id": "arxiv:2605.10966", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2605.10966v1", "topics": ["evaluation", "automation", "video", "tool-integration"], "updated_at": "2026-05-08"}, "text": "MMTB: Evaluating Terminal Agents on Multimedia-File Tasks\nResource type: paper\nReview status: curated\nScope: adjacent\n\nMultiMedia-TerminalBench evaluates agents that use terminal tools to work with audio and video files. The Terminus-MM harness adds multimedia perception to study how accessible evidence changes executable workflows and task outcomes.\nPlatforms: unspecified\nTopics: evaluation, automation, video, tool-integration\nPublished: 2026-05-08\nAuthors: Chiyeong Heo, Jaechang Kim, Junhyuk Kwon, Hoyoung Kim, Dongmin Park, Jonghyun Lee, Jungseul Ok\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent computer-automation benchmark using terminal workflows, explicitly separated from graphical-interface control.\n\nPrimary source: https://arxiv.org/abs/2605.10966v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-10966/"}
{"id": "arxiv:2605.11212", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "422d2fdf2d7f665df1a5672e03d837ba41fe9c6832d64c40317f77481f642130", "submitted_at": "2026-05-11T20:27:54Z", "updated_at": "2026-06-05T22:18:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.11212v3"}, "authors": ["Amirhossein Abaskohi", "Yuhang He", "Peter West", "Giuseppe Carenini", "Pranit Chawla", "Vibhav Vineet"], "canonical_source_url": "https://arxiv.org/abs/2605.11212", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-11212/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "422d2fdf2d7f665df1a5672e03d837ba41fe9c6832d64c40317f77481f642130", "submitted_at": "2026-05-11T20:27:54Z", "updated_at": "2026-06-05T22:18:46Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.11212v3"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.11212"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.11212"}], "name": "ReVision: Scaling Computer-Use Agents via Temporal Visual Redundancy Reduction", "publication": null, "published_at": "2026-05-11", "record_sha256": "91ddc5efc22942158ed9de3821e080424e8b50a4b7032467e9559bf83437667a", "relations": [], "resource_id": "arxiv:2605.11212", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.11212v3", "topics": ["efficiency", "interaction-history", "screen-understanding"], "updated_at": "2026-06-05"}, "text": "ReVision: Scaling Computer-Use Agents via Temporal Visual Redundancy Reduction\nResource type: paper\nReview status: curated\nScope: core\n\nTrains GUI agents with a learned selector that removes repeated visual patches across consecutive screenshots while preserving spatial structure. ReVision evaluates this history representation on OSWorld, WebTailBench and AgentNetBench.\nPlatforms: desktop, web\nTopics: efficiency, interaction-history, screen-understanding\nPublished: 2026-05-11\nAuthors: Amirhossein Abaskohi, Yuhang He, Peter West, Giuseppe Carenini, Pranit Chawla, Vibhav Vineet\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.11212v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-11212/"}
{"id": "arxiv:2605.12294", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "388eac2721929fc8211f40cc4edde61be0b6e6145fa60fbaa444ec80822fcc36", "submitted_at": "2026-05-12T15:48:44Z", "updated_at": "2026-05-12T15:48:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12294v1"}, "authors": ["Zerui Qin", "Sheng Yue", "Xingyuan Hua", "Yongjian Fu", "Ju Ren"], "canonical_source_url": "https://arxiv.org/abs/2605.12294", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-12294/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "388eac2721929fc8211f40cc4edde61be0b6e6145fa60fbaa444ec80822fcc36", "submitted_at": "2026-05-12T15:48:44Z", "updated_at": "2026-05-12T15:48:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12294v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.12294"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12294"}], "name": "Executable Agentic Memory for GUI Agent", "publication": null, "published_at": "2026-05-12", "record_sha256": "771818b6633dae56eea1bc35312251f6e7ade5c6c0af64a5238927bac1e85c1c", "relations": [], "resource_id": "arxiv:2605.12294", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.12294v1", "topics": ["memory", "planning", "long-horizon", "efficiency"], "updated_at": "2026-05-12"}, "text": "Executable Agentic Memory for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nExecutable Agentic Memory stores GUI routines in a knowledge graph built through state-aware exploration and action-group mining. A learned value model guides tree search over this executable memory for retrieval-based planning, with AndroidWorld and efficiency evaluation.\nPlatforms: android\nTopics: memory, planning, long-horizon, efficiency\nPublished: 2026-05-12\nAuthors: Zerui Qin, Sheng Yue, Xingyuan Hua, Yongjian Fu, Ju Ren\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.12294v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-12294/"}
{"id": "arxiv:2605.12481", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09128e90b925ac0cd2cf6e967853180bbef1e97a0960c22065f1b0b50c30d58b", "submitted_at": "2026-05-12T17:57:04Z", "updated_at": "2026-09-19T14:53:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.12481v2"}, "authors": ["Xuhao Hu", "Xi Zhang", "Haiyang Xu", "Kyle Qiao", "Jingyi Yang", "Xuanjing Huang", "Jing Shao", "Ming Yan", "Jieping Ye"], "canonical_source_url": "https://arxiv.org/abs/2605.12481", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-12481/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "09128e90b925ac0cd2cf6e967853180bbef1e97a0960c22065f1b0b50c30d58b", "submitted_at": "2026-05-12T17:57:04Z", "updated_at": "2026-09-19T14:53:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.12481v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.12481"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12481"}], "name": "ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents", "publication": null, "published_at": "2026-05-12", "record_sha256": "47a4e1a8882a783a9233fcf84f3589de8eefee6ca8666961a6ae4b8146bf8865", "relations": [], "resource_id": "arxiv:2605.12481", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.12481v2", "topics": ["tool-integration", "data-generation", "imitation-learning", "reinforcement-learning", "planning", "efficiency"], "updated_at": "2026-09-19"}, "text": "ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nToolCUA synthesizes interleaved GUI/tool trajectories and trains agents to choose between graphical actions and higher-level tools. Staged supervised and reinforcement learning culminate in online hybrid-environment training with a reward for efficient execution paths.\nPlatforms: desktop\nTopics: tool-integration, data-generation, imitation-learning, reinforcement-learning, planning, efficiency\nPublished: 2026-05-12\nAuthors: Xuhao Hu, Xi Zhang, Haiyang Xu, Kyle Qiao, Jingyi Yang, Xuanjing Huang, Jing Shao, Ming Yan, Jieping Ye\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.12481v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-12481/"}
{"id": "arxiv:2605.12493", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5893d1d9205776667a1e1d89720ec3baf55fafe80cc9c6201ca97088f44a10ac", "submitted_at": "2026-05-12T17:59:34Z", "updated_at": "2026-05-12T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12493v1"}, "authors": ["Di Wu", "Zixiang Ji", "Asmi Kawatkar", "Bryan Kwan", "Jia-Chen Gu", "Nanyun Peng", "Kai-Wei Chang"], "canonical_source_url": "https://arxiv.org/abs/2605.12493", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-12493/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5893d1d9205776667a1e1d89720ec3baf55fafe80cc9c6201ca97088f44a10ac", "submitted_at": "2026-05-12T17:59:34Z", "updated_at": "2026-05-12T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12493v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "identifiers": {"arxiv": "2605.12493"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12493"}], "name": "LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues", "publication": null, "published_at": "2026-05-12", "record_sha256": "b8bdaad7743a3e6f2e597c1cb45926f70ac7875cfeb6f664dc203419a788ad2d", "relations": [], "resource_id": "arxiv:2605.12493", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.12493v1", "topics": ["memory", "interaction-history", "state-tracking", "workflow", "evaluation"], "updated_at": "2026-05-12"}, "text": "LongMemEval-V2: Evaluating Long-Term Agent Memory Toward Experienced Colleagues\nResource type: paper\nReview status: curated\nScope: core\n\nLongMemEval-V2 asks memory systems to recover environment knowledge from long web-agent histories for question answering. Its AgentRunbook variants contrast indexed knowledge pools with coding-agent retrieval from stored trajectory files.\nPlatforms: web\nTopics: memory, interaction-history, state-tracking, workflow, evaluation\nPublished: 2026-05-12\nAuthors: Di Wu, Zixiang Ji, Asmi Kawatkar, Bryan Kwan, Jia-Chen Gu, Nanyun Peng, Kai-Wei Chang\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2605.12493v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-12493/"}
{"id": "arxiv:2605.12501", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb57ba3cd858513d1da58d64a7c85388f12e331c9ea9f89204be5f43e6fec38", "submitted_at": "2026-05-12T17:59:58Z", "updated_at": "2026-05-12T17:59:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12501v1"}, "authors": ["Miaosen Zhang", "Xiaohan Zhao", "Zhihong Tan", "Zhou Huoshen", "Yijia Fan", "Yifan Yang", "Kai Qiu", "Bei Liu", "Justin Wagle", "Chenzhong Yin", "Mingxi Cheng", "Ji Li", "Qi Dai", "Chong Luo", "Xu Yang", "Xin Geng", "Baining Guo"], "canonical_source_url": "https://arxiv.org/abs/2605.12501", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-12501/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb57ba3cd858513d1da58d64a7c85388f12e331c9ea9f89204be5f43e6fec38", "submitted_at": "2026-05-12T17:59:58Z", "updated_at": "2026-05-12T17:59:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12501v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.12501"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12501"}], "name": "Covering Human Action Space for Computer Use: Data Synthesis and Benchmark", "publication": null, "published_at": "2026-05-12", "record_sha256": "99491a4c5af3e5e465da6744f948ed92d9e37e1f7a6354588fc4cfc0807f8cf7", "relations": [], "resource_id": "arxiv:2605.12501", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.12501v1", "topics": ["grounding", "data-generation", "evaluation", "action-prediction"], "updated_at": "2026-05-12"}, "text": "Covering Human Action Space for Computer Use: Data Synthesis and Benchmark\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CUActSpot to evaluate clicks, drags, drawing and other screen interactions across GUI, text, tables, canvases and natural images. A renderer-based pipeline records geometry and screenshots, then generates matching instructions and action traces for training a grounding model.\nPlatforms: unspecified\nTopics: grounding, data-generation, evaluation, action-prediction\nPublished: 2026-05-12\nAuthors: Miaosen Zhang, Xiaohan Zhao, Zhihong Tan, Zhou Huoshen, Yijia Fan, Yifan Yang, Kai Qiu, Bei Liu, Justin Wagle, Chenzhong Yin, Mingxi Cheng, Ji Li, Qi Dai, Chong Luo, Xu Yang, Xin Geng, Baining Guo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.12501v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-12501/"}
{"id": "arxiv:2605.12549", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb60b76fca3a8825030b1a31a7b97ff31e06f30ac3852f243efe440f965e5578", "submitted_at": "2026-05-10T07:04:07Z", "updated_at": "2026-05-10T07:04:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12549v1"}, "authors": ["Jiaping Lin", "Fei Shen", "Junzhe Li", "Ping Nie", "Fei Yu", "Ming Li", "Haizhou Li"], "canonical_source_url": "https://arxiv.org/abs/2605.12549", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-12549/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cb60b76fca3a8825030b1a31a7b97ff31e06f30ac3852f243efe440f965e5578", "submitted_at": "2026-05-10T07:04:07Z", "updated_at": "2026-05-10T07:04:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.12549v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.12549"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.12549"}], "name": "What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs", "publication": null, "published_at": "2026-05-10", "record_sha256": "8771ac541ed643c7e8d1017790a170a56cbf4460b57107c74341b89dde55ff0f", "relations": [], "resource_id": "arxiv:2605.12549", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.12549v1", "topics": ["grounding", "training-free", "screen-understanding", "evaluation"], "updated_at": "2026-05-10"}, "text": "What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs\nResource type: paper\nReview status: curated\nScope: core\n\nRe-Prefill adds an attention-guided second prefill pass that revisits candidate screen regions before coordinate generation. The study evaluates this training-free intervention across GUI-grounding models and benchmarks.\nPlatforms: unspecified\nTopics: grounding, training-free, screen-understanding, evaluation\nPublished: 2026-05-10\nAuthors: Jiaping Lin, Fei Shen, Junzhe Li, Ping Nie, Fei Yu, Ming Li, Haizhou Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.12549v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-12549/"}
{"id": "arxiv:2605.14290", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572be49190554bbfb6accac4c87ef91608f88e90d06250f229db2922e19fcd56", "submitted_at": "2026-05-14T02:48:57Z", "updated_at": "2026-05-14T02:48:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14290v1"}, "authors": ["Julien Piet", "Annabella Chow", "Yiwei Hou", "Muxi Lyu", "Sylvie Venuto", "Jinhao Zhu", "Raluca Ada Popa", "David Wagner"], "canonical_source_url": "https://arxiv.org/abs/2605.14290", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-14290/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572be49190554bbfb6accac4c87ef91608f88e90d06250f229db2922e19fcd56", "submitted_at": "2026-05-14T02:48:57Z", "updated_at": "2026-05-14T02:48:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.14290v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.14290"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.14290"}], "name": "Web Agents Should Adopt the Plan-Then-Execute Paradigm", "publication": null, "published_at": "2026-05-14", "record_sha256": "ab4afe11528ac80cb00d4bf5479c3142c411b0c20917f5579f6ef0e8d7a5d61f", "relations": [], "resource_id": "arxiv:2605.14290", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2605.14290v1", "topics": ["planning", "safety-security", "tool-integration"], "updated_at": "2026-05-14"}, "text": "Web Agents Should Adopt the Plan-Then-Execute Paradigm\nResource type: paper\nReview status: curated\nScope: adjacent\n\nArgues for compiling web tasks into predefined execution graphs before exposing an agent to untrusted runtime content. Its WebArena task analysis identifies typed, semantically complete website tools as a prerequisite and distinguishes this interface proposal from reactive click-level control.\nPlatforms: web\nTopics: planning, safety-security, tool-integration\nPublished: 2026-05-14\nAuthors: Julien Piet, Annabella Chow, Yiwei Hou, Muxi Lyu, Sylvie Venuto, Jinhao Zhu, Raluca Ada Popa, David Wagner\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. Directly analyzes browser-agent prompt-injection boundaries and WebArena task structure, but advocates typed website operations rather than demonstrating a new GUI-control policy.\n\nPrimary source: https://arxiv.org/abs/2605.14290v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-14290/"}
{"id": "arxiv:2605.15542", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd85c0d0a8b6bc5c687bc301b1b84d6a019c5713012890e0a23c08ae2d83a1e6", "submitted_at": "2026-05-15T02:27:41Z", "updated_at": "2026-05-15T02:27:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.15542v1"}, "authors": ["Yichao Liu", "Huawen Shen", "Liu Yu", "Shiyu Liu", "Zeyu Chen", "Yu Zhou"], "canonical_source_url": "https://arxiv.org/abs/2605.15542", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-15542/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd85c0d0a8b6bc5c687bc301b1b84d6a019c5713012890e0a23c08ae2d83a1e6", "submitted_at": "2026-05-15T02:27:41Z", "updated_at": "2026-05-15T02:27:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.15542v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.15542"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.15542"}], "name": "DRS-GUI: Dynamic Region Search for Training-Free GUI Grounding", "publication": null, "published_at": "2026-05-15", "record_sha256": "afce34c1f8cfbf1fa3fdb6ecd722c1b67aee77261991ae3fd97de2f91dbd569a", "relations": [], "resource_id": "arxiv:2605.15542", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.15542v1", "topics": ["grounding", "training-free", "planning", "screen-understanding"], "updated_at": "2026-05-15"}, "text": "DRS-GUI: Dynamic Region Search for Training-Free GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nDRS-GUI searches screenshot regions through focus, shift and scatter operations selected by a tree-search planner. A region-quality reward guides this training-free perceptual search before final GUI grounding.\nPlatforms: unspecified\nTopics: grounding, training-free, planning, screen-understanding\nPublished: 2026-05-15\nAuthors: Yichao Liu, Huawen Shen, Liu Yu, Shiyu Liu, Zeyu Chen, Yu Zhou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.15542v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-15542/"}
{"id": "arxiv:2605.15777", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c4ee2439dd31a375b6f64cbc7f1d0ff75c87e1346b1627a53db145ae42682f8a", "submitted_at": "2026-05-15T09:35:15Z", "updated_at": "2026-05-24T16:35:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.15777v2"}, "authors": ["Kean Shi", "Zihang Li", "Tianyi Ma", "Zengji Tu", "Jialong Wu", "Xinbo Xu", "Qingyao Yang", "Ruoyu Wu", "Weichu Xie", "Ming Wu", "Jason Zeng", "Michael Heinrich", "Elvis Zhang", "Liang Chen", "Kuan Li", "Baobao Chang"], "canonical_source_url": "https://arxiv.org/abs/2605.15777", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-15777/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c4ee2439dd31a375b6f64cbc7f1d0ff75c87e1346b1627a53db145ae42682f8a", "submitted_at": "2026-05-15T09:35:15Z", "updated_at": "2026-05-24T16:35:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.15777v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "topics"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "8cfab6fca80090adb47c4ef634f89eedd49a6204dbdc2b5b8e0d6f04c32db289", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md"}], "identifiers": {"arxiv": "2605.15777"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.15777"}], "name": "SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?", "publication": null, "published_at": "2026-05-15", "record_sha256": "c6c78bab164a590983b6f91919c778575040527f8c94ad2396450642e3d99efe", "relations": [], "resource_id": "arxiv:2605.15777", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.15777v2", "topics": ["web-navigation", "evaluation", "long-horizon", "workflow", "cross-app", "state-tracking"], "updated_at": "2026-05-24"}, "text": "SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows?\nResource type: paper\nReview status: curated\nScope: core\n\nSaaS-Bench evaluates browser agents on professional workflows across self-hosted applications. Weighted checkpoints measure task completion and partial progress, with application-state checks and configured content judges documented in the official harness.\nPlatforms: web\nTopics: web-navigation, evaluation, long-horizon, workflow, cross-app, state-tracking\nPublished: 2026-05-15\nAuthors: Kean Shi, Zihang Li, Tianyi Ma, Zengji Tu, Jialong Wu, Xinbo Xu, Qingyao Yang, Ruoyu Wu, Weichu Xie, Ming Wu, Jason Zeng, Michael Heinrich, Elvis Zhang, Liang Chen, Kuan Li, Baobao Chang\nRecorded arXiv version: v2\n\nThe complete retained official abstract and cached official README were reviewed. The README establishes browser execution against running SaaS applications; no repository code was executed and benchmark results were not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.15777v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-15777/"}
{"id": "arxiv:2605.16402", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "671f33dd87a7ebf9233494ef6fa0efd9c9772a3951f3c218bf4de822825f7af2", "submitted_at": "2026-05-13T02:48:52Z", "updated_at": "2026-05-13T02:48:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16402v1"}, "authors": ["Haoren Zhao", "Tianyi Chen", "Zhen Wang"], "canonical_source_url": "https://arxiv.org/abs/2605.16402", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-16402/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "671f33dd87a7ebf9233494ef6fa0efd9c9772a3951f3c218bf4de822825f7af2", "submitted_at": "2026-05-13T02:48:52Z", "updated_at": "2026-05-13T02:48:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16402v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.16402"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16402"}], "name": "WinDeskGround: A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments", "publication": null, "published_at": "2026-05-13", "record_sha256": "2672e1aad43c2328210792a899cc0d50f6981cc58de3eddc5ea53720a22be3c7", "relations": [], "resource_id": "arxiv:2605.16402", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.16402v1", "topics": ["grounding", "evaluation", "robustness", "data-generation"], "updated_at": "2026-05-13"}, "text": "WinDeskGround: A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments\nResource type: paper\nReview status: curated\nScope: core\n\nWinDeskGround synthesizes multi-window desktop layouts with controlled occlusion, density and semantic similarity to test GUI grounding robustness. Its instruction-target benchmark measures how localization changes when interfaces depart from simple single-layer scenes.\nPlatforms: desktop\nTopics: grounding, evaluation, robustness, data-generation\nPublished: 2026-05-13\nAuthors: Haoren Zhao, Tianyi Chen, Zhen Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.16402v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-16402/"}
{"id": "arxiv:2605.16883", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52c1bf4c614ac05a55f2c2c66d3ac4eeb1bf31743814a45eb9f0ce064b6baaab", "submitted_at": "2026-05-16T08:51:57Z", "updated_at": "2026-05-16T08:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16883v1"}, "authors": ["Shilong Jin", "Lanjun Wang", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2605.16883", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-16883/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52c1bf4c614ac05a55f2c2c66d3ac4eeb1bf31743814a45eb9f0ce064b6baaab", "submitted_at": "2026-05-16T08:51:57Z", "updated_at": "2026-05-16T08:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.16883v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.16883"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.16883"}], "name": "SE-GA: Memory-Augmented Self-Evolution for GUI Agents", "publication": null, "published_at": "2026-05-16", "record_sha256": "6c6ad0042e00e431fc6b7c671009314ec04095ce06667c02edde0e3522227a8a", "relations": [], "resource_id": "arxiv:2605.16883", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.16883v1", "topics": ["memory", "interaction-history", "mobile-navigation", "grounding", "evaluation"], "updated_at": "2026-05-16"}, "text": "SE-GA: Memory-Augmented Self-Evolution for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSE-GA retrieves hierarchical memories during inference and uses the collected experience in a separate policy-improvement pipeline. Its evaluation includes offline grounding/action tasks and online Android interaction, which measure different capabilities.\nPlatforms: android\nTopics: memory, interaction-history, mobile-navigation, grounding, evaluation\nPublished: 2026-05-16\nAuthors: Shilong Jin, Lanjun Wang, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.16883v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-16883/"}
{"id": "arxiv:2605.17320", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51576fb9bde23f769d3200a0d593ad610eb27248917132aff1bcf6eb7ff03cea", "submitted_at": "2026-05-17T08:24:01Z", "updated_at": "2026-05-17T08:24:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17320v1"}, "authors": ["Yutong Huang", "Vikranth Srivatsa", "Alex Asch", "Hansin Tushar Patwa", "Yiying Zhang"], "canonical_source_url": "https://arxiv.org/abs/2605.17320", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-17320/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51576fb9bde23f769d3200a0d593ad610eb27248917132aff1bcf6eb7ff03cea", "submitted_at": "2026-05-17T08:24:01Z", "updated_at": "2026-05-17T08:24:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17320v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.17320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.17320"}], "name": "TClone: Low-Latency Forking of Live GUI Environments for Computer-Use Agents", "publication": null, "published_at": "2026-05-17", "record_sha256": "b928fa01ef1003bfc977ed43f6d360eee804487d6f7bbcb61349b8bde1c8aa55", "relations": [], "resource_id": "arxiv:2605.17320", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.17320v1", "topics": ["sandbox", "efficiency", "safety-security", "automation"], "updated_at": "2026-05-17"}, "text": "TClone: Low-Latency Forking of Live GUI Environments for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTClone versions live GUI workspaces through snapshots, isolated forks, rollback and selective commit or merge. It separates fast branch creation from durable checkpoints using copy-on-write sharing, filesystem versioning and asynchronous checkpointing, with end-to-end agent-loop latency measurements.\nPlatforms: unspecified\nTopics: sandbox, efficiency, safety-security, automation\nPublished: 2026-05-17\nAuthors: Yutong Huang, Vikranth Srivatsa, Alex Asch, Hansin Tushar Patwa, Yiying Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.17320v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-17320/"}
{"id": "arxiv:2605.17656", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff95e0b071508592190ba236741c1c01200dd38fdc74c31b2cc4d1bf8143ae9c", "submitted_at": "2026-05-17T21:14:32Z", "updated_at": "2026-05-17T21:14:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17656v1"}, "authors": ["Athar Parvez", "Muhammad Jawad Mufti", "Muqaddas Gull", "Omar Hammad"], "canonical_source_url": "https://arxiv.org/abs/2605.17656", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-17656/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff95e0b071508592190ba236741c1c01200dd38fdc74c31b2cc4d1bf8143ae9c", "submitted_at": "2026-05-17T21:14:32Z", "updated_at": "2026-05-17T21:14:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.17656v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "identifiers": {"arxiv": "2605.17656"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.17656"}], "name": "MUIAnno: An Expert-Annotated Dataset and Evaluation Benchmark for Mobile UI Understanding", "publication": null, "published_at": "2026-05-17", "record_sha256": "44042d6ce792b1b0303ffbe19545f420bea8647be0bc4076065e1645b373c6f8", "relations": [], "resource_id": "arxiv:2605.17656", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.17656v1", "topics": ["screen-understanding", "grounding", "evaluation", "data-generation"], "updated_at": "2026-05-17"}, "text": "MUIAnno: An Expert-Annotated Dataset and Evaluation Benchmark for Mobile UI Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nMUIAnno provides expert labels for elements in manually explored mobile-app screens, with a web annotation tool that emits structured JSON. It also supplies UI-element detection baselines for the dataset.\nPlatforms: unspecified\nTopics: screen-understanding, grounding, evaluation, data-generation\nPublished: 2026-05-17\nAuthors: Athar Parvez, Muhammad Jawad Mufti, Muqaddas Gull, Omar Hammad\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2605.17656v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-17656/"}
{"id": "arxiv:2605.18048", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "017909a697c6c661d3486d960042f4d691f70a660a60083e70512de34690cb51", "submitted_at": "2026-05-18T08:36:04Z", "updated_at": "2026-05-18T08:36:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18048v1"}, "authors": ["Jingjing Liu", "Ziye Huang", "Zihao Cheng", "Zeming Liu", "Jiahong Wu", "Yuhang Guo", "Kehai Chen", "Yunhong Wang", "Haifeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2605.18048", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-18048/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "017909a697c6c661d3486d960042f4d691f70a660a60083e70512de34690cb51", "submitted_at": "2026-05-18T08:36:04Z", "updated_at": "2026-05-18T08:36:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18048v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.18048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18048"}], "name": "DocOS: Towards Proactive Document-Guided Actions in GUI Agents", "publication": null, "published_at": "2026-05-18", "record_sha256": "45cc06cbff5a0e6c27a672a4520ad3f4dd3223c32b8a15c4fe5ca63b2090ca90", "relations": [], "resource_id": "arxiv:2605.18048", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.18048v1", "topics": ["web-navigation", "grounding", "evaluation", "planning"], "updated_at": "2026-05-18"}, "text": "DocOS: Towards Proactive Document-Guided Actions in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDocOS benchmarks agents that search online documentation and translate procedural instructions into GUI actions in interactive web environments. It separates failures to retrieve useful procedures from failures to ground retrieved instructions into executable actions.\nPlatforms: web\nTopics: web-navigation, grounding, evaluation, planning\nPublished: 2026-05-18\nAuthors: Jingjing Liu, Ziye Huang, Zihao Cheng, Zeming Liu, Jiahong Wu, Yuhang Guo, Kehai Chen, Yunhong Wang, Haifeng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.18048v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-18048/"}
{"id": "arxiv:2605.18652", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d96c8844274f79adea167169a0853e3a7cd82fee4a8a666913f97898a0836e7c", "submitted_at": "2026-05-18T16:57:36Z", "updated_at": "2026-05-18T16:57:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18652v1"}, "authors": ["Ziyun Zeng", "Hang Hua", "Bocheng Zou", "Mu Cai", "Rogerio Feris", "Jiebo Luo"], "canonical_source_url": "https://arxiv.org/abs/2605.18652", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-18652/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d96c8844274f79adea167169a0853e3a7cd82fee4a8a666913f97898a0836e7c", "submitted_at": "2026-05-18T16:57:36Z", "updated_at": "2026-05-18T16:57:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.18652v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.18652"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18652"}], "name": "MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents", "publication": null, "published_at": "2026-05-18", "record_sha256": "cbcdc53f80c54d943ffaf145d84ff88d9226b23732732df8dc2c3179aabc68b6", "relations": [], "resource_id": "arxiv:2605.18652", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.18652v1", "topics": ["memory", "interaction-history", "long-horizon", "data-generation", "evaluation"], "updated_at": "2026-05-18"}, "text": "MementoGUI: Learning Agentic Multimodal Memory Control for Long-Horizon GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMementoGUI learns a controller for selecting, compressing and retrieving textual and region-level visual memories while leaving the GUI-agent backbone fixed. It also converts computer-use traces into controller training data and evaluates semantic actions, progress and memory consistency on long-horizon tasks.\nPlatforms: web\nTopics: memory, interaction-history, long-horizon, data-generation, evaluation\nPublished: 2026-05-18\nAuthors: Ziyun Zeng, Hang Hua, Bocheng Zou, Mu Cai, Rogerio Feris, Jiebo Luo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.18652v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-18652/"}
{"id": "arxiv:2605.18758", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99e8e0039ec99ef8f6a6ef3900cb74a62ddd5ae4f8df30629110969290028e6a", "submitted_at": "2026-04-03T08:57:06Z", "updated_at": "2026-09-28T02:42:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.18758v2"}, "authors": ["Felix Henry", "Lihan Lin", "Jiangyou Zhu", "Yangfan", "Bingqian Zhang", "Min Chen", "Shiyu Huang"], "canonical_source_url": "https://arxiv.org/abs/2605.18758", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-18758/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "99e8e0039ec99ef8f6a6ef3900cb74a62ddd5ae4f8df30629110969290028e6a", "submitted_at": "2026-04-03T08:57:06Z", "updated_at": "2026-09-28T02:42:06Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.18758v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:39Z", "role": "primary", "sha256": "27c33576ed43c7bc63d4b290f6558553bd1d3aa112a7358138f4045e1bc14918", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.18758"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.18758"}], "name": "OmniGUI: Benchmarking GUI Agents in Omni-Modal Smartphone Environments", "publication": null, "published_at": "2026-04-03", "record_sha256": "582bac0df8bdebc81d003d69dfe87392dd23cf4404355a7d6dc7513d2b4de84d", "relations": [], "resource_id": "arxiv:2605.18758", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.18758v2", "topics": ["evaluation", "action-prediction", "video", "mobile-navigation"], "updated_at": "2026-09-28"}, "text": "OmniGUI: Benchmarking GUI Agents in Omni-Modal Smartphone Environments\nResource type: paper\nReview status: curated\nScope: core\n\nOmniGUI evaluates step-level action prediction from interleaved smartphone screenshots, synchronous audio and video. Expert demonstrations carry multimodal-dependency annotations, and initial baselines use native multimodal models as agent proxies to study temporal and audio-related failures.\nPlatforms: unspecified\nTopics: evaluation, action-prediction, video, mobile-navigation\nPublished: 2026-04-03\nAuthors: Felix Henry, Lihan Lin, Jiangyou Zhu, Yangfan, Bingqian Zhang, Min Chen, Shiyu Huang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.18758v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-18758/"}
{"id": "arxiv:2605.19769", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a93d880793018d1b47a39e0487e3be76550fd2b49de9535b2913e9f08159074", "submitted_at": "2026-05-19T12:40:29Z", "updated_at": "2026-05-19T12:40:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19769v1"}, "authors": ["Jinbiao Wei", "Qianran Ma", "Yilun Zhao", "Xiao Zhou", "Kangqi Ni", "Guo Gan", "Arman Cohan"], "canonical_source_url": "https://arxiv.org/abs/2605.19769", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-19769/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a93d880793018d1b47a39e0487e3be76550fd2b49de9535b2913e9f08159074", "submitted_at": "2026-05-19T12:40:29Z", "updated_at": "2026-05-19T12:40:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.19769v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.19769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.19769"}], "name": "OpenComputer: Verifiable Software Worlds for Computer-Use Agents", "publication": null, "published_at": "2026-05-19", "record_sha256": "a2aea10082e355274489f6146bab38b12e534a21018abddaa538015f1c3881f1", "relations": [], "resource_id": "arxiv:2605.19769", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.19769v1", "topics": ["evaluation", "data-generation", "observability"], "updated_at": "2026-05-19"}, "text": "OpenComputer: Verifiable Software Worlds for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OpenComputer, which builds desktop tasks around application-state verifiers and refines those verifiers using execution feedback. An evaluation harness records trajectories and assigns auditable partial credit across professional software workflows.\nPlatforms: desktop\nTopics: evaluation, data-generation, observability\nPublished: 2026-05-19\nAuthors: Jinbiao Wei, Qianran Ma, Yilun Zhao, Xiao Zhou, Kangqi Ni, Guo Gan, Arman Cohan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.19769v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-19769/"}
{"id": "arxiv:2605.20291", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6e11bae4d952d79276543ced6f81193b36868a6222fe51ee7b0816aa72fbab", "submitted_at": "2026-05-19T09:19:01Z", "updated_at": "2026-05-26T02:53:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.20291v2"}, "authors": ["Fatemeh Pesaran Zadeh", "Seyeon Choi", "Xing Han Lù", "Siva Reddy", "Gunhee Kim"], "canonical_source_url": "https://arxiv.org/abs/2605.20291", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-20291/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa6e11bae4d952d79276543ced6f81193b36868a6222fe51ee7b0816aa72fbab", "submitted_at": "2026-05-19T09:19:01Z", "updated_at": "2026-05-26T02:53:26Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.20291v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.20291"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.20291"}], "name": "Weasel: Out-of-Domain Generalization for Web Agents via Importance-Diversity Data Selection", "publication": null, "published_at": "2026-05-19", "record_sha256": "6dd5ef3b22a9ecaa8d97924857c6fe5a874734e9ba0690876d71b3b6d57f0a2d", "relations": [], "resource_id": "arxiv:2605.20291", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.20291v2", "topics": ["web-navigation", "imitation-learning", "efficiency", "robustness", "data-generation"], "updated_at": "2026-05-26"}, "text": "Weasel: Out-of-Domain Generalization for Web Agents via Importance-Diversity Data Selection\nResource type: paper\nReview status: curated\nScope: core\n\nWeasel selects web-agent training steps by balancing importance and diversity across states, sites and interaction patterns. It also prunes training accessibility trees around known action targets and regenerates rationales to match the model’s reasoning style.\nPlatforms: web\nTopics: web-navigation, imitation-learning, efficiency, robustness, data-generation\nPublished: 2026-05-19\nAuthors: Fatemeh Pesaran Zadeh, Seyeon Choi, Xing Han Lù, Siva Reddy, Gunhee Kim\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.20291v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-20291/"}
{"id": "arxiv:2605.21470", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b73bf7776e280a9040b1d7d1e29957913723253014c1dfe84f64916b2a08e5e", "submitted_at": "2026-05-20T17:54:27Z", "updated_at": "2026-05-29T17:54:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.21470v2"}, "authors": ["Caleb Winston", "Ron Yifeng Wang", "Azalia Mirhoseini", "Christos Kozyrakis"], "canonical_source_url": "https://arxiv.org/abs/2605.21470", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-21470/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9b73bf7776e280a9040b1d7d1e29957913723253014c1dfe84f64916b2a08e5e", "submitted_at": "2026-05-20T17:54:27Z", "updated_at": "2026-05-29T17:54:05Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.21470v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.21470"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.21470"}], "name": "Agent JIT Compilation for Latency-Optimizing Web Agent Planning and Scheduling", "publication": null, "published_at": "2026-05-20", "record_sha256": "d6bfd949f58e31a61b3be4de2efd2348d87fff4cf9a68d99d17e06c6f78017fb", "relations": [], "resource_id": "arxiv:2605.21470", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.21470v2", "topics": ["planning", "efficiency", "tool-integration", "web-navigation"], "updated_at": "2026-05-29"}, "text": "Agent JIT Compilation for Latency-Optimizing Web Agent Planning and Scheduling\nResource type: paper\nReview status: curated\nScope: core\n\nAgent JIT compilation turns browser task descriptions into validated executable plans containing model calls, tools and parallel work. A cost-based scheduler and pre/postcondition protocol target latency and incorrect tool use in web-agent execution.\nPlatforms: web\nTopics: planning, efficiency, tool-integration, web-navigation\nPublished: 2026-05-20\nAuthors: Caleb Winston, Ron Yifeng Wang, Azalia Mirhoseini, Christos Kozyrakis\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.21470v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-21470/"}
{"id": "arxiv:2605.24785", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e64d36e920a33851aacf93138a2081d0470424dfec6a8fd11915928660824851", "submitted_at": "2026-05-24T00:07:25Z", "updated_at": "2026-05-26T04:25:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.24785v2"}, "authors": ["Yubo Li", "Yidi Miao", "Yuntian Shen", "Yuxin Liu"], "canonical_source_url": "https://arxiv.org/abs/2605.24785", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-24785/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e64d36e920a33851aacf93138a2081d0470424dfec6a8fd11915928660824851", "submitted_at": "2026-05-24T00:07:25Z", "updated_at": "2026-05-26T04:25:32Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.24785v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2605.24785"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.24785"}], "name": "PANDO: Efficient Multimodal AI Agents via Online Skill Distillation", "publication": null, "published_at": "2026-05-24", "record_sha256": "da897226a16cbae7d427c5ea609fa47484589c81a76740fe07dc4cb5ab052b03", "relations": [], "resource_id": "arxiv:2605.24785", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.24785v2", "topics": ["efficiency", "memory", "web-navigation"], "updated_at": "2026-05-26"}, "text": "PANDO: Efficient Multimodal AI Agents via Online Skill Distillation\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes inefficiency in VisualWebArena trajectories and distills reusable online skills to reduce repeated actions and token overhead in multimodal web agents.\nPlatforms: web\nTopics: efficiency, memory, web-navigation\nPublished: 2026-05-24\nAuthors: Yubo Li, Yidi Miao, Yuntian Shen, Yuxin Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.24785v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-24785/"}
{"id": "arxiv:2605.25160", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb40c8bf005d56139d263e1b5dc4553ddcd68522124686b7f0865e6d8d1725ad", "submitted_at": "2026-05-24T16:33:14Z", "updated_at": "2026-06-18T02:24:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25160v2"}, "authors": ["Guohong Liu", "Jialei Ye", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Yunxin Liu", "Yuanchun Li"], "canonical_source_url": "https://arxiv.org/abs/2605.25160", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-25160/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "eb40c8bf005d56139d263e1b5dc4553ddcd68522124686b7f0865e6d8d1725ad", "submitted_at": "2026-05-24T16:33:14Z", "updated_at": "2026-06-18T02:24:07Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25160v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.25160"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.25160"}], "name": "ScaleWoB: Guiding GUI Agents with Coding Agents via Large-Scale Environmental Synthesis", "publication": null, "published_at": "2026-05-24", "record_sha256": "a0a0e79b5f4849a090efbaea057b092cbff8e33d6c7bf151063e44539064611f", "relations": [], "resource_id": "arxiv:2605.25160", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.25160v2", "topics": ["data-generation", "evaluation", "sandbox"], "updated_at": "2026-06-18"}, "text": "ScaleWoB: Guiding GUI Agents with Coding Agents via Large-Scale Environmental Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ScaleWoB, which synthesizes interactive GUI environments as backend-free webpages with verifiable rewards. Its mobile-oriented benchmark evaluates agent execution and compares selected synthetic tasks with real-app behavior.\nPlatforms: web\nTopics: data-generation, evaluation, sandbox\nPublished: 2026-05-24\nAuthors: Guohong Liu, Jialei Ye, Pengzhi Gao, Wei Liu, Jian Luan, Yunxin Liu, Yuanchun Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. Mobile, desktop and automotive interfaces are simulated as webpages; this does not assert native deployment on those platforms.\n\nPrimary source: https://arxiv.org/abs/2605.25160v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-25160/"}
{"id": "arxiv:2605.25624", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecc0712f55ee90c237f9a48e013e396b1082f932a485c22d8249d1da7acbf307", "submitted_at": "2026-05-25T09:28:03Z", "updated_at": "2026-06-08T08:32:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25624v2"}, "authors": ["Bowen Wang", "Dunjie Lu", "Junli Wang", "Tianyi Bai", "Shixuan Liu", "Zhipeng Zhang", "Haiquan Wang", "Hao Hu", "Tianbao Xie", "Shuai Bai", "Dayiheng Liu", "Que Shen", "Junyang Lin", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2605.25624", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-25624/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecc0712f55ee90c237f9a48e013e396b1082f932a485c22d8249d1da7acbf307", "submitted_at": "2026-05-25T09:28:03Z", "updated_at": "2026-06-08T08:32:35Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.25624v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.25624"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.25624"}], "name": "CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents", "publication": null, "published_at": "2026-05-25", "record_sha256": "53a7c7a9aba824d84c6c5bea606ee37c1076beffb8bd4dbb42c933e8464aaa53", "relations": [], "resource_id": "arxiv:2605.25624", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.25624v2", "topics": ["data-generation", "reinforcement-learning", "sandbox"], "updated_at": "2026-06-08"}, "text": "CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents CUA-Gym, which co-generates instructions, initial and target states, and executable reward checks through generator-discriminator iterations. Mock web applications expand training environments, with learned agents evaluated on OSWorld-Verified and WebArena.\nPlatforms: web, desktop\nTopics: data-generation, reinforcement-learning, sandbox\nPublished: 2026-05-25\nAuthors: Bowen Wang, Dunjie Lu, Junli Wang, Tianyi Bai, Shixuan Liu, Zhipeng Zhang, Haiquan Wang, Hao Hu, Tianbao Xie, Shuai Bai, Dayiheng Liu, Que Shen, Junyang Lin, Tao Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The abstract promises future open sourcing; this annotation does not independently verify artifact release or licensing.\n\nPrimary source: https://arxiv.org/abs/2605.25624v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-25624/"}
{"id": "arxiv:2605.26086", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27999553928305cfa4223011feb3f9c2decda8a861e5800865537a81c7c08832", "submitted_at": "2026-05-25T17:50:04Z", "updated_at": "2026-05-25T17:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26086v1"}, "authors": ["Yusong Lin", "Xinyuan Liang", "Haiyang Wang", "Qipeng Gu", "Siqi Cheng", "Jiangui Chen", "Shuzhe Wu", "Feiyang Pan", "Lue Fan", "Sanyuan Zhao", "Dandan Tu"], "canonical_source_url": "https://arxiv.org/abs/2605.26086", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-26086/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27999553928305cfa4223011feb3f9c2decda8a861e5800865537a81c7c08832", "submitted_at": "2026-05-25T17:50:04Z", "updated_at": "2026-05-25T17:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26086v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "identifiers": {"arxiv": "2605.26086"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26086"}], "name": "Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World", "publication": null, "published_at": "2026-05-25", "record_sha256": "ce782e16fc3852d260441052d8f128b4e78f0ab3f3e26a6295acb270254cd194", "relations": [], "resource_id": "arxiv:2605.26086", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.26086v1", "topics": ["evaluation", "interaction-history", "long-horizon", "robustness", "data-generation", "workflow"], "updated_at": "2026-05-25"}, "text": "Claw-Anything: Benchmarking Always-On Personal Assistants with Broader Access to User's Digital World\nResource type: paper\nReview status: curated\nScope: core\n\nClaw-Anything benchmarks persistent personal assistance across activity histories, backend services and integrated GUI/CLI interaction. Its simulated event histories introduce irrelevant or conflicting context, and a separate generation pipeline supplies training environments.\nPlatforms: unspecified\nTopics: evaluation, interaction-history, long-horizon, robustness, data-generation, workflow\nPublished: 2026-05-25\nAuthors: Yusong Lin, Xinyuan Liang, Haiyang Wang, Qipeng Gu, Siqi Cheng, Jiangui Chen, Shuzhe Wu, Feiyang Pan, Lue Fan, Sanyuan Zhao, Dandan Tu\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2605.26086v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-26086/"}
{"id": "arxiv:2605.26114", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "145c7012fb15a85b203924ecd559c8b20d18b51bf79805a6751fec75d6f68dec", "submitted_at": "2026-05-25T17:59:49Z", "updated_at": "2026-09-23T17:59:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.26114v3"}, "authors": ["Dingbang Wu", "Rui Hao", "Haiyang Wang", "Shuzhe Wu", "Han Xiao", "Zhenghong Li", "Bojiang Zhou", "Zheng Ju", "Zichen Liu", "Lue Fan", "Zhaoxiang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2605.26114", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-26114/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "145c7012fb15a85b203924ecd559c8b20d18b51bf79805a6751fec75d6f68dec", "submitted_at": "2026-05-25T17:59:49Z", "updated_at": "2026-09-23T17:59:57Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.26114v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.26114"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26114"}], "name": "MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research", "publication": null, "published_at": "2026-05-25", "record_sha256": "26172989b7f02fe0b2d545716e7aedeefb718b1583c261fdfc11d4c50fcf8388", "relations": [], "resource_id": "arxiv:2605.26114", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.26114v3", "topics": ["sandbox", "evaluation", "reinforcement-learning"], "updated_at": "2026-09-23"}, "text": "MobileGym: A Verifiable and Highly Parallel Simulation Platform for Mobile GUI Agent Research\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MobileGym, a browser-hosted simulation of mobile app interactions whose configurable JSON state supports deterministic judging and parallel rollouts. MobileGym-Bench adds parameterized tasks, with experiments examining online training and transfer to real devices.\nPlatforms: web\nTopics: sandbox, evaluation, reinforcement-learning\nPublished: 2026-05-25\nAuthors: Dingbang Wu, Rui Hao, Haiyang Wang, Shuzhe Wu, Han Xiao, Zhenghong Li, Bojiang Zhou, Zheng Ju, Zichen Liu, Lue Fan, Zhaoxiang Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced. The simulator is browser-hosted; the abstract does not identify a native mobile operating system.\n\nPrimary source: https://arxiv.org/abs/2605.26114v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-26114/"}
{"id": "arxiv:2605.26144", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64f8e09fc62f080a75aec1928b82187063cfc3ad8467290608347622410b6460", "submitted_at": "2026-05-22T20:29:12Z", "updated_at": "2026-09-29T00:12:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2605.26144v4"}, "authors": ["JunJia Guo", "Yuhang Yao", "Chang Li", "Zimeng Pan", "Jiawei", "Zhou", "Jingdi Chen"], "canonical_source_url": "https://arxiv.org/abs/2605.26144", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-26144/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64f8e09fc62f080a75aec1928b82187063cfc3ad8467290608347622410b6460", "submitted_at": "2026-05-22T20:29:12Z", "updated_at": "2026-09-29T00:12:17Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2605.26144v4"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "identifiers": {"arxiv": "2605.26144"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26144"}], "name": "VISTA: An End-to-End Benchmark for Visual Spec-to-Web-App Coding Agents", "publication": null, "published_at": "2026-05-22", "record_sha256": "e1c025d856f279a01a347fbfd5053a239f6c19752a23bce94edcd9dad2e51ca1", "relations": [], "resource_id": "arxiv:2605.26144", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2605.26144v4", "topics": ["evaluation", "ui-design", "automation"], "updated_at": "2026-09-29"}, "text": "VISTA: An End-to-End Benchmark for Visual Spec-to-Web-App Coding Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nVISTA evaluates coding agents that turn design handoffs into working web and Android applications. Its evaluator locates annotated interactive components and probes their behavior, keeping visual localization distinct from functional correctness.\nPlatforms: web, android\nTopics: evaluation, ui-design, automation\nPublished: 2026-05-22\nAuthors: JunJia Guo, Yuhang Yao, Chang Li, Zimeng Pan, Jiawei, Zhou, Jingdi Chen\nRecorded arXiv version: v4\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent GUI-application generation benchmark with real component interaction probes; its main evaluated agents are application coders, not general GUI task executors.\n\nPrimary source: https://arxiv.org/abs/2605.26144v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-26144/"}
{"id": "arxiv:2605.26546", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97467a9a1d3c456a0078e96ebb128efa43412a3139a7a68beb4e6dd7b5f0674d", "submitted_at": "2026-05-26T04:53:53Z", "updated_at": "2026-05-26T04:53:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26546v1"}, "authors": ["Runxi Huang", "Liyu Zhang", "Shengzhong Liu", "Xiaomin Ouyang"], "canonical_source_url": "https://arxiv.org/abs/2605.26546", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-26546/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97467a9a1d3c456a0078e96ebb128efa43412a3139a7a68beb4e6dd7b5f0674d", "submitted_at": "2026-05-26T04:53:53Z", "updated_at": "2026-05-26T04:53:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.26546v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.26546"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.26546"}], "name": "MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration", "publication": null, "published_at": "2026-05-26", "record_sha256": "5e5b1638fb5766bda05153cbf1d51b480fb6f9d664fc6cfc9bd6460ee512dee3", "relations": [], "resource_id": "arxiv:2605.26546", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.26546v1", "topics": ["mobile-navigation", "efficiency", "memory", "robustness"], "updated_at": "2026-05-26"}, "text": "MobileExplorer: Accelerating On-Device Inference for Mobile GUI Agents via Online Exploration\nResource type: paper\nReview status: curated\nScope: core\n\nMobileExplorer probes relevant UI elements while an on-device model is reasoning, then supplies the exploration trace as compact context. A two-level rollback mechanism restores the working state when simple backtracking fails.\nPlatforms: android\nTopics: mobile-navigation, efficiency, memory, robustness\nPublished: 2026-05-26\nAuthors: Runxi Huang, Liyu Zhang, Shengzhong Liu, Xiaomin Ouyang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.26546v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-26546/"}
{"id": "arxiv:2605.27761", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c89b8b5ef5aa17fe747201285104544648449804cc2229a77577d217763f497", "submitted_at": "2026-05-26T23:19:42Z", "updated_at": "2026-05-26T23:19:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27761v1"}, "authors": ["Yifan Sui", "Xin Huang", "Hongbing Li", "Fang Xu", "Jiahe Lv", "Haolong Yan", "Yeqing Shen", "Litao Liu", "Zhimin Fan", "Ziyang Meng", "Jia Wang", "Junbo Qi", "Kaijun Tan", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang", "Osamu Yoshie"], "canonical_source_url": "https://arxiv.org/abs/2605.27761", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-27761/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0c89b8b5ef5aa17fe747201285104544648449804cc2229a77577d217763f497", "submitted_at": "2026-05-26T23:19:42Z", "updated_at": "2026-05-26T23:19:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.27761v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.27761"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.27761"}], "name": "AndroidDaily: A Verifiable Benchmark for Mobile GUI Agents on Real-World Closed-Source Applications", "publication": null, "published_at": "2026-05-26", "record_sha256": "99dbe09a5e30f98e64edf1242ab5f5fa8ee580714a3df887f59819a52ecb498b", "relations": [], "resource_id": "arxiv:2605.27761", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.27761v1", "topics": ["mobile-navigation", "evaluation", "long-horizon", "robustness"], "updated_at": "2026-05-26"}, "text": "AndroidDaily: A Verifiable Benchmark for Mobile GUI Agents on Real-World Closed-Source Applications\nResource type: paper\nReview status: curated\nScope: core\n\nAndroidDaily evaluates mobile GUI agents on daily workflows in closed-source Android apps. Its GRADE evaluator follows visual trajectories against operational obligations, output-quality requirements and negative constraints, producing process-level diagnostics without hidden app state.\nPlatforms: android\nTopics: mobile-navigation, evaluation, long-horizon, robustness\nPublished: 2026-05-26\nAuthors: Yifan Sui, Xin Huang, Hongbing Li, Fang Xu, Jiahe Lv, Haolong Yan, Yeqing Shen, Litao Liu, Zhimin Fan, Ziyang Meng, Jia Wang, Junbo Qi, Kaijun Tan, Zheng Ge, Xiangyu Zhang, Daxin Jiang, Osamu Yoshie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.27761v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-27761/"}
{"id": "arxiv:2605.28116", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4f39433695b9574b275fc844ca6b10a43baf66aed57048bf841f41bf08d4ac9a", "submitted_at": "2026-05-27T08:09:53Z", "updated_at": "2026-05-27T08:09:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28116v1"}, "authors": ["Ruoqi Guo", "Yi Liu", "Gelei Deng", "Yiheng Xiong", "Yuekang Li", "Ying Zhang", "Leo Yu Zhang", "Lida Zhao", "Ji Jie", "Yuxiao Lu"], "canonical_source_url": "https://arxiv.org/abs/2605.28116", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-28116/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4f39433695b9574b275fc844ca6b10a43baf66aed57048bf841f41bf08d4ac9a", "submitted_at": "2026-05-27T08:09:53Z", "updated_at": "2026-05-27T08:09:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28116v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.28116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28116"}], "name": "MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content", "publication": null, "published_at": "2026-05-27", "record_sha256": "75eb154ab63a6d214fce1c906329e0be66d356a985e1d6d97f305888718e2ccf", "relations": [], "resource_id": "arxiv:2605.28116", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.28116v1", "topics": ["safety-security", "data-generation", "evaluation", "robustness"], "updated_at": "2026-05-27"}, "text": "MIRAGE: Context-Aware Prompt Injection against Mobile GUI Agents via User-Generated Content\nResource type: paper\nReview status: curated\nScope: core\n\nMIRAGE constructs mobile screenshot injections in ordinary user-content regions through localization, contextual payload generation and realism curation. Its evaluation separates visual realism from attack effectiveness against pixel-observing GUI agents.\nPlatforms: unspecified\nTopics: safety-security, data-generation, evaluation, robustness\nPublished: 2026-05-27\nAuthors: Ruoqi Guo, Yi Liu, Gelei Deng, Yiheng Xiong, Yuekang Li, Ying Zhang, Leo Yu Zhang, Lida Zhao, Ji Jie, Yuxiao Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.28116v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-28116/"}
{"id": "arxiv:2605.28258", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e03e0ef06c6610eaefeec08b7891a1ab4d5073c0793650d4845735259978723", "submitted_at": "2026-05-27T10:08:48Z", "updated_at": "2026-10-02T05:26:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28258v2"}, "authors": ["Yixu Huang", "Bo Li", "Na Li", "Zhe Wang", "Kaijie Chen", "Haonan Ge", "Qingyi Si", "Yuanzhe Shen", "Ruihan Yang", "Guangjing Wang", "Hongcheng Guo"], "canonical_source_url": "https://arxiv.org/abs/2605.28258", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-28258/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4e03e0ef06c6610eaefeec08b7891a1ab4d5073c0793650d4845735259978723", "submitted_at": "2026-05-27T10:08:48Z", "updated_at": "2026-10-02T05:26:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28258v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.28258"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28258"}], "name": "GUI Agents for Continual Game Generation", "publication": null, "published_at": "2026-05-27", "record_sha256": "bdab83686508e4e3b34474d9d59d8ec94f7bc0c439eeef2bd48160a58653732e", "relations": [], "resource_id": "arxiv:2605.28258", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2605.28258v2", "topics": ["evaluation", "automation", "workflow"], "updated_at": "2026-10-02"}, "text": "GUI Agents for Continual Game Generation\nResource type: paper\nReview status: curated\nScope: adjacent\n\nPlaytestArena evaluates generated browser games using expected in-play behaviors adjudicated by GUI agents. Play2Code pairs a coding agent with a rubric-blind GUI playtester whose interaction traces and feedback guide iterative refinement, with a separate final judge.\nPlatforms: web\nTopics: evaluation, automation, workflow\nPublished: 2026-05-27\nAuthors: Yixu Huang, Bo Li, Na Li, Zhe Wang, Kaijie Chen, Haonan Ge, Qingyi Si, Yuanzhe Shen, Ruihan Yang, Guangjing Wang, Hongcheng Guo\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. GUI agents actually load and play browser-game builds, providing an executable interface-testing signal for game-code generation; the central application remains game development.\n\nPrimary source: https://arxiv.org/abs/2605.28258v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-28258/"}
{"id": "arxiv:2605.28534", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78a5f5d7df9a8b6f1dbdb5693849c728a4b97e1f25c6e2e24d0c2418628c64e7", "submitted_at": "2026-05-27T14:29:41Z", "updated_at": "2026-05-27T14:29:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28534v1"}, "authors": ["Zheng Wu", "Chengcheng Han", "Zhengxi Lu", "Tianjie Ju", "Yanyu Chen", "Qi Gu", "Xunliang Cai", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2605.28534", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-28534/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78a5f5d7df9a8b6f1dbdb5693849c728a4b97e1f25c6e2e24d0c2418628c64e7", "submitted_at": "2026-05-27T14:29:41Z", "updated_at": "2026-05-27T14:29:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.28534v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:11:36Z", "role": "primary", "sha256": "235f1fff1385719d603377b680d74f6803bbe46d112dd6c008bee6b42f105429", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.28534"}, "kind": "paper", "last_verified_at": "2026-10-17T17:11:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28534"}], "name": "GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-aware Exemplar Reselection", "publication": null, "published_at": "2026-05-27", "record_sha256": "e3ba18877b4e2712f136b77cf97904b82e0d85a8ce6702a85f3dc6efa40eb813", "relations": [], "resource_id": "arxiv:2605.28534", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.28534v1", "topics": ["data-generation", "planning"], "updated_at": "2026-05-27"}, "text": "GUI-CIDER: Mid-training GUI Agents via Causal Internalization and Density-aware Exemplar Reselection\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-CIDER turns GUI trajectories into textual planning and causal interface knowledge, reselects exemplars for causal structure and low redundancy, then mid-trains GUI agents on that corpus.\nPlatforms: unspecified\nTopics: data-generation, planning\nPublished: 2026-05-27\nAuthors: Zheng Wu, Chengcheng Han, Zhengxi Lu, Tianjie Ju, Yanyu Chen, Qi Gu, Xunliang Cai, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.28534v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-28534/"}
{"id": "arxiv:2605.28646", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb9b67f7e51edecfbe57af47b6bad445c05129c11c377d32ea84369b430f895", "submitted_at": "2026-05-27T15:51:22Z", "updated_at": "2026-05-29T04:03:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28646v2"}, "authors": ["Yanqiu Zhao", "Dongying Zheng", "Kaibo Huang", "Yukun Wei", "Zhongliang Yang", "Linna Zhou"], "canonical_source_url": "https://arxiv.org/abs/2605.28646", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-28646/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6eb9b67f7e51edecfbe57af47b6bad445c05129c11c377d32ea84369b430f895", "submitted_at": "2026-05-27T15:51:22Z", "updated_at": "2026-05-29T04:03:46Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2605.28646v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.28646"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.28646"}], "name": "MaskClaw: Edge-Side Personalized Privacy Arbitration for GUI Agents with Behavior-Driven Skill Evolution", "publication": null, "published_at": "2026-05-27", "record_sha256": "b9c3e782eee68b505df7f6c1bda1ac47a7974c59f53ce526f3bc4e7b6b7d3a19", "relations": [], "resource_id": "arxiv:2605.28646", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.28646v2", "topics": ["safety-security", "memory", "human-in-the-loop", "evaluation"], "updated_at": "2026-05-29"}, "text": "MaskClaw: Edge-Side Personalized Privacy Arbitration for GUI Agents with Behavior-Driven Skill Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nMaskClaw makes local Allow, Mask or Ask decisions before GUI screenshots leave a trusted environment, using visual evidence and personalized policy memory. Its benchmark and designed evolution scenarios test how corrections and edits can become privacy skills checked by a sandbox gate.\nPlatforms: unspecified\nTopics: safety-security, memory, human-in-the-loop, evaluation\nPublished: 2026-05-27\nAuthors: Yanqiu Zhao, Dongying Zheng, Kaibo Huang, Yukun Wei, Zhongliang Yang, Linna Zhou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.28646v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-28646/"}
{"id": "arxiv:2605.29486", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a41cb1a93e6aa6ba8fbc7dca55415577ab5880a6592e9b0bee5b12b5a7732cbf", "submitted_at": "2026-05-28T07:14:15Z", "updated_at": "2026-09-28T15:27:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.29486v3"}, "authors": ["Yuxuan Liu", "Xin Lai", "Junyi Li", "Pengyuan Lyu", "Jason", "Yiduo Guo", "Zhengyao Fang", "Yang Ding", "Yi Zhang", "Weinong Wang", "Huawen Shen", "Xingran Zhou", "Liang Wu", "Fei Tang", "Sunqi Fan", "Shangpin Peng", "Zheng Ruan", "Anran Zhang", "Chengquan Zhang", "Han Hu", "Benyou Wang", "Ji-Rong Wen", "Rui Yan", "Zhengyang Tang"], "canonical_source_url": "https://arxiv.org/abs/2605.29486", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-29486/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a41cb1a93e6aa6ba8fbc7dca55415577ab5880a6592e9b0bee5b12b5a7732cbf", "submitted_at": "2026-05-28T07:14:15Z", "updated_at": "2026-09-28T15:27:12Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2605.29486v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "identifiers": {"arxiv": "2605.29486"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.29486"}], "name": "PhoneWorld: From Real-App Trajectories to Dynamic and Verifiable Environments for Phone-Use Agents", "publication": null, "published_at": "2026-05-28", "record_sha256": "2188fbdaca19816759036694939310d297999b48062065cf272bbc416bbc31a1", "relations": [], "resource_id": "arxiv:2605.29486", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.29486v3", "topics": ["data-generation", "sandbox", "reinforcement-learning", "evaluation"], "updated_at": "2026-09-28"}, "text": "PhoneWorld: From Real-App Trajectories to Dynamic and Verifiable Environments for Phone-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PhoneWorld, which converts observed app pages, transitions and state changes into executable Android environments through a build-inspect-repair process. Generated tasks and programmatic verifiers support benchmarking, trajectory collection and online training.\nPlatforms: android\nTopics: data-generation, sandbox, reinforcement-learning, evaluation\nPublished: 2026-05-28\nAuthors: Yuxuan Liu, Xin Lai, Junyi Li, Pengyuan Lyu, Jason, Yiduo Guo, Zhengyao Fang, Yang Ding, Yi Zhang, Weinong Wang, Huawen Shen, Xingran Zhou, Liang Wu, Fei Tang, Sunqi Fan, Shangpin Peng, Zheng Ruan, Anran Zhang, Chengquan Zhang, Han Hu, Benyou Wang, Ji-Rong Wen, Rui Yan, Zhengyang Tang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experimental results and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.29486v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-29486/"}
{"id": "arxiv:2605.29534", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "461790f351a2816004027cc58d07f79ee7c670ed5064c59b1651a4efb84ff2bf", "submitted_at": "2026-05-28T07:49:09Z", "updated_at": "2026-05-28T07:49:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29534v1"}, "authors": ["Yuxiang Chai", "Han Xiao", "Xinyu Fu", "Jinpeng Chen", "Rui Liu", "Hongsheng Li"], "canonical_source_url": "https://arxiv.org/abs/2605.29534", "citation_url": "https://awesomecua.github.io/resources/arxiv-2605-29534/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "461790f351a2816004027cc58d07f79ee7c670ed5064c59b1651a4efb84ff2bf", "submitted_at": "2026-05-28T07:49:09Z", "updated_at": "2026-05-28T07:49:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2605.29534v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2605.29534"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2605.29534"}], "name": "UI-KOBE: Knowledge-Oriented Behavior Exploration for Lightweight Graph-Guided GUI Agents", "publication": null, "published_at": "2026-05-28", "record_sha256": "5d92a25468ccaddd6ca2310fd506dd792b45add3ced9404990aef28384628164", "relations": [], "resource_id": "arxiv:2605.29534", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2605.29534v1", "topics": ["mobile-navigation", "planning", "state-tracking", "memory", "efficiency"], "updated_at": "2026-05-28"}, "text": "UI-KOBE: Knowledge-Oriented Behavior Exploration for Lightweight Graph-Guided GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUI-KOBE explores mobile applications to build graphs of UI states and executable transitions. A lightweight agent then matches its current screenshot to the graph and chooses a transition, local action, completion signal or fallback action.\nPlatforms: unspecified\nTopics: mobile-navigation, planning, state-tracking, memory, efficiency\nPublished: 2026-05-28\nAuthors: Yuxiang Chai, Han Xiao, Xinyu Fu, Jinpeng Chen, Rui Liu, Hongsheng Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2605.29534v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2605-29534/"}
{"id": "arxiv:2606.01533", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc6e8af676b7de5393469d379746e674ccd662f570399622d0d5b1b0e882939", "submitted_at": "2026-06-01T01:29:36Z", "updated_at": "2026-06-01T01:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01533v1"}, "authors": ["Jing Yu Koh", "Ruslan Salakhutdinov", "Daniel Fried"], "canonical_source_url": "https://arxiv.org/abs/2606.01533", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-01533/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc6e8af676b7de5393469d379746e674ccd662f570399622d0d5b1b0e882939", "submitted_at": "2026-06-01T01:29:36Z", "updated_at": "2026-06-01T01:29:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01533v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.01533"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.01533"}], "name": "Multi-Agent Computer Use", "publication": null, "published_at": "2026-06-01", "record_sha256": "2d489d32cee9c4d379f39974a52d1b2bb5981410ee9af26283d0bf34b8261db1", "relations": [], "resource_id": "arxiv:2606.01533", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.01533v1", "topics": ["planning", "long-horizon", "workflow", "state-tracking", "efficiency"], "updated_at": "2026-06-01"}, "text": "Multi-Agent Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nMACU represents computer-use objectives as a dependency graph whose manager dispatches parallel subagents and revises tasks as new information arrives. Desktop and web experiments examine coordination, information handoff and task-completion time.\nPlatforms: desktop, web\nTopics: planning, long-horizon, workflow, state-tracking, efficiency\nPublished: 2026-06-01\nAuthors: Jing Yu Koh, Ruslan Salakhutdinov, Daniel Fried\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.01533v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-01533/"}
{"id": "arxiv:2606.01790", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73364572c0f21976048c89cd7732a16c7a952b08986792d02681a5ad0288c3ca", "submitted_at": "2026-06-01T07:10:00Z", "updated_at": "2026-06-01T07:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01790v1"}, "authors": ["Yuhang Han", "Wenzheng Yang", "Yujie Chen", "Xiangqi Jin", "Yaojie Zhang", "Siteng Huang", "Linfeng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.01790", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-01790/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73364572c0f21976048c89cd7732a16c7a952b08986792d02681a5ad0288c3ca", "submitted_at": "2026-06-01T07:10:00Z", "updated_at": "2026-06-01T07:10:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.01790v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2606.01790"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.01790"}], "name": "STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models", "publication": null, "published_at": "2026-06-01", "record_sha256": "e9a3e3edc9b49da2c07adef50be3aea49ad2c87d59d9d1d263591c39e8aebdb7", "relations": [], "resource_id": "arxiv:2606.01790", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.01790v1", "topics": ["efficiency", "training-free", "screen-understanding", "interaction-history"], "updated_at": "2026-06-01"}, "text": "STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models\nResource type: paper\nReview status: curated\nScope: core\n\nSTaR-KV compresses GUI-agent KV caches by calibrating visual-token importance across attention subspaces and interaction time. Its training-free scoring combines spatial specialization, temporal redundancy and an adaptive temperature, evaluated at matched cache budgets.\nPlatforms: unspecified\nTopics: efficiency, training-free, screen-understanding, interaction-history\nPublished: 2026-06-01\nAuthors: Yuhang Han, Wenzheng Yang, Yujie Chen, Xiangqi Jin, Yaojie Zhang, Siteng Huang, Linfeng Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.01790v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-01790/"}
{"id": "arxiv:2606.03103", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf0fcafeb2d40e489ec80784a3541114b02c6a9077a330265d5c49775eca4582", "submitted_at": "2026-06-02T03:42:34Z", "updated_at": "2026-06-02T03:42:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03103v1"}, "authors": ["Wenkai Wang", "Tao Xiong", "Jingchen Ni", "Yunpeng Bao", "Xiyun Li", "Tianqi Liu", "Hongcan Guo", "Zilong Huang", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.03103", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-03103/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf0fcafeb2d40e489ec80784a3541114b02c6a9077a330265d5c49775eca4582", "submitted_at": "2026-06-02T03:42:34Z", "updated_at": "2026-06-02T03:42:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03103v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:23Z", "role": "primary", "sha256": "9a71e95803ce58c462c84e2b45a891555e31d1fb409e643bb3ca4675a44c96e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2605.12233%2C2605.12493%2C2605.12978%2C2605.13119%2C2605.14133%2C2605.16909%2C2605.17260%2C2605.17324%2C2605.17448%2C2605.17656%2C2605.17721%2C2605.19447%2C2605.19587%2C2605.20894%2C2605.22642%2C2605.23204%2C2605.25233%2C2605.25901%2C2605.26086%2C2605.26144%2C2605.26252%2C2605.26646%2C2605.27209%2C2605.27705%2C2605.27898%2C2605.28158%2C2605.28255%2C2605.29341%2C2605.29463%2C2605.29801%2C2605.30010%2C2605.30052%2C2605.30126%2C2605.30434%2C2605.30907%2C2605.30914%2C2605.31264%2C2605.31557%2C2606.00530%2C2606.00543%2C2606.00856%2C2606.01057%2C2606.01961%2C2606.02540%2C2606.03103%2C2606.03889%2C2606.03895%2C2606.04315%2C2606.04455%2C2606.05525&max_results=50"}], "identifiers": {"arxiv": "2606.03103"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:23Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03103"}], "name": "DeskCraft: Benchmarking Desktop Agents on Professional Workflows and Human-in-the-Loop Collaboration", "publication": null, "published_at": "2026-06-02", "record_sha256": "6edbd416136e24f2aeef3fbbad6ad04b6c14a5df594e38ed466ef29706b6dad6", "relations": [], "resource_id": "arxiv:2606.03103", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.03103v1", "topics": ["evaluation", "human-in-the-loop", "long-horizon", "workflow"], "updated_at": "2026-06-02"}, "text": "DeskCraft: Benchmarking Desktop Agents on Professional Workflows and Human-in-the-Loop Collaboration\nResource type: paper\nReview status: curated\nScope: core\n\nDeskCraft evaluates professional desktop workflows and human-agent collaboration through mid-execution clarification, interruption and post-completion feedback. Its tasks cover creative applications and expose difficulties in long-horizon execution and proactive clarification.\nPlatforms: desktop\nTopics: evaluation, human-in-the-loop, long-horizon, workflow\nPublished: 2026-06-02\nAuthors: Wenkai Wang, Tao Xiong, Jingchen Ni, Yunpeng Bao, Xiyun Li, Tianqi Liu, Hongcan Guo, Zilong Huang, Shengyu Zhang\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2606.03103v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-03103/"}
{"id": "arxiv:2606.03203", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b23c12cafdd2d4a3ef63027c914ed3894c61c555edd89468c56ae40c3ec6cea9", "submitted_at": "2026-06-02T06:02:35Z", "updated_at": "2026-06-02T06:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03203v1"}, "authors": ["Jia Yu", "Zilong Wang", "Xinyang Jiang", "Dongsheng Li", "Shuo Wang"], "canonical_source_url": "https://arxiv.org/abs/2606.03203", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-03203/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b23c12cafdd2d4a3ef63027c914ed3894c61c555edd89468c56ae40c3ec6cea9", "submitted_at": "2026-06-02T06:02:35Z", "updated_at": "2026-06-02T06:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03203v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.03203"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03203"}], "name": "MedCUA-Bench: A Screenshot-Only Benchmark for Clinical Computer-Use Agents", "publication": null, "published_at": "2026-06-02", "record_sha256": "0c2d700b3747dc4ea296f2b220ad4471e741ecc13e3430a21bba4aba8230b378", "relations": [], "resource_id": "arxiv:2606.03203", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.03203v1", "topics": ["evaluation", "safety-security", "science", "workflow"], "updated_at": "2026-06-02"}, "text": "MedCUA-Bench: A Screenshot-Only Benchmark for Clinical Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMedCUA-Bench evaluates screenshot-only clinical interface workflows using paired intent- and step-level goals. It combines reconstructed medical interfaces with completion and clinical-safety checks, and separately reports results on a real OpenEMR system.\nPlatforms: unspecified\nTopics: evaluation, safety-security, science, workflow\nPublished: 2026-06-02\nAuthors: Jia Yu, Zilong Wang, Xinyang Jiang, Dongsheng Li, Shuo Wang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.03203v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-03203/"}
{"id": "arxiv:2606.03854", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7490150e82a5abfc510c60401cccf21c821d829e66cc5f2b40fd2fbbf4b0df4", "submitted_at": "2026-06-02T16:30:33Z", "updated_at": "2026-06-02T16:30:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03854v1"}, "authors": ["Yuhao Yang", "Tianyu Fan", "Chao Huang"], "canonical_source_url": "https://arxiv.org/abs/2606.03854", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-03854/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f7490150e82a5abfc510c60401cccf21c821d829e66cc5f2b40fd2fbbf4b0df4", "submitted_at": "2026-06-02T16:30:33Z", "updated_at": "2026-06-02T16:30:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.03854v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.03854"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.03854"}], "name": "CLI-Anything: Towards Agent-Native Computer Use", "publication": null, "published_at": "2026-06-02", "record_sha256": "c80c86c54f44aeb7d8e1ad78171e38048bb35a828ebeb93dfbb2846e2d1223f5", "relations": [], "resource_id": "arxiv:2606.03854", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2606.03854v1", "topics": ["automation", "tool-integration", "workflow"], "updated_at": "2026-06-02"}, "text": "CLI-Anything: Towards Agent-Native Computer Use\nResource type: paper\nReview status: curated\nScope: adjacent\n\nCLI-Anything argues for application interfaces built around structured commands, explicit state and execution feedback. It describes transforming existing software into command-line harnesses and organizing that approach through CLI-Hub.\nPlatforms: unspecified\nTopics: automation, tool-integration, workflow\nPublished: 2026-06-02\nAuthors: Yuhao Yang, Tianyu Fan, Chao Huang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced. Adjacent computer-interface architecture: the described application command harnesses concern computer operation, but no GUI execution or comparative benchmark result is established in the abstract.\n\nPrimary source: https://arxiv.org/abs/2606.03854v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-03854/"}
{"id": "arxiv:2606.04391", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d891956bde9b8672859fcfcd2d54dfd3e4720964c25d77cc28bc1200ef3a56c", "submitted_at": "2026-06-03T03:11:50Z", "updated_at": "2026-06-03T03:11:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04391v1"}, "authors": ["Jiaxi Li", "Ke Deng", "Yun Wang", "Jingyuan Huang", "Yucheng Shi", "Qiaoyu Tan", "Jin Lu", "Ninghao Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.04391", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-04391/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d891956bde9b8672859fcfcd2d54dfd3e4720964c25d77cc28bc1200ef3a56c", "submitted_at": "2026-06-03T03:11:50Z", "updated_at": "2026-06-03T03:11:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04391v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.04391"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.04391"}], "name": "Online Skill Learning for Web Agents via State-Grounded Dynamic Retrieval", "publication": null, "published_at": "2026-06-03", "record_sha256": "5214a72544f390762af7a6621f6586cf71b2d76d0d0c7f0f236ccce8a8253e88", "relations": [], "resource_id": "arxiv:2606.04391", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.04391v1", "topics": ["web-navigation", "memory", "interaction-history", "planning", "workflow"], "updated_at": "2026-06-03"}, "text": "Online Skill Learning for Web Agents via State-Grounded Dynamic Retrieval\nResource type: paper\nReview status: curated\nScope: core\n\nState-Grounded Dynamic Retrieval extracts reusable sub-procedures from completed web trajectories and represents them as paired text and code. Skills are retrieved against both the current webpage state and the task goal at intermediate execution steps.\nPlatforms: web\nTopics: web-navigation, memory, interaction-history, planning, workflow\nPublished: 2026-06-03\nAuthors: Jiaxi Li, Ke Deng, Yun Wang, Jingyuan Huang, Yucheng Shi, Qiaoyu Tan, Jin Lu, Ninghao Liu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.04391v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-04391/"}
{"id": "arxiv:2606.04701", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d5a90333c4f6be0b8d203a2f5938f7e7eefcb09fbe7ea47b85eba8c7c1c422a", "submitted_at": "2026-06-03T10:25:46Z", "updated_at": "2026-06-03T10:25:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04701v1"}, "authors": ["Jiashu Yao", "Heyan Huang", "Daiqing Wu", "Wangke Chen", "Huaxi Ai", "Haoyu Wen", "Zeming Liu", "Yuhang Guo"], "canonical_source_url": "https://arxiv.org/abs/2606.04701", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-04701/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d5a90333c4f6be0b8d203a2f5938f7e7eefcb09fbe7ea47b85eba8c7c1c422a", "submitted_at": "2026-06-03T10:25:46Z", "updated_at": "2026-06-03T10:25:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.04701v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.04701"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.04701"}], "name": "Benchmarking Living-Screen-Native GUI Agents on Short-Video Platforms", "publication": null, "published_at": "2026-06-03", "record_sha256": "74c936a706af85849da607cd8ac4ccb0ed4de7a2d2974e32b5f64eda558b7384", "relations": [], "resource_id": "arxiv:2606.04701", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.04701v1", "topics": ["evaluation", "video", "efficiency", "screen-understanding"], "updated_at": "2026-06-03"}, "text": "Benchmarking Living-Screen-Native GUI Agents on Short-Video Platforms\nResource type: paper\nReview status: curated\nScope: core\n\nLivingScreen evaluates agents in a browser-based short-video environment where observing content consumes time and affects subsequent actions. Its task suite scores accuracy together with information efficiency to diagnose over- and under-observation on continuously changing screens.\nPlatforms: web\nTopics: evaluation, video, efficiency, screen-understanding\nPublished: 2026-06-03\nAuthors: Jiashu Yao, Heyan Huang, Daiqing Wu, Wangke Chen, Huaxi Ai, Haoyu Wen, Zeming Liu, Yuhang Guo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.04701v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-04701/"}
{"id": "arxiv:2606.05233", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b649c0cef63bc15558aa4498a65043644c109faaf6403b610c2132fd3f5706d", "submitted_at": "2026-06-03T01:21:59Z", "updated_at": "2026-06-03T01:21:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05233v1"}, "authors": ["Nicholas Saban"], "canonical_source_url": "https://arxiv.org/abs/2606.05233", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-05233/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b649c0cef63bc15558aa4498a65043644c109faaf6403b610c2132fd3f5706d", "submitted_at": "2026-06-03T01:21:59Z", "updated_at": "2026-06-03T01:21:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.05233v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.05233"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.05233"}], "name": "Domain-Conditioned Safety in Frontier Computer-Using Agents: A 793-Episode Browser Benchmark, a Coding-Domain Cross-Reference, and a Reproducibility Audit of Recent Red-Teaming", "publication": null, "published_at": "2026-06-03", "record_sha256": "c98bd8ac440ca7ac870f811ef1ab6731733561dd7280375abf0c04e8c0d3fb97", "relations": [], "resource_id": "arxiv:2606.05233", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.05233v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-03"}, "text": "Domain-Conditioned Safety in Frontier Computer-Using Agents: A 793-Episode Browser Benchmark, a Coding-Domain Cross-Reference, and a Reproducibility Audit of Recent Red-Teaming\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CUA-HandCrafted, a browser-agent red-teaming benchmark that varies attack templates, tasks and system prompts. Its domain comparison and reproducibility audit caution against transferring browser-surface safety conclusions to coding agents or unreleased optimized injections.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2026-06-03\nAuthors: Nicholas Saban\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.05233v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-05233/"}
{"id": "arxiv:2606.05597", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a8afe44ffa73739505070c5f8004d8b4bbe98455fc0f5f3cef4b450bf441c74", "submitted_at": "2026-06-04T02:18:44Z", "updated_at": "2026-08-07T17:58:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.05597v3"}, "authors": ["Hao Bai", "Rui Yang", "Chenlu Ye", "Spencer Whitehead", "Aviral Kumar", "Tong Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.05597", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-05597/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8a8afe44ffa73739505070c5f8004d8b4bbe98455fc0f5f3cef4b450bf441c74", "submitted_at": "2026-06-04T02:18:44Z", "updated_at": "2026-08-07T17:58:32Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.05597v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.05597"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.05597"}], "name": "AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents", "publication": null, "published_at": "2026-06-04", "record_sha256": "57950fdb51e6aa4cc868efb6ae937f7f6ba35de15094ebac00ba52da40704982", "relations": [], "resource_id": "arxiv:2606.05597", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.05597v3", "topics": ["reinforcement-learning", "efficiency", "web-navigation"], "updated_at": "2026-08-07"}, "text": "AsyncWebRL: Efficient Asynchronous Reinforcement Learning for Multi-Step Visual Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAsyncWebRL overlaps web-agent rollouts, updates and policy refresh while managing screenshot handling and a persistent rollout pool. It also replaces trajectory-length normalization in multi-step GRPO to reduce verbose failed trajectories, with throughput and out-of-distribution evaluation.\nPlatforms: web\nTopics: reinforcement-learning, efficiency, web-navigation\nPublished: 2026-06-04\nAuthors: Hao Bai, Rui Yang, Chenlu Ye, Spencer Whitehead, Aviral Kumar, Tong Zhang\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.05597v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-05597/"}
{"id": "arxiv:2606.06322", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27ffcb10d966946de9fa2353c4b96798e5f090bdfdc5610f5e8a5c6233cf70ea", "submitted_at": "2026-06-04T15:57:29Z", "updated_at": "2026-06-04T15:57:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06322v1"}, "authors": ["Nathan Bout", "Maxime Langevin", "Ronan Riochet"], "canonical_source_url": "https://arxiv.org/abs/2606.06322", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-06322/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "27ffcb10d966946de9fa2353c4b96798e5f090bdfdc5610f5e8a5c6233cf70ea", "submitted_at": "2026-06-04T15:57:29Z", "updated_at": "2026-06-04T15:57:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06322v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "identifiers": {"arxiv": "2606.06322"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06322"}], "name": "DragOn: A Benchmark and Dataset for Drag-Based GUI Interactions", "publication": null, "published_at": "2026-06-04", "record_sha256": "30ce1ae8a92347047a90503ab613ada6d53c614dba0159693f895a640b878f91", "relations": [], "resource_id": "arxiv:2606.06322", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.06322v1", "topics": ["grounding", "data-generation", "evaluation", "screen-understanding"], "updated_at": "2026-06-04"}, "text": "DragOn: A Benchmark and Dataset for Drag-Based GUI Interactions\nResource type: paper\nReview status: curated\nScope: core\n\nDragOn supplies drag-grounding training data and a held-out benchmark for highlighting text, selecting cells, resizing elements and moving sliders. It evaluates existing vision-language models and a model fine-tuned on the drag tasks.\nPlatforms: unspecified\nTopics: grounding, data-generation, evaluation, screen-understanding\nPublished: 2026-06-04\nAuthors: Nathan Bout, Maxime Langevin, Ronan Riochet\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.06322v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-06322/"}
{"id": "arxiv:2606.06560", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd66adb4e99df2348f59e238d219d48cb6e758e694839367d2b9a64675429d87", "submitted_at": "2026-06-04T14:01:32Z", "updated_at": "2026-06-04T14:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06560v1"}, "authors": ["Victor Muryn", "Maksym Shamrai", "Sofiia Mazepa", "Yehor Khodysko"], "canonical_source_url": "https://arxiv.org/abs/2606.06560", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-06560/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fd66adb4e99df2348f59e238d219d48cb6e758e694839367d2b9a64675429d87", "submitted_at": "2026-06-04T14:01:32Z", "updated_at": "2026-06-04T14:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.06560v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.06560"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06560"}], "name": "MacArena: Benchmarking Computer Use Agents on an Online macOS Environment", "publication": null, "published_at": "2026-06-04", "record_sha256": "c041ae786197e3f13b96759d1951acfe31dda0959348858611d57f83c3ce3178", "relations": [], "resource_id": "arxiv:2606.06560", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.06560v1", "topics": ["evaluation", "robustness"], "updated_at": "2026-06-04"}, "text": "MacArena: Benchmarking Computer Use Agents on an Online macOS Environment\nResource type: paper\nReview status: curated\nScope: core\n\nMacArena runs computer-use tasks through Apple’s virtualization framework on Apple Silicon, combining task ports with macOS-native workloads. It distinguishes the ported and native subsets when examining agent performance and cross-platform transfer.\nPlatforms: desktop, macos\nTopics: evaluation, robustness\nPublished: 2026-06-04\nAuthors: Victor Muryn, Maksym Shamrai, Sofiia Mazepa, Yehor Khodysko\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.06560v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-06560/"}
{"id": "arxiv:2606.06708", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a449bb2706d5d18cbd176dcfd2f14d814cf1c1a22a08e9210c99e0722e10ea93", "submitted_at": "2026-06-04T20:48:37Z", "updated_at": "2026-08-04T18:57:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.06708v2"}, "authors": ["Shubham Gaur", "Ian Lane"], "canonical_source_url": "https://arxiv.org/abs/2606.06708", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-06708/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a449bb2706d5d18cbd176dcfd2f14d814cf1c1a22a08e9210c99e0722e10ea93", "submitted_at": "2026-06-04T20:48:37Z", "updated_at": "2026-08-04T18:57:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.06708v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.06708"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.06708"}], "name": "Signal-Driven Observation for Long-Horizon Web Agents", "publication": null, "published_at": "2026-06-04", "record_sha256": "abc20a16cd209d597b698c253d16acb45dfea7809d684e40d831d1d0776583cd", "relations": [], "resource_id": "arxiv:2606.06708", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.06708v2", "topics": ["efficiency", "screen-understanding", "long-horizon"], "updated_at": "2026-08-04"}, "text": "Signal-Driven Observation for Long-Horizon Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProposes Signal-Driven Observation: a separate call selects relevant DOM elements and selectors, and a detector requests new observations only after meaningful browser events or failures. The paper frames observation timing and compression as an architectural research problem for long-horizon web agents.\nPlatforms: web\nTopics: efficiency, screen-understanding, long-horizon\nPublished: 2026-06-04\nAuthors: Shubham Gaur, Ian Lane\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. The abstract presents an architecture proposal and open problems, without reporting an empirical evaluation.\n\nPrimary source: https://arxiv.org/abs/2606.06708v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-06708/"}
{"id": "arxiv:2606.07027", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d530152b210a69cb458fecb036c10621c72daccfbcac7c8306c7e248c87cf14", "submitted_at": "2026-06-05T08:17:28Z", "updated_at": "2026-06-12T07:28:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.07027v2"}, "authors": ["Haojie Hao", "Longkun Hao", "Yihang Lou", "Yan Bai", "Zhenyang Li", "Zhichao Yang", "Dongshuo Huang", "Hongyu Lin", "Lanqing Hong", "Jiakai Wang", "Xianglong Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.07027", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-07027/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d530152b210a69cb458fecb036c10621c72daccfbcac7c8306c7e248c87cf14", "submitted_at": "2026-06-05T08:17:28Z", "updated_at": "2026-06-12T07:28:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.07027v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.07027"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.07027"}], "name": "StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents", "publication": null, "published_at": "2026-06-05", "record_sha256": "46683990cfd4a27697db73c49ccad13b0775ab93c0e5725e7f04fc9b6ee059b2", "relations": [], "resource_id": "arxiv:2606.07027", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.07027v2", "topics": ["reward-models", "state-tracking", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-12"}, "text": "StainFlow: Entity-Stain Tracking and Evidence Linking for Process Rewards in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStainFlow tracks visually verifiable task entities through GUI trajectories to identify phases and retrieve focused evidence windows. Its process-reward model is tested for both online policy learning and trajectory-completion judgments.\nPlatforms: android\nTopics: reward-models, state-tracking, reinforcement-learning, evaluation\nPublished: 2026-06-05\nAuthors: Haojie Hao, Longkun Hao, Yihang Lou, Yan Bai, Zhenyang Li, Zhichao Yang, Dongshuo Huang, Hongyu Lin, Lanqing Hong, Jiakai Wang, Xianglong Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.07027v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-07027/"}
{"id": "arxiv:2606.07594", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33c343e9bc02976fc3e32057cdcdebc3abeae5ef049b877443e81917a8abc550", "submitted_at": "2026-05-28T17:59:31Z", "updated_at": "2026-05-28T17:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07594v1"}, "authors": ["Bo Zhang", "Borui Zhang", "Chenghao Jiang", "Minglei Shi", "Xiaofeng Wang", "Zheng Zhu", "Jie Zhou", "Jiwen Lu"], "canonical_source_url": "https://arxiv.org/abs/2606.07594", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-07594/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "33c343e9bc02976fc3e32057cdcdebc3abeae5ef049b877443e81917a8abc550", "submitted_at": "2026-05-28T17:59:31Z", "updated_at": "2026-05-28T17:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.07594v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "identifiers": {"arxiv": "2606.07594"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.07594"}], "name": "Syll: Open-Source Personal Automation with Cross-Surface Execution", "publication": null, "published_at": "2026-05-28", "record_sha256": "ffce948e2f1112b5993b338fefb89aca77980225e2081457b3f029d831209dd2", "relations": [], "resource_id": "arxiv:2606.07594", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.07594v1", "topics": ["automation", "human-in-the-loop", "workflow", "memory", "recording", "observability"], "updated_at": "2026-05-28"}, "text": "Syll: Open-Source Personal Automation with Cross-Surface Execution\nResource type: paper\nReview status: curated\nScope: core\n\nSyll combines tool calls, shell execution and visual GUI control in a personal-automation runtime. It turns user demonstrations into reusable skills and exposes execution logs, keyframes and approval checkpoints for inspection.\nPlatforms: desktop, macos\nTopics: automation, human-in-the-loop, workflow, memory, recording, observability\nPublished: 2026-05-28\nAuthors: Bo Zhang, Borui Zhang, Chenghao Jiang, Minglei Shi, Xiaofeng Wang, Zheng Zhu, Jie Zhou, Jiwen Lu\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2606.07594v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-07594/"}
{"id": "arxiv:2606.09399", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29dec745923e652654521f65a32f919a99d4577992006efc1a5e469bf775d81", "submitted_at": "2026-06-08T12:18:22Z", "updated_at": "2026-06-08T12:18:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09399v1"}, "authors": ["Radeen Mostafa", "Sawradip Saha"], "canonical_source_url": "https://arxiv.org/abs/2606.09399", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-09399/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d29dec745923e652654521f65a32f919a99d4577992006efc1a5e469bf775d81", "submitted_at": "2026-06-08T12:18:22Z", "updated_at": "2026-06-08T12:18:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09399v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2606.09399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09399"}], "name": "RunAgent SuperBrowser: A Theory of Autonomous Web Navigation Grounded in Human Browsing Behaviour", "publication": null, "published_at": "2026-06-08", "record_sha256": "1c5a7e865f5181011326de6b36cf96dcdbc1877615a11db01d02511fa5953239", "relations": [], "resource_id": "arxiv:2606.09399", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.09399v1", "topics": ["web-navigation", "grounding", "planning", "memory", "efficiency"], "updated_at": "2026-06-08"}, "text": "RunAgent SuperBrowser: A Theory of Autonomous Web Navigation Grounded in Human Browsing Behaviour\nResource type: paper\nReview status: curated\nScope: core\n\nSUPERBROWSER combines screenshot-based target proposals with separate orchestration, planning and action roles. A bounded ledger and context-eviction process retain recent actions, task facts and checkpoints during browser execution.\nPlatforms: web\nTopics: web-navigation, grounding, planning, memory, efficiency\nPublished: 2026-06-08\nAuthors: Radeen Mostafa, Sawradip Saha\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.09399v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-09399/"}
{"id": "arxiv:2606.09426", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b751efd63d98fc81166a1a80cb0c162d68bf1ec4ecca0a1a1a7717ad62e680", "submitted_at": "2026-06-08T12:39:23Z", "updated_at": "2026-07-06T14:25:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.09426v3"}, "authors": ["Wanli Li", "Bowen Zhou", "Yunyao Yu", "Zhou Xu", "Yifan Yang", "Dongsheng Li", "Caihua Shan"], "canonical_source_url": "https://arxiv.org/abs/2606.09426", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-09426/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73b751efd63d98fc81166a1a80cb0c162d68bf1ec4ecca0a1a1a7717ad62e680", "submitted_at": "2026-06-08T12:39:23Z", "updated_at": "2026-07-06T14:25:02Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.09426v3"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.09426"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09426"}], "name": "WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces", "publication": null, "published_at": "2026-06-08", "record_sha256": "8fb8ea40eb8c051e6ecb29cf513250096e9a026635a89f58df062122cc6f45cf", "relations": [], "resource_id": "arxiv:2606.09426", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.09426v3", "topics": ["evaluation", "long-horizon", "workflow", "tool-integration"], "updated_at": "2026-07-06"}, "text": "WeaveBench: A Long-Horizon, Real-World Benchmark for Computer-Use Agents with Hybrid Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nWeaveBench requires agents to combine GUI, command-line and code operations within long tasks on an Ubuntu desktop. A trajectory-aware judge checks deliverables and interaction evidence, including shortcut behavior that outcome-only grading can miss.\nPlatforms: desktop, linux\nTopics: evaluation, long-horizon, workflow, tool-integration\nPublished: 2026-06-08\nAuthors: Wanli Li, Bowen Zhou, Yunyao Yu, Zhou Xu, Yifan Yang, Dongsheng Li, Caihua Shan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.09426v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-09426/"}
{"id": "arxiv:2606.09447", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a852c77650fa54f234794c6cd27777f4e7219b8389b509d807e27db16138a1b", "submitted_at": "2026-06-08T12:55:42Z", "updated_at": "2026-06-08T12:55:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09447v1"}, "authors": ["Bojie Rong", "Zheyu Shen", "Qiaoping Wang", "Pengfei Kang", "Yang Xu", "Yawen Wei", "Hanyu Wu", "Zhi Zhao", "Leihao Pei", "Linquan Jiang"], "canonical_source_url": "https://arxiv.org/abs/2606.09447", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-09447/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a852c77650fa54f234794c6cd27777f4e7219b8389b509d807e27db16138a1b", "submitted_at": "2026-06-08T12:55:42Z", "updated_at": "2026-06-08T12:55:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09447v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.09447"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09447"}], "name": "AliyunConsoleAgent: Training Web Agents in Real-World Cloud Environments via Distillation and Reinforcement Learning", "publication": null, "published_at": "2026-06-08", "record_sha256": "b989ba0d8c8b68dfa607c0454fe78d5951c71cba4c5d5cd4a22e7a14d80dd628", "relations": [], "resource_id": "arxiv:2606.09447", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.09447v1", "topics": ["web-navigation", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-08"}, "text": "AliyunConsoleAgent: Training Web Agents in Real-World Cloud Environments via Distillation and Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nAliyunConsoleAgent trains browser agents to verify documented procedures against changing cloud-console UIs through trajectory distillation and online GRPO. Provisioned rollout environments and backend audit-log checks support training and outcome evaluation.\nPlatforms: web\nTopics: web-navigation, imitation-learning, reinforcement-learning, evaluation\nPublished: 2026-06-08\nAuthors: Bojie Rong, Zheyu Shen, Qiaoping Wang, Pengfei Kang, Yang Xu, Yawen Wei, Hanyu Wu, Zhi Zhao, Leihao Pei, Linquan Jiang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.09447v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-09447/"}
{"id": "arxiv:2606.09764", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "495420bd87ba06ae2ec1763865b1145271e964b4bd40145c8e591653a1581c42", "submitted_at": "2026-06-08T17:27:13Z", "updated_at": "2026-06-08T17:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09764v1"}, "authors": ["Lawrence Keunho Jang", "Mareks Woodside", "Geronimo Carom", "Andrew Keunwoo Jang", "Jing Yu Koh", "Ruslan Salakhutdinov"], "canonical_source_url": "https://arxiv.org/abs/2606.09764", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-09764/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "495420bd87ba06ae2ec1763865b1145271e964b4bd40145c8e591653a1581c42", "submitted_at": "2026-06-08T17:27:13Z", "updated_at": "2026-06-08T17:27:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.09764v1"}}, "environments": ["ios"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:02Z", "role": "primary", "sha256": "e436adf2735cb537ee4b24aa0c3f36930a50c2f683cc6868e7c40df602bdb6b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2604.00892%2C2604.01221%2C2604.01676%2C2604.02345%2C2604.03465%2C2604.05157%2C2604.06134%2C2604.06240%2C2604.06995%2C2604.07277%2C2604.07831%2C2604.07929%2C2604.08005%2C2604.09581%2C2604.10261%2C2604.10577%2C2604.10988%2C2604.11201%2C2604.18847%2C2604.21375%2C2604.24441%2C2604.26020%2C2605.00551%2C2605.02168%2C2605.07505%2C2605.10347%2C2605.12294%2C2605.15040%2C2605.16883%2C2605.27134%2C2605.28258%2C2606.01790%2C2606.02031%2C2606.02208%2C2606.03203%2C2606.03236%2C2606.03951%2C2606.04391%2C2606.04627%2C2606.04701%2C2606.05233%2C2606.05597%2C2606.05697%2C2606.06560%2C2606.06708%2C2606.07027%2C2606.07074%2C2606.09399%2C2606.09447%2C2606.09764&max_results=50"}], "identifiers": {"arxiv": "2606.09764"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.09764"}], "name": "iOSWorld: A Benchmark for Personally Intelligent Phone Agents", "publication": null, "published_at": "2026-06-08", "record_sha256": "2cef6e9c87141bcc3f40a5a7bfa9ea6d9a376dba9b1085d979da1feff6313809", "relations": [], "resource_id": "arxiv:2606.09764", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.09764v1", "topics": ["evaluation", "cross-app", "memory"], "updated_at": "2026-06-08"}, "text": "iOSWorld: A Benchmark for Personally Intelligent Phone Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces a native iOS simulator benchmark with persistent user data, multi-app tasks, and personalization-dependent phone workflows.\nPlatforms: ios\nTopics: evaluation, cross-app, memory\nPublished: 2026-06-08\nAuthors: Lawrence Keunho Jang, Mareks Woodside, Geronimo Carom, Andrew Keunwoo Jang, Jing Yu Koh, Ruslan Salakhutdinov\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.09764v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-09764/"}
{"id": "arxiv:2606.10423", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "386bd82be97d5cbe64fdb3478e427362e710d78d423c9e52661a8a96ea2fdecf", "submitted_at": "2026-06-09T04:53:19Z", "updated_at": "2026-06-09T04:53:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10423v1"}, "authors": ["Jayoo Hwang", "Xiaowen Zhang", "Vedant Padwal"], "canonical_source_url": "https://arxiv.org/abs/2606.10423", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-10423/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "386bd82be97d5cbe64fdb3478e427362e710d78d423c9e52661a8a96ea2fdecf", "submitted_at": "2026-06-09T04:53:19Z", "updated_at": "2026-06-09T04:53:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.10423v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.10423"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.10423"}], "name": "WebChallenger: A Reliable and Efficient Generalist Web Agent", "publication": null, "published_at": "2026-06-09", "record_sha256": "ea9c471d3a24d9960b59d2e80f55c2fc5e7a4dfe288867c884e5316b8ab2a624", "relations": [], "resource_id": "arxiv:2606.10423", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.10423v1", "topics": ["web-navigation", "memory", "efficiency", "training-free", "workflow"], "updated_at": "2026-06-09"}, "text": "WebChallenger: A Reliable and Efficient Generalist Web Agent\nResource type: paper\nReview status: curated\nScope: core\n\nWebChallenger builds a hierarchical DOM-derived page representation for selective observation, reusable site maps and compound actions. The framework combines these mechanisms without model fine-tuning and evaluates generalization across multiple web-agent benchmarks.\nPlatforms: web\nTopics: web-navigation, memory, efficiency, training-free, workflow\nPublished: 2026-06-09\nAuthors: Jayoo Hwang, Xiaowen Zhang, Vedant Padwal\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.10423v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-10423/"}
{"id": "arxiv:2606.10522", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7523d3901755d8feeecbaf2f02dcc00fd0a40102e1498d49b9f789b0a7c02d75", "submitted_at": "2026-06-09T07:52:10Z", "updated_at": "2026-07-06T06:23:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.10522v2"}, "authors": ["Can Lin", "Tao Feng", "Hangjie Yuan", "Dan Zhang", "Yifan Zhu", "Zhonghong Ou"], "canonical_source_url": "https://arxiv.org/abs/2606.10522", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-10522/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7523d3901755d8feeecbaf2f02dcc00fd0a40102e1498d49b9f789b0a7c02d75", "submitted_at": "2026-06-09T07:52:10Z", "updated_at": "2026-07-06T06:23:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.10522v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.10522"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.10522"}], "name": "GUI-AC: Enhancing Continual Learning in GUI Agents", "publication": null, "published_at": "2026-06-09", "record_sha256": "710a4b81005f786dd9b8c2093cd9dd597723e17adb8946efa81baf1f9e2d108a", "relations": [], "resource_id": "arxiv:2606.10522", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.10522v2", "topics": ["reinforcement-learning", "grounding", "robustness"], "updated_at": "2026-07-06"}, "text": "GUI-AC: Enhancing Continual Learning in GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-AC uses grounding certainty to adjust advantage weighting and policy-clipping bounds during continual reinforcement fine-tuning. Its mechanisms aim to suppress noisy confidence updates while retaining exploration under shifting GUI domains and resolutions.\nPlatforms: unspecified\nTopics: reinforcement-learning, grounding, robustness\nPublished: 2026-06-09\nAuthors: Can Lin, Tao Feng, Hangjie Yuan, Dan Zhang, Yifan Zhu, Zhonghong Ou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.10522v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-10522/"}
{"id": "arxiv:2606.11042", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "037a3197137d27318320a3fd23fb5e606a1aa748133e57213cd02fd10aad4c4d", "submitted_at": "2026-06-09T16:10:16Z", "updated_at": "2026-07-17T12:22:12Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2606.11042v4"}, "authors": ["Liya Zhu", "Jingzhe Ding", "Jian Zhang", "Jianbo Xue", "Shihao Liang", "Ge Zhang", "Yi Zhu", "Duju Zeng", "Xiang Gao", "Qingshui Gu", "Mailun Gao", "Huimin Che", "Yan Zhao", "Peiheng Zhou", "Haojun Wang", "Chaobo Xian", "Lili Le", "Chi Wu", "Yiwei Liu", "Shengda Long", "Jiale Yang", "Fangzhi Xu", "Sijin Wu", "Haodong Duan", "Chao He", "Zhaojian Li", "Minchao Wang", "Huan Zhou", "Jiani Hou", "Chuqian Yu", "Weiran Shi", "Hongwan Gao", "Jiamin Chen", "Guanhong Chen", "Tingqin Luo", "Kaiyuan Zhang", "Zhixin Yao", "Qing Hua", "Yuhao Jiang", "Jin Chen", "Pu Chen", "Zhenyu Hu", "Xingyu Li", "Zhengxuan Jiang", "Meng Cao", "Tianfeng Long", "Haozhe Wang", "Mingzhang Wang", "Yichen Zhang", "Yiming Dai", "Chenchen Zhang", "Jiaying Wang", "Xinying Liu", "Xingzu Liu", "Lingling Zhang", "Xinjie Chen", "Yujia Qin", "Wangchunshu Zhou", "Zhiyong Wu", "Yang Liu", "Jiaheng Liu", "Lei Zhang", "Shen Yan", "Wenhao Huang", "Zaiyuan Wang", "Xiaolong Chang"], "canonical_source_url": "https://arxiv.org/abs/2606.11042", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-11042/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "037a3197137d27318320a3fd23fb5e606a1aa748133e57213cd02fd10aad4c4d", "submitted_at": "2026-06-09T16:10:16Z", "updated_at": "2026-07-17T12:22:12Z", "version": 4, "versioned_url": "https://arxiv.org/abs/2606.11042v4"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.11042"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.11042"}], "name": "Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields", "publication": null, "published_at": "2026-06-09", "record_sha256": "d07aa8dfc7fde468bcdd827caab22b24a0663c746b69c99d0c505c2bf945b8b0", "relations": [], "resource_id": "arxiv:2606.11042", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.11042v4", "topics": ["evaluation", "long-horizon", "workflow", "robustness"], "updated_at": "2026-07-17"}, "text": "Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields\nResource type: paper\nReview status: curated\nScope: core\n\nWorkflow-GYM evaluates long GUI workflows in professional software rather than isolated interface actions. Its analysis examines failures such as omitted stages, accumulated errors and goal drift during specialized end-to-end work.\nPlatforms: unspecified\nTopics: evaluation, long-horizon, workflow, robustness\nPublished: 2026-06-09\nAuthors: Liya Zhu, Jingzhe Ding, Jian Zhang, Jianbo Xue, Shihao Liang, Ge Zhang, Yi Zhu, Duju Zeng, Xiang Gao, Qingshui Gu, Mailun Gao, Huimin Che, Yan Zhao, Peiheng Zhou, Haojun Wang, Chaobo Xian, Lili Le, Chi Wu, Yiwei Liu, Shengda Long, Jiale Yang, Fangzhi Xu, Sijin Wu, Haodong Duan, Chao He, Zhaojian Li, Minchao Wang, Huan Zhou, Jiani Hou, Chuqian Yu, Weiran Shi, Hongwan Gao, Jiamin Chen, Guanhong Chen, Tingqin Luo, Kaiyuan Zhang, Zhixin Yao, Qing Hua, Yuhao Jiang, Jin Chen, Pu Chen, Zhenyu Hu, Xingyu Li, Zhengxuan Jiang, Meng Cao, Tianfeng Long, Haozhe Wang, Mingzhang Wang, Yichen Zhang, Yiming Dai, Chenchen Zhang, Jiaying Wang, Xinying Liu, Xingzu Liu, Lingling Zhang, Xinjie Chen, Yujia Qin, Wangchunshu Zhou, Zhiyong Wu, Yang Liu, Jiaheng Liu, Lei Zhang, Shen Yan, Wenhao Huang, Zaiyuan Wang, Xiaolong Chang\nRecorded arXiv version: v4\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.11042v4\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-11042/"}
{"id": "arxiv:2606.11078", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8ed399c8b489a80c5313a00232d2ecca0fbc8228b1f56745c02d4fedb6f6908", "submitted_at": "2026-06-09T16:39:10Z", "updated_at": "2026-06-09T16:39:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.11078v1"}, "authors": ["Jaewoo Lee", "Zaid Khan", "Archiki Prasad", "Justin Chih-Yao Chen", "Supriyo Chakraborty", "Kartik Balasubramaniam", "Sambit Sahu", "Elias Stengel-Eskin", "Hyunji Lee", "Mohit Bansal"], "canonical_source_url": "https://arxiv.org/abs/2606.11078", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-11078/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a8ed399c8b489a80c5313a00232d2ecca0fbc8228b1f56745c02d4fedb6f6908", "submitted_at": "2026-06-09T16:39:10Z", "updated_at": "2026-06-09T16:39:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.11078v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.11078"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.11078"}], "name": "A History-Aware Visually Grounded Critic for Computer Use Agents", "publication": null, "published_at": "2026-06-09", "record_sha256": "fb44084216f61b57f82ec22f5c725ee02aa0d6aa412d0f4d8e7386ef0f6d51ac", "relations": [], "resource_id": "arxiv:2606.11078", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.11078v1", "topics": ["interaction-history", "memory", "grounding", "reward-models", "long-horizon", "evaluation"], "updated_at": "2026-06-09"}, "text": "A History-Aware Visually Grounded Critic for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nHiViG combines a compact history of completed actions with a visually grounded critic that checks proposed coordinates before execution. Its evaluations separate history-related planning failures from errors caused by targeting the wrong screen element.\nPlatforms: web, desktop\nTopics: interaction-history, memory, grounding, reward-models, long-horizon, evaluation\nPublished: 2026-06-09\nAuthors: Jaewoo Lee, Zaid Khan, Archiki Prasad, Justin Chih-Yao Chen, Supriyo Chakraborty, Kartik Balasubramaniam, Sambit Sahu, Elias Stengel-Eskin, Hyunji Lee, Mohit Bansal\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.11078v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-11078/"}
{"id": "arxiv:2606.12485", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "408c8e4bc46072641819f7ab1d97ba98604c06ba7e91049b8d0e29e1a8ff2db1", "submitted_at": "2026-06-10T08:56:27Z", "updated_at": "2026-08-17T02:54:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12485v2"}, "authors": ["Longkun Hao", "Hongyu Lin", "Hao Li", "Zhuowen Liu", "Zhichao Yang", "Haojie Hao", "Dongshuo Huang", "Haitao Yang", "Hongyu Ge", "Ming jie Xie", "Yanjun Wu", "Zi Hao Yin", "Yan Bai", "Yihang Lou"], "canonical_source_url": "https://arxiv.org/abs/2606.12485", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-12485/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "408c8e4bc46072641819f7ab1d97ba98604c06ba7e91049b8d0e29e1a8ff2db1", "submitted_at": "2026-06-10T08:56:27Z", "updated_at": "2026-08-17T02:54:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12485v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.12485"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.12485"}], "name": "Speculative Rollback Correction for Quality-Diverse Web Agent Imitation", "publication": null, "published_at": "2026-06-10", "record_sha256": "c9012af0c92cac33fb786d448368c148d4e568c3394735efe0886238d9b187ce", "relations": [], "resource_id": "arxiv:2606.12485", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.12485v2", "topics": ["imitation-learning", "data-generation", "robustness"], "updated_at": "2026-08-17"}, "text": "Speculative Rollback Correction for Quality-Diverse Web Agent Imitation\nResource type: paper\nReview status: curated\nScope: core\n\nSpeculative Rollback Correction lets a student web agent execute short branches before teacher review localizes harmful deviations. Verified continuations and localized corrections feed supervised training while an archive retains diverse successful solutions.\nPlatforms: web\nTopics: imitation-learning, data-generation, robustness\nPublished: 2026-06-10\nAuthors: Longkun Hao, Hongyu Lin, Hao Li, Zhuowen Liu, Zhichao Yang, Haojie Hao, Dongshuo Huang, Haitao Yang, Hongyu Ge, Ming jie Xie, Yanjun Wu, Zi Hao Yin, Yan Bai, Yihang Lou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.12485v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-12485/"}
{"id": "arxiv:2606.12666", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fed1c6e109cc5d34aebe3bb333955e8bf792264d4c1ed0893d76534a77649afd", "submitted_at": "2026-06-10T20:48:39Z", "updated_at": "2026-06-16T03:15:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12666v2"}, "authors": ["Siyu Shen", "Fenghao Xu", "Wenrui Diao", "Kehuan Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.12666", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-12666/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fed1c6e109cc5d34aebe3bb333955e8bf792264d4c1ed0893d76534a77649afd", "submitted_at": "2026-06-10T20:48:39Z", "updated_at": "2026-06-16T03:15:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.12666v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.12666"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.12666"}], "name": "CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agents", "publication": null, "published_at": "2026-06-10", "record_sha256": "7a8d68116ca92d32eb8cb251b40bbd6ee77f00d14e73dcfc79b8f78fd5b952ba", "relations": [], "resource_id": "arxiv:2606.12666", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.12666v2", "topics": ["safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-06-16"}, "text": "CAPED: Context-Aware Privacy Exposure Defense for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCAPED filters mobile screenshots before upload by combining task requirements, screen context and parsed UI elements to mask incidental private content. AndroidWorld and a controlled seeded privacy evaluation examine leakage reduction alongside task utility and prototype costs.\nPlatforms: android\nTopics: safety-security, screen-understanding, mobile-navigation\nPublished: 2026-06-10\nAuthors: Siyu Shen, Fenghao Xu, Wenrui Diao, Kehuan Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.12666v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-12666/"}
{"id": "arxiv:2606.13239", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edf7e63803efba89551aa297bff1da9b0babf8ec898a2971fa6c8a20cc58dea8", "submitted_at": "2026-06-11T11:53:32Z", "updated_at": "2026-06-30T03:31:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13239v2"}, "authors": ["Jiaxin Ai", "Tao Hu", "Xuemeng Yang", "Shu Zou", "Hairong Zhang", "Daocheng Fu", "Yu Yang", "Hongbin Zhou", "Nianchen Deng", "Pinlong Cai", "Zhongyuan Wang", "Botian Shi", "Kaipeng Zhang", "Licheng Wen"], "canonical_source_url": "https://arxiv.org/abs/2606.13239", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-13239/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edf7e63803efba89551aa297bff1da9b0babf8ec898a2971fa6c8a20cc58dea8", "submitted_at": "2026-06-11T11:53:32Z", "updated_at": "2026-06-30T03:31:20Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13239v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.13239"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.13239"}], "name": "ComAct: Reframing Professional Software Manipulation via COM-as-Action Paradigm", "publication": null, "published_at": "2026-06-11", "record_sha256": "d38ff12731317fc65d4f21e6a5dc12042d08e376dcdf185ae16f7aa2d2e58f2c", "relations": [], "resource_id": "arxiv:2606.13239", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2606.13239v2", "topics": ["automation", "evaluation", "tool-integration", "science"], "updated_at": "2026-06-30"}, "text": "ComAct: Reframing Professional Software Manipulation via COM-as-Action Paradigm\nResource type: paper\nReview status: curated\nScope: adjacent\n\nComAct proposes COM program synthesis as an alternative interface for manipulating professional software and introduces ComCADBench with GUI-based comparison agents. ComActor adds a staged training and correction approach, supported by a Windows-container training platform.\nPlatforms: desktop, windows\nTopics: automation, evaluation, tool-integration, science\nPublished: 2026-06-11\nAuthors: Jiaxin Ai, Tao Hu, Xuemeng Yang, Shu Zou, Hairong Zhang, Daocheng Fu, Yu Yang, Hongbin Zhou, Nianchen Deng, Pinlong Cai, Zhongyuan Wang, Botian Shi, Kaipeng Zhang, Licheng Wen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.13239v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-13239/"}
{"id": "arxiv:2606.13385", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88b7e2ee8717661e4f8b417916e35b844171ecd8a8a2add657cf341c8f759536", "submitted_at": "2026-06-11T14:12:43Z", "updated_at": "2026-07-27T14:19:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13385v2"}, "authors": ["Zihao Wang", "Yiming Li", "Yutong Wu", "Kangjie Chen", "Zheyu Liu", "Fok Kar Wai", "Pin-Yu Chen", "Vrizlynn L. L. Thing", "Bo Li", "Dacheng Tao", "Tianwei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.13385", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-13385/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88b7e2ee8717661e4f8b417916e35b844171ecd8a8a2add657cf341c8f759536", "submitted_at": "2026-06-11T14:12:43Z", "updated_at": "2026-07-27T14:19:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.13385v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.13385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.13385"}], "name": "Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents", "publication": null, "published_at": "2026-06-11", "record_sha256": "2827cb00bebef84ba6e37ef3c17bf59500990e412c5478cf9b325f1206b458ba", "relations": [], "resource_id": "arxiv:2606.13385", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.13385v2", "topics": ["evaluation", "safety-security", "web-navigation"], "updated_at": "2026-07-27"}, "text": "Who Pays the Price? Stakeholder-Centric Prompt Injection Benchmarking for Real-world Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStakeBench evaluates prompt-injection harm to different stakeholders in shopping web-agent workflows. Its executable cases separate attack objectives and assess both outcomes and execution processes to reveal how the same attack pattern can affect different parties.\nPlatforms: web\nTopics: evaluation, safety-security, web-navigation\nPublished: 2026-06-11\nAuthors: Zihao Wang, Yiming Li, Yutong Wu, Kangjie Chen, Zheyu Liu, Fok Kar Wai, Pin-Yu Chen, Vrizlynn L. L. Thing, Bo Li, Dacheng Tao, Tianwei Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.13385v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-13385/"}
{"id": "arxiv:2606.14106", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb87da27071ab21519599b7d16359a38118aa358f620de94e66b15cad040828", "submitted_at": "2026-06-12T04:39:01Z", "updated_at": "2026-06-12T04:39:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14106v1"}, "authors": ["Seoyoung Choi", "Minseok Ko", "Hyunseok Lee", "Kunwoong Kim", "Woomin Song", "Chanseok Jeon", "Jinwoo Shin"], "canonical_source_url": "https://arxiv.org/abs/2606.14106", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-14106/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1fb87da27071ab21519599b7d16359a38118aa358f620de94e66b15cad040828", "submitted_at": "2026-06-12T04:39:01Z", "updated_at": "2026-06-12T04:39:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14106v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.14106"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14106"}], "name": "Naive Visual Memory is Not Enough: A Failure-Mode Study of GUI Agents", "publication": null, "published_at": "2026-06-12", "record_sha256": "443c1533984516fb121d9b836bff0bb47db45e485d8b2c95727e59768b8836b3", "relations": [], "resource_id": "arxiv:2606.14106", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.14106v1", "topics": ["memory", "interaction-history", "grounding", "robustness", "evaluation"], "updated_at": "2026-06-12"}, "text": "Naive Visual Memory is Not Enough: A Failure-Mode Study of GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study finds that full-image visual memory can reduce some state errors while worsening action-level failures. Its AGMem alternative stores local crops associated with successful actions or recovery and evaluates their use in desktop GUI tasks.\nPlatforms: desktop\nTopics: memory, interaction-history, grounding, robustness, evaluation\nPublished: 2026-06-12\nAuthors: Seoyoung Choi, Minseok Ko, Hyunseok Lee, Kunwoong Kim, Woomin Song, Chanseok Jeon, Jinwoo Shin\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.14106v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-14106/"}
{"id": "arxiv:2606.14579", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a5e949a760f236c08ae3cebc32d8a52220b1101f5324770d40bd7d054351f6a", "submitted_at": "2026-06-12T15:58:18Z", "updated_at": "2026-06-12T15:58:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14579v1"}, "authors": ["Xinyu Qiu", "Yunzhu Zhang", "Heng Jia", "Shuheng Shen", "Changhua Meng", "Linchao Zhu"], "canonical_source_url": "https://arxiv.org/abs/2606.14579", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-14579/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1a5e949a760f236c08ae3cebc32d8a52220b1101f5324770d40bd7d054351f6a", "submitted_at": "2026-06-12T15:58:18Z", "updated_at": "2026-06-12T15:58:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14579v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.14579"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14579"}], "name": "VISTA: View-Consistent Self-Verified Training for GUI Grounding", "publication": null, "published_at": "2026-06-12", "record_sha256": "23a6e1e199931115746ee99480b8df9b80dca2d904e9bfdfd774832b755a957c", "relations": [], "resource_id": "arxiv:2606.14579", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.14579v1", "topics": ["grounding", "reinforcement-learning", "robustness", "data-generation"], "updated_at": "2026-06-12"}, "text": "VISTA: View-Consistent Self-Verified Training for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nVISTA forms GRPO comparison groups from target-preserving crops of the same GUI instance with exactly remapped target boxes. A conditional, advantage-weighted cross-view anchor supplies additional supervision without entering the group baseline, and evaluation includes view-consistency robustness.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, robustness, data-generation\nPublished: 2026-06-12\nAuthors: Xinyu Qiu, Yunzhu Zhang, Heng Jia, Shuheng Shen, Changhua Meng, Linchao Zhu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.14579v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-14579/"}
{"id": "arxiv:2606.14832", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "280ec12bcb246f1f07cf3627011945e8a9ff4afed93b1fb2937f2d6b1bc1323b", "submitted_at": "2026-06-12T15:01:32Z", "updated_at": "2026-06-12T15:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14832v1"}, "authors": ["Chenxin Li", "Zhengyao Fang", "Zhengyang Tang", "Pengyuan Lyu", "Xingran Zhou", "Xin Lai", "Fei Tang", "Liang Wu", "Yiduo Guo", "Weinong Wang", "Junyi Li", "Yi Zhang", "Yang Ding", "Huawen Shen", "Sunqi Fan", "Shangpin Peng", "Zheng Ruan", "Anran Zhang", "Benyou Wang", "Chengquan Zhang", "Han Hu"], "canonical_source_url": "https://arxiv.org/abs/2606.14832", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-14832/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "280ec12bcb246f1f07cf3627011945e8a9ff4afed93b1fb2937f2d6b1bc1323b", "submitted_at": "2026-06-12T15:01:32Z", "updated_at": "2026-06-12T15:01:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.14832v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.14832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.14832"}], "name": "PhoneHarness: Harnessing Phone-Use Agents through Mixed GUI, CLI, and Tool Actions", "publication": null, "published_at": "2026-06-12", "record_sha256": "8260206560332a7b3a5c6714b99fbf7d7fcc5f002f1beb379e6e2fe08325fdf7", "relations": [], "resource_id": "arxiv:2606.14832", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.14832v1", "topics": ["mobile-navigation", "tool-integration", "evaluation", "workflow"], "updated_at": "2026-06-12"}, "text": "PhoneHarness: Harnessing Phone-Use Agents through Mixed GUI, CLI, and Tool Actions\nResource type: paper\nReview status: curated\nScope: core\n\nPhoneHarness routes phone workflows across GUI actions, device-side commands and host tools while retaining auditable execution traces. Its accompanying benchmark checks observable task side effects, separating mixed-action execution from plausible final-answer generation.\nPlatforms: unspecified\nTopics: mobile-navigation, tool-integration, evaluation, workflow\nPublished: 2026-06-12\nAuthors: Chenxin Li, Zhengyao Fang, Zhengyang Tang, Pengyuan Lyu, Xingran Zhou, Xin Lai, Fei Tang, Liang Wu, Yiduo Guo, Weinong Wang, Junyi Li, Yi Zhang, Yang Ding, Huawen Shen, Sunqi Fan, Shangpin Peng, Zheng Ruan, Anran Zhang, Benyou Wang, Chengquan Zhang, Han Hu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.14832v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-14832/"}
{"id": "arxiv:2606.15017", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50f1acd5727f84946cf95d181f270dbc4f46f15d24339e6cb257fea1b23b21a0", "submitted_at": "2026-06-12T23:30:14Z", "updated_at": "2026-08-30T17:49:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.15017v2"}, "authors": ["Sina Hajimiri", "Masih Aminbeidokhti", "Jose Dolz", "Ismail Ben Ayed", "Issam H. Laradji", "Spandana Gella", "Nicolas Gontier"], "canonical_source_url": "https://arxiv.org/abs/2606.15017", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-15017/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50f1acd5727f84946cf95d181f270dbc4f46f15d24339e6cb257fea1b23b21a0", "submitted_at": "2026-06-12T23:30:14Z", "updated_at": "2026-08-30T17:49:13Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.15017v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.15017"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.15017"}], "name": "Are Online Skill and Memory Modules Always Worth Their Tokens? A Budget-Constrained Study of Web Agents", "publication": null, "published_at": "2026-06-12", "record_sha256": "0bfc2d8abd76f33eb29b4912d5dd85cf6d357cfb98b113a9b209af7766b807bf", "relations": [], "resource_id": "arxiv:2606.15017", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.15017v2", "topics": ["memory", "efficiency", "evaluation", "robustness"], "updated_at": "2026-08-30"}, "text": "Are Online Skill and Memory Modules Always Worth Their Tokens? A Budget-Constrained Study of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCompares online skill and memory augmentations with a vanilla web agent given the same total inference-token budget. Across WebArena and WorkArena settings, it studies when module overhead removes aggregate gains and how rerun variability changes conclusions.\nPlatforms: web\nTopics: memory, efficiency, evaluation, robustness\nPublished: 2026-06-12\nAuthors: Sina Hajimiri, Masih Aminbeidokhti, Jose Dolz, Ismail Ben Ayed, Issam H. Laradji, Spandana Gella, Nicolas Gontier\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.15017v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-15017/"}
{"id": "arxiv:2606.15034", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7bf0297fba942441b19bcfeba147838bdef888b428f1150c2462d8bf3d986882", "submitted_at": "2026-06-13T00:32:24Z", "updated_at": "2026-10-09T19:45:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.15034v3"}, "authors": ["Mina Mohammadmirzaei", "Jeffrey Flanigan"], "canonical_source_url": "https://arxiv.org/abs/2606.15034", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-15034/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7bf0297fba942441b19bcfeba147838bdef888b428f1150c2462d8bf3d986882", "submitted_at": "2026-06-13T00:32:24Z", "updated_at": "2026-10-09T19:45:15Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2606.15034v3"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.15034"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.15034"}], "name": "OSGuard: A Benchmark for Safety in Computer-Use Agents", "publication": null, "published_at": "2026-06-13", "record_sha256": "bd1afa86a5a97f4f28e906b72a5b98418bb8708740595124d233fc261bd29931", "relations": [], "resource_id": "arxiv:2606.15034", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.15034v3", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-09"}, "text": "OSGuard: A Benchmark for Safety in Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates action-level guardrails and full computer tasks with environment-dependent safety constraints, separating safe completion from nominal success.\nPlatforms: desktop\nTopics: safety-security, evaluation\nPublished: 2026-06-13\nAuthors: Mina Mohammadmirzaei, Jeffrey Flanigan\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.15034v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-15034/"}
{"id": "arxiv:2606.16659", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6311b22ddb2262c74efbe79acc8856335f6870aa24197c507f3f7b048d6f7901", "submitted_at": "2026-06-15T12:53:55Z", "updated_at": "2026-06-15T12:53:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16659v1"}, "authors": ["Y. H. Zhou", "Z. M. Ma", "Y. J. Zhou", "Y. T. Li", "H. X. Xiang", "Y. M. Cheng", "T. L. Chen", "K. J. Zhang", "Z. H. Nan", "J. H. Ni", "Z. Wu", "Q. Y. Pan", "S. Zhang", "S. Cheng", "M. Y. Luo"], "canonical_source_url": "https://arxiv.org/abs/2606.16659", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-16659/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6311b22ddb2262c74efbe79acc8856335f6870aa24197c507f3f7b048d6f7901", "submitted_at": "2026-06-15T12:53:55Z", "updated_at": "2026-06-15T12:53:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16659v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.16659"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16659"}], "name": "FraudSMSWalker: Benchmarking Agentic Large Language Models for SMS-to-Webpage Fraud Detection", "publication": null, "published_at": "2026-06-15", "record_sha256": "5db979400b22323b0460947c876394ed883844d8dbd417f871bb868b95574839", "relations": [], "resource_id": "arxiv:2606.16659", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.16659v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-06-15"}, "text": "FraudSMSWalker: Benchmarking Agentic Large Language Models for SMS-to-Webpage Fraud Detection\nResource type: paper\nReview status: curated\nScope: core\n\nFraudSMSWalker tests browser agents on bilingual SMS-to-webpage fraud judgments while hiding URL and reputation shortcuts. Its controlled chains include difficult benign login/payment cases and evaluate whether classifications remain supported by the visible message and page evidence.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2026-06-15\nAuthors: Y. H. Zhou, Z. M. Ma, Y. J. Zhou, Y. T. Li, H. X. Xiang, Y. M. Cheng, T. L. Chen, K. J. Zhang, Z. H. Nan, J. H. Ni, Z. Wu, Q. Y. Pan, S. Zhang, S. Cheng, M. Y. Luo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.16659v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-16659/"}
{"id": "arxiv:2606.16748", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc529f96d6df52e3560f66df8e0630dabdae3ab30cadc8b59123f347eba187a", "submitted_at": "2026-06-15T14:08:09Z", "updated_at": "2026-06-15T14:08:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16748v1"}, "authors": ["Lawrence Keunho Jang", "Andrew Keunwoo Jang", "Jing Yu Koh", "Ruslan Salakhutdinov"], "canonical_source_url": "https://arxiv.org/abs/2606.16748", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-16748/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "acc529f96d6df52e3560f66df8e0630dabdae3ab30cadc8b59123f347eba187a", "submitted_at": "2026-06-15T14:08:09Z", "updated_at": "2026-06-15T14:08:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16748v1"}}, "environments": ["desktop", "linux", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.16748"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16748"}], "name": "MyPCBench: A Benchmark for Personally Intelligent Computer-Use Agents", "publication": null, "published_at": "2026-06-15", "record_sha256": "c591b1cf7680bf57f9ba00695ad23961da3baa6139ccc052c583ffafc006eb20", "relations": [], "resource_id": "arxiv:2606.16748", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.16748v1", "topics": ["evaluation", "cross-app", "workflow"], "updated_at": "2026-06-15"}, "text": "MyPCBench: A Benchmark for Personally Intelligent Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMyPCBench tests personal-assistant tasks on a Linux desktop with simulated web applications and data for a single fictional persona. A shared computer-and-shell interface enables evaluation of cross-application workflows that depend on personal context and account-like state.\nPlatforms: desktop, linux, web\nTopics: evaluation, cross-app, workflow\nPublished: 2026-06-15\nAuthors: Lawrence Keunho Jang, Andrew Keunwoo Jang, Jing Yu Koh, Ruslan Salakhutdinov\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.16748v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-16748/"}
{"id": "arxiv:2606.16802", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7487f31097bc8de90dfbe836051d47f407c26e31a42bd751bedfd38cf4f1742c", "submitted_at": "2026-06-15T14:42:33Z", "updated_at": "2026-06-15T14:42:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16802v1"}, "authors": ["Anqi Zou", "Han Deng", "Chengyu Zhang", "Junquan Hu", "Yu Wang", "Yuxiang Xing", "Aokai Zhang", "Hanling Zhang", "Zhaoyang Liu", "Ben Fei", "Zhihui Wang", "Wanli Ouyang"], "canonical_source_url": "https://arxiv.org/abs/2606.16802", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-16802/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7487f31097bc8de90dfbe836051d47f407c26e31a42bd751bedfd38cf4f1742c", "submitted_at": "2026-06-15T14:42:33Z", "updated_at": "2026-06-15T14:42:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.16802v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.16802"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.16802"}], "name": "LabOSBench: Benchmarking Computer Use Agents for Scientific Instrument Control", "publication": null, "published_at": "2026-06-15", "record_sha256": "a9a56d5dabc4e7b26c54905a34f501ed384e7dd3f0624b09047d82ab4f825919", "relations": [], "resource_id": "arxiv:2606.16802", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.16802v1", "topics": ["evaluation", "science", "long-horizon"], "updated_at": "2026-06-15"}, "text": "LabOSBench: Benchmarking Computer Use Agents for Scientific Instrument Control\nResource type: paper\nReview status: curated\nScope: core\n\nLabOSBench evaluates GUI agents on browser-based scientific instrument simulators with configurable tasks and execution-based checks. Its subtask and end-to-end tests examine feedback-driven adjustment and long workflows without controlling physical instruments.\nPlatforms: web\nTopics: evaluation, science, long-horizon\nPublished: 2026-06-15\nAuthors: Anqi Zou, Han Deng, Chengyu Zhang, Junquan Hu, Yu Wang, Yuxiang Xing, Aokai Zhang, Hanling Zhang, Zhaoyang Liu, Ben Fei, Zhihui Wang, Wanli Ouyang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.16802v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-16802/"}
{"id": "arxiv:2606.17321", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7e6e0b66f2923cb66ab7aa404bedf9f4e1c73ae00df2e9fbfa7a5c3795cac4c", "submitted_at": "2026-06-15T22:04:11Z", "updated_at": "2026-06-15T22:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17321v1"}, "authors": ["Jaehun Jung", "Ximing Lu", "Brandon Cui", "Muhammad Khalifa", "Shaokun Zhang", "Hao Zhang", "Jin Xu", "Amala Sanjay Deshmukh", "Karan Sapra", "Andrew Tao", "Yejin Choi", "Jan Kautz", "Mingjie Liu", "Yi Dong"], "canonical_source_url": "https://arxiv.org/abs/2606.17321", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-17321/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d7e6e0b66f2923cb66ab7aa404bedf9f4e1c73ae00df2e9fbfa7a5c3795cac4c", "submitted_at": "2026-06-15T22:04:11Z", "updated_at": "2026-06-15T22:04:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17321v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.17321"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.17321"}], "name": "ProCUA-SFT Technical Report", "publication": null, "published_at": "2026-06-15", "record_sha256": "dedf79737f62155aafa6f633418e9d664c8e983cf08ee4a86595af884c527eb8", "relations": [], "resource_id": "arxiv:2606.17321", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.17321v1", "topics": ["data-generation", "imitation-learning", "evaluation"], "updated_at": "2026-06-15"}, "text": "ProCUA-SFT Technical Report\nResource type: paper\nReview status: curated\nScope: core\n\nProCUA-SFT generates tasks and synthetic trajectories on live desktops seeded with document content, checking task preconditions before rollout. The resulting step-prefix examples are formatted to match inference context and used for supervised GUI-agent fine-tuning.\nPlatforms: desktop\nTopics: data-generation, imitation-learning, evaluation\nPublished: 2026-06-15\nAuthors: Jaehun Jung, Ximing Lu, Brandon Cui, Muhammad Khalifa, Shaokun Zhang, Hao Zhang, Jin Xu, Amala Sanjay Deshmukh, Karan Sapra, Andrew Tao, Yejin Choi, Jan Kautz, Mingjie Liu, Yi Dong\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.17321v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-17321/"}
{"id": "arxiv:2606.17929", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "19fd8d8363ef6a7c58e42309f25dea1b9c5b6bca09b5d823f8a7cc7a80d61f0f", "submitted_at": "2026-06-16T13:40:21Z", "updated_at": "2026-06-16T13:40:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17929v1"}, "authors": ["Bojie Li"], "canonical_source_url": "https://arxiv.org/abs/2606.17929", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-17929/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "19fd8d8363ef6a7c58e42309f25dea1b9c5b6bca09b5d823f8a7cc7a80d61f0f", "submitted_at": "2026-06-16T13:40:21Z", "updated_at": "2026-06-16T13:40:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.17929v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.17929"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.17929"}], "name": "PreAct: Computer-Using Agents that Get Faster on Repeated Tasks", "publication": null, "published_at": "2026-06-16", "record_sha256": "294239690a41f6e427c6da25bd39959d55cf187cbe46af9e1062dd72d1aba7e8", "relations": [], "resource_id": "arxiv:2606.17929", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.17929v1", "topics": ["recording", "memory", "efficiency", "automation"], "updated_at": "2026-06-16"}, "text": "PreAct: Computer-Using Agents that Get Faster on Repeated Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nPreAct compiles successful screen-interaction runs into guarded state-machine programs for repeated tasks. It validates programs from clean states before storage and returns control to the agent when replayed screen expectations fail, with tests across mobile, desktop and web.\nPlatforms: cross-platform\nTopics: recording, memory, efficiency, automation\nPublished: 2026-06-16\nAuthors: Bojie Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.17929v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-17929/"}
{"id": "arxiv:2606.18101", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b935d8993be0d03ae778375f908b68b3ec3c614a7406b6b3f6a00f2a89ef1f", "submitted_at": "2026-06-16T16:02:41Z", "updated_at": "2026-06-17T08:58:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.18101v2"}, "authors": ["Jingyuan Huang", "Zuming Huang", "Yucheng Shi", "Tianze Yang", "Xiaoming Zhai", "Wei Chu", "Ninghao Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.18101", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-18101/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "49b935d8993be0d03ae778375f908b68b3ec3c614a7406b6b3f6a00f2a89ef1f", "submitted_at": "2026-06-16T16:02:41Z", "updated_at": "2026-06-17T08:58:52Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.18101v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.18101"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18101"}], "name": "Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding", "publication": null, "published_at": "2026-06-16", "record_sha256": "ec21053e0c88d89a4a264c2d485dd3f694eeb6fb48a70bee5947c9a84149bf58", "relations": [], "resource_id": "arxiv:2606.18101", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.18101v2", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": "2026-06-17"}, "text": "Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nProposes quality-aware self-distillation for GUI coordinate generation under student-produced prefixes. A correctness-aware gate suppresses coordinate-token supervision that cannot reach the target box, while teacher-probability scaling adjusts the remaining signal strength.\nPlatforms: unspecified\nTopics: grounding, imitation-learning, evaluation\nPublished: 2026-06-16\nAuthors: Jingyuan Huang, Zuming Huang, Yucheng Shi, Tianze Yang, Xiaoming Zhai, Wei Chu, Ninghao Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.18101v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-18101/"}
{"id": "arxiv:2606.18448", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2605cedb4359aa0cf1e694455903871ed334dd37b7b8e15edf31bac9d5e51b13", "submitted_at": "2026-06-16T19:57:07Z", "updated_at": "2026-06-16T19:57:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18448v1"}, "authors": ["Ziyan Jiang", "Li An", "Yujian Liu", "Jiabao Ji", "Qiucheng Wu", "Jacob Andreas", "Yang Zhang", "Shiyu Chang"], "canonical_source_url": "https://arxiv.org/abs/2606.18448", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-18448/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2605cedb4359aa0cf1e694455903871ed334dd37b7b8e15edf31bac9d5e51b13", "submitted_at": "2026-06-16T19:57:07Z", "updated_at": "2026-06-16T19:57:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18448v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.18448"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18448"}], "name": "VISUALSKILL: Multimodal Skills for Computer-Use Agents", "publication": null, "published_at": "2026-06-16", "record_sha256": "2ecd265ce3836bac212d857886695c7fab98612d5a6d27a659a2cc8d36111e17", "relations": [], "resource_id": "arxiv:2606.18448", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.18448v1", "topics": ["memory", "tool-integration", "screen-understanding", "long-horizon"], "updated_at": "2026-06-16"}, "text": "VISUALSKILL: Multimodal Skills for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nVISUALSKILL organizes application-specific text and figures into hierarchical skills retrieved on demand through an MCP tool. Skills combine authored documentation with live UI exploration, and matched multimodal/text-only comparisons test their utility for interface identification and workflow-state verification.\nPlatforms: unspecified\nTopics: memory, tool-integration, screen-understanding, long-horizon\nPublished: 2026-06-16\nAuthors: Ziyan Jiang, Li An, Yujian Liu, Jiabao Ji, Qiucheng Wu, Jacob Andreas, Yang Zhang, Shiyu Chang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.18448v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-18448/"}
{"id": "arxiv:2606.18671", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c52695e860206d02735839853416965ec3d42fac134ebfa0e8f2f5f5ea043914", "submitted_at": "2026-06-17T04:13:20Z", "updated_at": "2026-06-17T04:13:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18671v1"}, "authors": ["Yujin Zhang", "Daye Nam"], "canonical_source_url": "https://arxiv.org/abs/2606.18671", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-18671/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c52695e860206d02735839853416965ec3d42fac134ebfa0e8f2f5f5ea043914", "submitted_at": "2026-06-17T04:13:20Z", "updated_at": "2026-06-17T04:13:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18671v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.18671"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18671"}], "name": "HANSEL: Extracting Breadcrumbs from Web Agent Trajectories for Interactive Verification", "publication": null, "published_at": "2026-06-17", "record_sha256": "6e666e4247e3376c2ab896d5dcd9ed8a263ad9bee5f5eabe1d24ec25f1bfedb9", "relations": [], "resource_id": "arxiv:2606.18671", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.18671v1", "topics": ["human-in-the-loop", "observability", "recording", "evaluation"], "updated_at": "2026-06-17"}, "text": "HANSEL: Extracting Breadcrumbs from Web Agent Trajectories for Interactive Verification\nResource type: paper\nReview status: curated\nScope: core\n\nHANSEL extracts evidence pages and snippets from browser-agent traces and preserves relevant page state in interactive verification views. It flags untraceable answers and evaluates evidence selection and human verification effort separately.\nPlatforms: web\nTopics: human-in-the-loop, observability, recording, evaluation\nPublished: 2026-06-17\nAuthors: Yujin Zhang, Daye Nam\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.18671v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-18671/"}
{"id": "arxiv:2606.18890", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1929dfb5074e159d83006c7bbb470237a0c557e0723ac631116aa5e0369f9247", "submitted_at": "2026-06-17T10:07:51Z", "updated_at": "2026-06-17T10:07:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18890v1"}, "authors": ["Zhimin Fan", "Hongwei Yu", "Yeqing Shen", "Haolong Yan", "Guozhen Peng", "Tianhao Peng", "Yudong Zhang", "Xiaowen Zhang", "Kaijun Tan", "Zheng Ge", "Xiangyu Zhang", "Daxin Jiang"], "canonical_source_url": "https://arxiv.org/abs/2606.18890", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-18890/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1929dfb5074e159d83006c7bbb470237a0c557e0723ac631116aa5e0369f9247", "submitted_at": "2026-06-17T10:07:51Z", "updated_at": "2026-06-17T10:07:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.18890v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.18890"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.18890"}], "name": "Skill-Guided Continuation Distillation for GUI Agents", "publication": null, "published_at": "2026-06-17", "record_sha256": "6054c25cd17d9b12e3f5134017da0e825c089d2b311da4f70c3ff741b6e615f1", "relations": [], "resource_id": "arxiv:2606.18890", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.18890v1", "topics": ["imitation-learning", "data-generation", "robustness", "planning"], "updated_at": "2026-06-17"}, "text": "Skill-Guided Continuation Distillation for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSkill-Guided Continuation Distillation first lets a policy reach its own off-trajectory states, then uses skill guidance to collect successful continuations. Mixing those continuations with expert traces supplies supervision for recovery states absent from demonstrations.\nPlatforms: unspecified\nTopics: imitation-learning, data-generation, robustness, planning\nPublished: 2026-06-17\nAuthors: Zhimin Fan, Hongwei Yu, Yeqing Shen, Haolong Yan, Guozhen Peng, Tianhao Peng, Yudong Zhang, Xiaowen Zhang, Kaijun Tan, Zheng Ge, Xiangyu Zhang, Daxin Jiang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.18890v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-18890/"}
{"id": "arxiv:2606.19388", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bf8b07e08b275d90fcbfc63d867f54f5579509132cc9cf694d01447b7f91b78", "submitted_at": "2026-06-16T02:36:22Z", "updated_at": "2026-06-16T02:36:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19388v1"}, "authors": ["Li Gu", "Zihuan Jiang", "Linqiang Guo", "Zhixiang Chi", "Ziqiang Wang", "Huan Liu", "Yuanhao Yu", "Tse-Hsun Chen", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2606.19388", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-19388/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9bf8b07e08b275d90fcbfc63d867f54f5579509132cc9cf694d01447b7f91b78", "submitted_at": "2026-06-16T02:36:22Z", "updated_at": "2026-06-16T02:36:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19388v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.19388"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19388"}], "name": "Beyond the GUI Paradigm: Do Mobile Agents Need the Phone Screen?", "publication": null, "published_at": "2026-06-16", "record_sha256": "c3b0992dfaf99c9f9969d1e50d59b942972b89da027b43d29f4630cc1da44ddc", "relations": [], "resource_id": "arxiv:2606.19388", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2606.19388v1", "topics": ["evaluation", "automation", "cross-app", "tool-integration"], "updated_at": "2026-06-16"}, "text": "Beyond the GUI Paradigm: Do Mobile Agents Need the Phone Screen?\nResource type: paper\nReview status: curated\nScope: adjacent\n\nThis study compares general coding agents using mobile command-line access with reproducible GUI agents on AndroidWorld and MobileWorld. It also supplies oracle CLI analyses and a separate task suite for operations such as bulk changes, filtering and hidden device state.\nPlatforms: android\nTopics: evaluation, automation, cross-app, tool-integration\nPublished: 2026-06-16\nAuthors: Li Gu, Zihuan Jiang, Linqiang Guo, Zhixiang Chi, Ziqiang Wang, Huan Liu, Yuanhao Yu, Tse-Hsun Chen, Yang Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.19388v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-19388/"}
{"id": "arxiv:2606.19926", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbbb1bd57057e465ce9d991d3f6e2bb52112fc133a5846fc8cb6ce51eddc2121", "submitted_at": "2026-06-18T08:26:09Z", "updated_at": "2026-06-18T08:26:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19926v1"}, "authors": ["Guangyi Liu", "Gao Wu", "Congxiao Liu", "Pengxiang Zhao", "Liang Liu", "Mading Li", "Qi Zhang", "Mengyan Wang", "Liang Guo", "Yong Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.19926", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-19926/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cbbb1bd57057e465ce9d991d3f6e2bb52112fc133a5846fc8cb6ce51eddc2121", "submitted_at": "2026-06-18T08:26:09Z", "updated_at": "2026-06-18T08:26:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.19926v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.19926"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19926"}], "name": "MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management", "publication": null, "published_at": "2026-06-18", "record_sha256": "b5bae45c3519e539e5cfb02bd9d7a6c4a7fe6271c178c3d1a0ccb8f1fb866ed5", "relations": [], "resource_id": "arxiv:2606.19926", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.19926v1", "topics": ["memory", "interaction-history", "long-horizon", "imitation-learning", "data-generation"], "updated_at": "2026-06-18"}, "text": "MemGUI-Agent: An End-to-End Long-Horizon Mobile GUI Agent with Proactive Context Management\nResource type: paper\nReview status: curated\nScope: core\n\nMemGUI-Agent makes context management an action chosen by the same policy that controls the mobile UI. Structured history, UI-state and recent-step fields are trained from annotated trajectories to retain cross-app facts while limiting context growth.\nPlatforms: unspecified\nTopics: memory, interaction-history, long-horizon, imitation-learning, data-generation\nPublished: 2026-06-18\nAuthors: Guangyi Liu, Gao Wu, Congxiao Liu, Pengxiang Zhao, Liang Liu, Mading Li, Qi Zhang, Mengyan Wang, Liang Guo, Yong Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.19926v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-19926/"}
{"id": "arxiv:2606.19930", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cab809618ea4e218da26b4ae11bf937859c3a4bc240c91dd27b3a0d333737069", "submitted_at": "2026-06-18T08:29:33Z", "updated_at": "2026-08-07T03:54:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.19930v2"}, "authors": ["Guangyi Liu", "Pengxiang Zhao", "Gao Wu", "Yiwen Yin", "Mading Li", "Liang Liu", "Congxiao Liu", "Zhang Qi", "Mengyan Wang", "Liang Guo", "Jiangning Zhang", "Yong Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.19930", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-19930/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cab809618ea4e218da26b4ae11bf937859c3a4bc240c91dd27b3a0d333737069", "submitted_at": "2026-06-18T08:29:33Z", "updated_at": "2026-08-07T03:54:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.19930v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.19930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.19930"}], "name": "MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization", "publication": null, "published_at": "2026-06-18", "record_sha256": "06b69a857f76d6bce58b316b15842f4d617d9add80b1a797b5fadfb03fadbc1f", "relations": [], "resource_id": "arxiv:2606.19930", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.19930v2", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation", "evaluation"], "updated_at": "2026-08-07"}, "text": "MobileForge: Annotation-Free Adaptation for Mobile GUI Agents with Hierarchical Feedback-Guided Policy Optimization\nResource type: paper\nReview status: curated\nScope: core\n\nMobileForge couples real-app exploration and generated tasks with hierarchical feedback for mobile-agent adaptation. Its optimizer turns trajectory outcomes, step feedback and corrective hints into contextualized GRPO updates, with evaluation on AndroidWorld and a GUI-only MobileWorld split.\nPlatforms: android\nTopics: mobile-navigation, reinforcement-learning, data-generation, evaluation\nPublished: 2026-06-18\nAuthors: Guangyi Liu, Pengxiang Zhao, Gao Wu, Yiwen Yin, Mading Li, Liang Liu, Congxiao Liu, Zhang Qi, Mengyan Wang, Liang Guo, Jiangning Zhang, Yong Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.19930v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-19930/"}
{"id": "arxiv:2606.20363", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b370491b37d5bf25c624f2bac654e438459c8084b5da292fb18a0bf9142adaa2", "submitted_at": "2026-06-18T15:25:42Z", "updated_at": "2026-06-18T15:25:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20363v1"}, "authors": ["Yuexing Hao", "Xiaomin Li"], "canonical_source_url": "https://arxiv.org/abs/2606.20363", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-20363/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b370491b37d5bf25c624f2bac654e438459c8084b5da292fb18a0bf9142adaa2", "submitted_at": "2026-06-18T15:25:42Z", "updated_at": "2026-06-18T15:25:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20363v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.20363"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20363"}], "name": "Automating SKILL.md Generation for Computer-Using Agents via Interaction Trajectory Mining", "publication": null, "published_at": "2026-06-18", "record_sha256": "383a2021d98ddb5b9881bd7e519932eaac187b1761f95f325ab56c4bea704439", "relations": [], "resource_id": "arxiv:2606.20363", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.20363v1", "topics": ["memory", "data-generation", "reinforcement-learning", "evaluation"], "updated_at": "2026-06-18"}, "text": "Automating SKILL.md Generation for Computer-Using Agents via Interaction Trajectory Mining\nResource type: paper\nReview status: curated\nScope: core\n\nSegments and clusters GUI trajectories into inspectable skills, then trains a skill-aware policy from those annotations. Its diagnostic results distinguish readable source-domain clusters from reliable transfer and report limited or absent downstream gains.\nPlatforms: unspecified\nTopics: memory, data-generation, reinforcement-learning, evaluation\nPublished: 2026-06-18\nAuthors: Yuexing Hao, Xiaomin Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.20363v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-20363/"}
{"id": "arxiv:2606.20622", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e761b20caa6882605ecfd6cd001aeda4a2d36b2e702eea9a6d70451fc0ea9eb", "submitted_at": "2026-05-26T13:32:14Z", "updated_at": "2026-05-26T13:32:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20622v1"}, "authors": ["Daniel Beechey", "Derek Yuen", "Jianheng Liu", "Dezhao Luo", "Tiantian He", "Weilin Luo", "Jun Wang", "Kun Shao"], "canonical_source_url": "https://arxiv.org/abs/2606.20622", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-20622/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9e761b20caa6882605ecfd6cd001aeda4a2d36b2e702eea9a6d70451fc0ea9eb", "submitted_at": "2026-05-26T13:32:14Z", "updated_at": "2026-05-26T13:32:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20622v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:51.952206+00:00", "role": "primary", "sha256": "5cd7fb95d5ee8e8c473e0c9d5b892bf50ee425604c6a320e8e9a12eff17120a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=300&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:14:00.181958+00:00", "role": "primary", "sha256": "e4711466886372760daa83384cb4d129d54fdcfe320be6e91fb7b4971d1a61de", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2606.20622v1"}], "identifiers": {"arxiv": "2606.20622"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:51.952206+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20622"}], "name": "Darwin Mobile Agent: A Roadmap for Self-Evolution", "publication": null, "published_at": "2026-05-26", "record_sha256": "f9cb5564f0b0fcf17e27ba1ba1d90fd93bcf14e144e25905ea197bb043821887", "relations": [], "resource_id": "arxiv:2606.20622", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.20622v1", "topics": ["reinforcement-learning", "mobile-navigation", "evaluation"], "updated_at": "2026-05-26"}, "text": "Darwin Mobile Agent: A Roadmap for Self-Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes asynchronous reinforcement-learning infrastructure using parallel cloud-hosted Android instances, explicit task lifecycles, and persistent-state management for mobile GUI interaction.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation, evaluation\nPublished: 2026-05-26\nAuthors: Daniel Beechey, Derek Yuen, Jianheng Liu, Dezhao Luo, Tiantian He, Weilin Luo, Jun Wang, Kun Shao\nRecorded arXiv version: v1\n\nThe implemented stage uses selected tasks, an external judge and heuristic history management. Autonomous curricula, verification and memory are roadmap directions, not completed capabilities established by this record. No runtime or results were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (2 Agent-Environment Interface and Policy Optimisation; 3 Scalable Asynchronous Mobile Environment; 3.2 Persistent State Management; 3.3 Task Lifecycle; 4.2 Rollout and Optimisation; 4.3 Anatomy of the Self-Evolving Agent); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2606.20622v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-20622/"}
{"id": "arxiv:2606.20717", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fea441692131ec677e4a9ae54f9ecef1f31d422e20884b496b124e99392e59c1", "submitted_at": "2026-06-16T15:31:33Z", "updated_at": "2026-10-15T04:29:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20717v2"}, "authors": ["Xuelong Dai", "Jianyu Ma", "Boyang Ma", "Biwei Yan", "Yijun Yang", "Yue Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.20717", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-20717/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fea441692131ec677e4a9ae54f9ecef1f31d422e20884b496b124e99392e59c1", "submitted_at": "2026-06-16T15:31:33Z", "updated_at": "2026-10-15T04:29:19Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20717v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.20717"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20717"}], "name": "MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents", "publication": null, "published_at": "2026-06-16", "record_sha256": "8c3231b76d8331a3a8779e5602feec0ab429f70dce20d57378fa2f109232fbfa", "relations": [], "resource_id": "arxiv:2606.20717", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.20717v2", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-10-15"}, "text": "MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMIRAGE evaluates targeted next-action hijacking from adversarial images restricted to legitimate third-party webpage regions. It combines diffusion-based image generation with execution-aware optimization and tests the constrained visual-injection setting on web-agent frameworks.\nPlatforms: web\nTopics: safety-security, robustness, evaluation\nPublished: 2026-06-16\nAuthors: Xuelong Dai, Jianyu Ma, Boyang Ma, Biwei Yan, Yijun Yang, Yue Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.20717v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-20717/"}
{"id": "arxiv:2606.20785", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "414eced85da9f76828e4398b977f9b8dac77236c97d1594108c8f5dfc3be7cbb", "submitted_at": "2026-06-18T17:53:03Z", "updated_at": "2026-07-22T17:47:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20785v2"}, "authors": ["Ahmed Awadallah", "Sahil Gupta", "Yash Lara", "Yadong Lu", "Hussein Mozannar", "Akshay Nambi", "Zach Nussbaum", "Yash Pandya", "Aravind Rajeswaran", "Corby Rosset", "Alexey Taymanov", "Luiz do Valle", "Vibhav Vineet", "Spencer Whitehead", "Andrew Zhao"], "canonical_source_url": "https://arxiv.org/abs/2606.20785", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-20785/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "414eced85da9f76828e4398b977f9b8dac77236c97d1594108c8f5dfc3be7cbb", "submitted_at": "2026-06-18T17:53:03Z", "updated_at": "2026-07-22T17:47:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.20785v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.20785"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20785"}], "name": "Fara-1.5: Scalable Learning Environments for Computer Use Agents", "publication": null, "published_at": "2026-06-18", "record_sha256": "16ab490c447050c6e4c5a11949887975a6f552e6e64e136c9e37bbd3f58e7863", "relations": [], "resource_id": "arxiv:2606.20785", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.20785v2", "topics": ["data-generation", "web-navigation", "evaluation"], "updated_at": "2026-07-22"}, "text": "Fara-1.5: Scalable Learning Environments for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds browser-agent training trajectories with live and synthetic environments, solvers, user simulation, and complementary correctness and efficiency verifiers.\nPlatforms: web\nTopics: data-generation, web-navigation, evaluation\nPublished: 2026-06-18\nAuthors: Ahmed Awadallah, Sahil Gupta, Yash Lara, Yadong Lu, Hussein Mozannar, Akshay Nambi, Zach Nussbaum, Yash Pandya, Aravind Rajeswaran, Corby Rosset, Alexey Taymanov, Luiz do Valle, Vibhav Vineet, Spencer Whitehead, Andrew Zhao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.20785v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-20785/"}
{"id": "arxiv:2606.20910", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b2c6080f0de11c06fe73c30195c7c8529e07de283dc45fa47709072846d385a", "submitted_at": "2026-06-18T20:07:39Z", "updated_at": "2026-06-18T20:07:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20910v1"}, "authors": ["Dayeon Kang", "Hyejun Jeong", "Jade Sheffey", "Pubali Datta", "Amir Houmansadr"], "canonical_source_url": "https://arxiv.org/abs/2606.20910", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-20910/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8b2c6080f0de11c06fe73c30195c7c8529e07de283dc45fa47709072846d385a", "submitted_at": "2026-06-18T20:07:39Z", "updated_at": "2026-06-18T20:07:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.20910v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.20910"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.20910"}], "name": "Whose Agent Are You? Multi-Layer Fingerprinting and Attribution of Autonomous Web Agents", "publication": null, "published_at": "2026-06-18", "record_sha256": "2cbbf1b7961baba9528b82117eb99b1ccdedfa25bce96900f4885d3b83368f6b", "relations": [], "resource_id": "arxiv:2606.20910", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.20910v1", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2026-06-18"}, "text": "Whose Agent Are You? Multi-Layer Fingerprinting and Attribution of Autonomous Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds a logging framework that combines network/protocol fingerprints with browser interaction behavior to attribute web-agent traffic. A decision-tree classifier is evaluated across six agent implementations, human browsing and conventional crawlers.\nPlatforms: web\nTopics: safety-security, observability, evaluation\nPublished: 2026-06-18\nAuthors: Dayeon Kang, Hyejun Jeong, Jade Sheffey, Pubali Datta, Amir Houmansadr\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.20910v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-20910/"}
{"id": "arxiv:2606.21654", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69e668a9491a31e8d8849bab8cc63eede9878fff10ccb0ae0a211610e90819b3", "submitted_at": "2026-06-19T18:00:14Z", "updated_at": "2026-08-19T17:29:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.21654v2"}, "authors": ["Vincent Siu", "Manasi Sharma", "Dawn Song", "Daniel Yue Zhang", "Chenguang Wang", "Ying Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.21654", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-21654/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69e668a9491a31e8d8849bab8cc63eede9878fff10ccb0ae0a211610e90819b3", "submitted_at": "2026-06-19T18:00:14Z", "updated_at": "2026-08-19T17:29:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.21654v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.21654"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.21654"}], "name": "ChainWorld: Composing Long-Horizon Desktop Workloads from Atomic OSWorld Tasks", "publication": null, "published_at": "2026-06-19", "record_sha256": "c37f1dffa249700e6f3267903d7ff2ed1b50db4ae5072f51be64bcc17ddc690f", "relations": [], "resource_id": "arxiv:2606.21654", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.21654v2", "topics": ["evaluation", "long-horizon", "state-tracking", "workflow"], "updated_at": "2026-08-19"}, "text": "ChainWorld: Composing Long-Horizon Desktop Workloads from Atomic OSWorld Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nChainWorld composes compatible atomic OSWorld tasks into multi-objective desktop chains while preserving their evaluators. It compares giving the full sequence at once with revealing goals turn by turn, exposing different state and session-management failures.\nPlatforms: desktop\nTopics: evaluation, long-horizon, state-tracking, workflow\nPublished: 2026-06-19\nAuthors: Vincent Siu, Manasi Sharma, Dawn Song, Daniel Yue Zhang, Chenguang Wang, Ying Liu\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.21654v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-21654/"}
{"id": "arxiv:2606.22557", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dc52f58bb2237d12041fa8b994b20bd1e41290aaab0c70d2c5dd79fcaca37b6", "submitted_at": "2026-06-21T15:32:57Z", "updated_at": "2026-06-21T15:32:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22557v1"}, "authors": ["Yikun Fu", "Bowen Fu", "Zhenyu Wu", "Shuang Cheng", "Xiaowei Sun", "Bowen Yang", "Zehao Li", "Yibo Zhao", "Zichen Ding", "Zhoumianze Liu", "Shijie Wang", "Biqing Qi", "Bowen Zhou"], "canonical_source_url": "https://arxiv.org/abs/2606.22557", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-22557/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dc52f58bb2237d12041fa8b994b20bd1e41290aaab0c70d2c5dd79fcaca37b6", "submitted_at": "2026-06-21T15:32:57Z", "updated_at": "2026-06-21T15:32:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22557v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}], "identifiers": {"arxiv": "2606.22557"}, "kind": "paper", "last_verified_at": "2026-10-16T06:32:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.22557"}], "name": "MacAgentBench: Benchmarking AI Agents on Real-World macOS Desktop", "publication": null, "published_at": "2026-06-21", "record_sha256": "cf4e59e9f9bfd906c33da205e8a8e8e8c7411debfe43fdff24337d33b03e23e3", "relations": [], "resource_id": "arxiv:2606.22557", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.22557v1", "topics": ["evaluation", "cross-app", "tool-integration"], "updated_at": "2026-06-21"}, "text": "MacAgentBench: Benchmarking AI Agents on Real-World macOS Desktop\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces macOS tasks spanning GUI and command-line interaction, with deterministic checks and intermediate-progress scoring for application workflows.\nPlatforms: desktop, macos\nTopics: evaluation, cross-app, tool-integration\nPublished: 2026-06-21\nAuthors: Yikun Fu, Bowen Fu, Zhenyu Wu, Shuang Cheng, Xiaowei Sun, Bowen Yang, Zehao Li, Yibo Zhao, Zichen Ding, Zhoumianze Liu, Shijie Wang, Biqing Qi, Bowen Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.22557v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-22557/"}
{"id": "arxiv:2606.22948", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbd8892e06a5e847c16aa3101b98b3d669a783b84ee8e3d4d4d810754914028b", "submitted_at": "2026-06-22T07:28:55Z", "updated_at": "2026-06-22T07:28:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22948v1"}, "authors": ["Yincheng Zhou", "Athena Zhuoming Zhong", "Shijie Zhang", "Kevin Zhang", "Teresa Xiaotao Shang", "Shanghang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.22948", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-22948/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fbd8892e06a5e847c16aa3101b98b3d669a783b84ee8e3d4d4d810754914028b", "submitted_at": "2026-06-22T07:28:55Z", "updated_at": "2026-06-22T07:28:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.22948v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.22948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.22948"}], "name": "ENVS: Environment-Native Verified Search for Long-Horizon GUI Agents", "publication": null, "published_at": "2026-06-22", "record_sha256": "c863ee2adabad3fe16196ded8ba2ab8f65aae54ed28904a58fa0935008d2da43", "relations": [], "resource_id": "arxiv:2606.22948", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.22948v1", "topics": ["data-generation", "reinforcement-learning", "long-horizon", "robustness", "evaluation"], "updated_at": "2026-06-22"}, "text": "ENVS: Environment-Native Verified Search for Long-Horizon GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nENVS branches over GUI actions in live desktop VMs, verifies successful leaves, and converts them into balanced step-level training supervision. It also introduces OSWorld-Noisy to test recovery from dynamic interruptions while retaining the underlying tasks.\nPlatforms: desktop\nTopics: data-generation, reinforcement-learning, long-horizon, robustness, evaluation\nPublished: 2026-06-22\nAuthors: Yincheng Zhou, Athena Zhuoming Zhong, Shijie Zhang, Kevin Zhang, Teresa Xiaotao Shang, Shanghang Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.22948v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-22948/"}
{"id": "arxiv:2606.23189", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da8190bc5afc0bae400fc8ab5e7d80a7f34488c3400dce4477ca61635b19d5b7", "submitted_at": "2026-06-22T11:36:58Z", "updated_at": "2026-09-10T18:49:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.23189v2"}, "authors": ["Anmol Goel", "Iryna Gurevych"], "canonical_source_url": "https://arxiv.org/abs/2606.23189", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-23189/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da8190bc5afc0bae400fc8ab5e7d80a7f34488c3400dce4477ca61635b19d5b7", "submitted_at": "2026-06-22T11:36:58Z", "updated_at": "2026-09-10T18:49:54Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.23189v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.23189"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.23189"}], "name": "Capable but Careless: Do Computer-Use Agents Follow Contextual Integrity?", "publication": null, "published_at": "2026-06-22", "record_sha256": "b75d0aff9628d73d7b7fe9921e929df0e923714cadaf1cf55a802ceee5907687", "relations": [], "resource_id": "arxiv:2606.23189", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.23189v2", "topics": ["evaluation", "safety-security", "cross-app"], "updated_at": "2026-09-10"}, "text": "Capable but Careless: Do Computer-Use Agents Follow Contextual Integrity?\nResource type: paper\nReview status: curated\nScope: core\n\nAgentCIBench evaluates whether computer-use agents disclose information in ways that violate its original context. Deterministically scored scenarios test nearby sensitive UI content, ambiguous requests and recipient mismatches, including end-to-end agent execution.\nPlatforms: unspecified\nTopics: evaluation, safety-security, cross-app\nPublished: 2026-06-22\nAuthors: Anmol Goel, Iryna Gurevych\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.23189v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-23189/"}
{"id": "arxiv:2606.24515", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0aeb8c28b570133f575c2f336b6dab0dcbecdfaec78549115407b21722816e8a", "submitted_at": "2026-06-23T12:46:29Z", "updated_at": "2026-06-23T12:46:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24515v1"}, "authors": ["Marta Sumyk", "Oleksandr Kosovan"], "canonical_source_url": "https://arxiv.org/abs/2606.24515", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-24515/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0aeb8c28b570133f575c2f336b6dab0dcbecdfaec78549115407b21722816e8a", "submitted_at": "2026-06-23T12:46:29Z", "updated_at": "2026-06-23T12:46:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24515v1"}}, "environments": ["desktop", "macos", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.24515"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24515"}], "name": "Reinforcement Learning for Computer-Use Agents with Autonomous Evaluation", "publication": null, "published_at": "2026-06-23", "record_sha256": "6e34baa110134b041614c96b1db1f280775d85edffdc36fefefd81d29e5f6662", "relations": [], "resource_id": "arxiv:2606.24515", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.24515v1", "topics": ["reinforcement-learning", "reward-models", "evaluation", "robustness"], "updated_at": "2026-06-23"}, "text": "Reinforcement Learning for Computer-Use Agents with Autonomous Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nThe framework uses a vision-language judge’s final-screen assessment as a terminal reward for GUI reinforcement learning. It models evaluator noise and corrects the reward estimator used by PPO across desktop benchmarks.\nPlatforms: desktop, macos, windows\nTopics: reinforcement-learning, reward-models, evaluation, robustness\nPublished: 2026-06-23\nAuthors: Marta Sumyk, Oleksandr Kosovan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.24515v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-24515/"}
{"id": "arxiv:2606.24525", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3d0341f41a8ecba57fb7ed9e4b266bb4a176639c742be2b4fc98555754f5744", "submitted_at": "2026-06-23T12:57:02Z", "updated_at": "2026-06-23T12:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24525v1"}, "authors": ["Jiachen Qian"], "canonical_source_url": "https://arxiv.org/abs/2606.24525", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-24525/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a3d0341f41a8ecba57fb7ed9e4b266bb4a176639c742be2b4fc98555754f5744", "submitted_at": "2026-06-23T12:57:02Z", "updated_at": "2026-06-23T12:57:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24525v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.24525"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24525"}], "name": "VisCritic: Visual State Comparison as Process Reward for GUI Agents", "publication": null, "published_at": "2026-06-23", "record_sha256": "266f73094c88e62e96f8b9666ece7baa54b9e97b982be636289cf9cf9b31f04c", "relations": [], "resource_id": "arxiv:2606.24525", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.24525v1", "topics": ["reward-models", "state-tracking", "data-generation", "evaluation"], "updated_at": "2026-06-23"}, "text": "VisCritic: Visual State Comparison as Process Reward for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nVisCritic compares before/after screenshots through change-aware visual features and an action-conditioned critic. Weakly supervised samples from existing traces train judgments of action success, progress and error type for GUI-agent verification.\nPlatforms: unspecified\nTopics: reward-models, state-tracking, data-generation, evaluation\nPublished: 2026-06-23\nAuthors: Jiachen Qian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.24525v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-24525/"}
{"id": "arxiv:2606.24551", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f777f05b1c169ee4ad7ae7d77e9d1fc80a0696e9a1f6eefb5028a5c3e6fe5d", "submitted_at": "2026-06-22T17:05:07Z", "updated_at": "2026-06-22T17:05:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24551v1"}, "authors": ["Xiao Zhou", "Siyue Zhang", "Yilun Zhao", "Jinbiao Wei", "Tingyu Song", "Arman Cohan", "Chen Zhao"], "canonical_source_url": "https://arxiv.org/abs/2606.24551", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-24551/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "77f777f05b1c169ee4ad7ae7d77e9d1fc80a0696e9a1f6eefb5028a5c3e6fe5d", "submitted_at": "2026-06-22T17:05:07Z", "updated_at": "2026-06-22T17:05:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.24551v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.24551"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.24551"}], "name": "GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents", "publication": null, "published_at": "2026-06-22", "record_sha256": "6181db7aa83335e1664d5958239de8562ada116e7e7e410139ff6dc739a1a825", "relations": [], "resource_id": "arxiv:2606.24551", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.24551v1", "topics": ["evaluation", "tool-integration", "robustness", "workflow"], "updated_at": "2026-06-22"}, "text": "GUI vs. CLI: Execution Bottlenecks in Screen-Only and Skill-Mediated Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis benchmark matches desktop goals, initial states and final-state verifiers across screen-only GUI and skill-mediated CLI agents. Its experiments separate graphical execution failures from missing CLI skill coverage, including a verifier-guided skill-augmentation condition.\nPlatforms: desktop\nTopics: evaluation, tool-integration, robustness, workflow\nPublished: 2026-06-22\nAuthors: Xiao Zhou, Siyue Zhang, Yilun Zhao, Jinbiao Wei, Tingyu Song, Arman Cohan, Chen Zhao\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.24551v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-24551/"}
{"id": "arxiv:2606.25705", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c5a2d0e461b011e2d5b9dbd73012d007261d8df37bd1d687fa3bd86dc6f2246", "submitted_at": "2026-06-24T11:18:01Z", "updated_at": "2026-06-24T11:18:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25705v1"}, "authors": ["Aradhana Nayak", "Mussadiq Nazeer", "Wang Peng", "Feng Liu"], "canonical_source_url": "https://arxiv.org/abs/2606.25705", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-25705/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c5a2d0e461b011e2d5b9dbd73012d007261d8df37bd1d687fa3bd86dc6f2246", "submitted_at": "2026-06-24T11:18:01Z", "updated_at": "2026-06-24T11:18:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.25705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25705"}], "name": "GUI agent: Guided Exploration of User-Sensitive Screens", "publication": null, "published_at": "2026-06-24", "record_sha256": "83037e7c0411d721783a5ec03e459963c890484e28b4b05077891f161ff9cb97", "relations": [], "resource_id": "arxiv:2606.25705", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.25705v1", "topics": ["safety-security", "human-in-the-loop", "data-generation"], "updated_at": "2026-06-24"}, "text": "GUI agent: Guided Exploration of User-Sensitive Screens\nResource type: paper\nReview status: curated\nScope: core\n\nThis short paper develops an explorer agent that expands from a demonstrated GUI task to discover requests likely to reach user-sensitive screens. The proposed exploration supports identifying situations where task execution should be handed back to the user.\nPlatforms: unspecified\nTopics: safety-security, human-in-the-loop, data-generation\nPublished: 2026-06-24\nAuthors: Aradhana Nayak, Mussadiq Nazeer, Wang Peng, Feng Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.25705v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-25705/"}
{"id": "arxiv:2606.25760", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "628d08bdbf23ae9e599bf767f0503cecedcde72dfbfcc183f9ea9f044c6f2d44", "submitted_at": "2026-06-24T12:34:28Z", "updated_at": "2026-09-29T20:55:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.25760v2"}, "authors": ["Divake Kumar", "Sina Tayebati", "Devashri Naik", "Amanda Sofie Rios", "Nilesh Ahuja", "Omesh Tickoo", "Ranganath Krishnan", "Amit Ranjan Trivedi"], "canonical_source_url": "https://arxiv.org/abs/2606.25760", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-25760/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "628d08bdbf23ae9e599bf767f0503cecedcde72dfbfcc183f9ea9f044c6f2d44", "submitted_at": "2026-06-24T12:34:28Z", "updated_at": "2026-09-29T20:55:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.25760v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.25760"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25760"}], "name": "Uncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets", "publication": null, "published_at": "2026-06-24", "record_sha256": "9c0c7afcadc3b55377b8696a08b2c644f086955dbcc9965458cab8d64f94f3aa", "relations": [], "resource_id": "arxiv:2606.25760", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.25760v2", "topics": ["grounding", "evaluation", "robustness", "safety-security"], "updated_at": "2026-09-29"}, "text": "Uncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets\nResource type: paper\nReview status: curated\nScope: core\n\nArgus benchmarks post-hoc uncertainty methods for single-step GUI grounding across open-weight and closed-source observation regimes. It tests calibration, rejection, error severity and conformal click regions, finding that uncertainty rankings transfer less reliably across model/interface regimes than across datasets for one model.\nPlatforms: unspecified\nTopics: grounding, evaluation, robustness, safety-security\nPublished: 2026-06-24\nAuthors: Divake Kumar, Sina Tayebati, Devashri Naik, Amanda Sofie Rios, Nilesh Ahuja, Omesh Tickoo, Ranganath Krishnan, Amit Ranjan Trivedi\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.25760v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-25760/"}
{"id": "arxiv:2606.25964", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a8efe4b9476d7f55ca4da17621ece653be2e33bf02be0183a339a12f5fac3ca", "submitted_at": "2026-06-24T15:35:29Z", "updated_at": "2026-06-24T15:35:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25964v1"}, "authors": ["Chengheng Li-Chen", "Zhiqian Zhou", "Hao Chen", "Nicolas Chauvin"], "canonical_source_url": "https://arxiv.org/abs/2606.25964", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-25964/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7a8efe4b9476d7f55ca4da17621ece653be2e33bf02be0183a339a12f5fac3ca", "submitted_at": "2026-06-24T15:35:29Z", "updated_at": "2026-06-24T15:35:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.25964v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.25964"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.25964"}], "name": "WinDOM: Self-Family Distillation for Small-Model GUI Grounding", "publication": null, "published_at": "2026-06-24", "record_sha256": "740df38caf1e56953cc7a6d541a09eea9773e51e65d1c00b3a17fea56c72554c", "relations": [], "resource_id": "arxiv:2606.25964", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.25964v1", "topics": ["grounding", "data-generation", "imitation-learning", "reinforcement-learning"], "updated_at": "2026-06-24"}, "text": "WinDOM: Self-Family Distillation for Small-Model GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nWinDOM obtains bounding-box supervision from a Windows 11 web reimplementation driven through Playwright, and studies rejection-sampling distillation for small GUI grounders. It treats the depth of the supervised cold start as a GRPO choice, comparing same-size EMA and larger same-family teachers.\nPlatforms: web\nTopics: grounding, data-generation, imitation-learning, reinforcement-learning\nPublished: 2026-06-24\nAuthors: Chengheng Li-Chen, Zhiqian Zhou, Hao Chen, Nicolas Chauvin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.25964v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-25964/"}
{"id": "arxiv:2606.27330", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42166ac3980296a53b56bc4a076634beec36ba885b3260b5d36115b5a5261cd0", "submitted_at": "2026-06-25T17:44:48Z", "updated_at": "2026-06-25T17:44:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.27330v1"}, "authors": ["Tianyi Men", "Zhuoran Jin", "Pengfei Cao", "Yubo Chen", "Kang Liu", "Jun Zhao"], "canonical_source_url": "https://arxiv.org/abs/2606.27330", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-27330/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "42166ac3980296a53b56bc4a076634beec36ba885b3260b5d36115b5a5261cd0", "submitted_at": "2026-06-25T17:44:48Z", "updated_at": "2026-06-25T17:44:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.27330v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.27330"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.27330"}], "name": "Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning", "publication": null, "published_at": "2026-06-25", "record_sha256": "84688dac4f84e36f61617d6426ceb75dfb9777d340433c8618dd04a026a470ec", "relations": [], "resource_id": "arxiv:2606.27330", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.27330v1", "topics": ["planning", "data-generation", "imitation-learning", "web-navigation"], "updated_at": "2026-06-25"}, "text": "Empowering GUI Agents via Autonomous Experience Exploration and Hindsight Experience Utilization for Task Planning\nResource type: paper\nReview status: curated\nScope: core\n\nPEEU explores web environments and turns hindsight experience into aligned high-level planning supervision. A hierarchical task-decomposition analysis examines why atomic-skill competence alone does not ensure planning generalization across websites.\nPlatforms: web\nTopics: planning, data-generation, imitation-learning, web-navigation\nPublished: 2026-06-25\nAuthors: Tianyi Men, Zhuoran Jin, Pengfei Cao, Yubo Chen, Kang Liu, Jun Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.27330v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-27330/"}
{"id": "arxiv:2606.28480", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6164c42541650e5120adc881b19524cb3b268f08d6eb77fbe91d8c1a02823786", "submitted_at": "2026-06-26T17:59:51Z", "updated_at": "2026-06-26T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.28480v1"}, "authors": ["Shoufa Chen", "Luyuan Wang", "Xuan Yang", "Zhiheng Liu", "Yuren Cong", "Yuanfeng Ji", "Feiyan Zhou", "Xiaohui Zhang", "Fanny Yang", "Belinda Zeng"], "canonical_source_url": "https://arxiv.org/abs/2606.28480", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-28480/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6164c42541650e5120adc881b19524cb3b268f08d6eb77fbe91d8c1a02823786", "submitted_at": "2026-06-26T17:59:51Z", "updated_at": "2026-06-26T17:59:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.28480v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:28Z", "role": "primary", "sha256": "7eaa9a5430d3a45049e612a117da345c1daafc54277a35a7cf56b42fbb1e8618", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2606.05646%2C2606.05670%2C2606.06556%2C2606.06891%2C2606.07367%2C2606.07594%2C2606.07682%2C2606.07723%2C2606.07867%2C2606.10394%2C2606.10525%2C2606.11290%2C2606.12203%2C2606.12736%2C2606.13040%2C2606.13368%2C2606.13578%2C2606.13669%2C2606.13802%2C2606.13898%2C2606.14397%2C2606.15057%2C2606.15300%2C2606.15441%2C2606.17162%2C2606.18425%2C2606.19336%2C2606.19409%2C2606.19704%2C2606.20950%2C2606.21033%2C2606.21565%2C2606.23521%2C2606.23654%2C2606.24775%2C2606.25084%2C2606.26080%2C2606.27406%2C2606.27634%2C2606.28430%2C2606.28480%2C2606.29955%2C2606.30616%2C2607.00692%2C2607.01084%2C2607.02032%2C2607.02448%2C2607.02490%2C2607.02577%2C2607.02911&max_results=50"}], "identifiers": {"arxiv": "2606.28480"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.28480"}], "name": "TUA-Bench: A Benchmark for General-Purpose Terminal-Use Agents", "publication": null, "published_at": "2026-06-26", "record_sha256": "1c996c92ffa47e30f9450e4b968b70f3dda76187e312f0233101be841ffad31b", "relations": [], "resource_id": "arxiv:2606.28480", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2606.28480v1", "topics": ["evaluation", "automation"], "updated_at": "2026-06-26"}, "text": "TUA-Bench: A Benchmark for General-Purpose Terminal-Use Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nTUA-Bench evaluates terminal-use agents on general digital tasks, including document editing, email, web information seeking and specialist software workflows. Tasks run in real terminals with deterministic setup and execution-based scoring.\nPlatforms: unspecified\nTopics: evaluation, automation\nPublished: 2026-06-26\nAuthors: Shoufa Chen, Luyuan Wang, Xuan Yang, Zhiheng Liu, Yuren Cong, Yuanfeng Ji, Feiyan Zhou, Xiaohui Zhang, Fanny Yang, Belinda Zeng\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified. Adjacent general computer-operation benchmark: the evaluated interface is a terminal; GUI benchmarks are a contrast, not its execution modality.\n\nPrimary source: https://arxiv.org/abs/2606.28480v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-28480/"}
{"id": "arxiv:2606.29472", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54de4e93f4f3e39d904b4661a193765e02c29f8ec7787264aee5616579c32210", "submitted_at": "2026-06-28T15:59:31Z", "updated_at": "2026-06-28T15:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29472v1"}, "authors": ["Bojie Li", "Noah Shi"], "canonical_source_url": "https://arxiv.org/abs/2606.29472", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-29472/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "54de4e93f4f3e39d904b4661a193765e02c29f8ec7787264aee5616579c32210", "submitted_at": "2026-06-28T15:59:31Z", "updated_at": "2026-06-28T15:59:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29472v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.29472"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29472"}], "name": "Agent-Computer Observation Interfaces Enable Dynamic Computer Use", "publication": null, "published_at": "2026-06-28", "record_sha256": "dd26da8a5d11fade605e48fc0259cbd1eb717ef3652b11a7a9ff3c1347002871", "relations": [], "resource_id": "arxiv:2606.29472", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.29472v1", "topics": ["evaluation", "video", "screen-understanding", "memory", "training-free"], "updated_at": "2026-06-28"}, "text": "Agent-Computer Observation Interfaces Enable Dynamic Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nThe Agent-Computer Observation Interface separates continuous perception from discrete computer actions using captured frames, audio transcription and persistent visual narration. DynaCU-Bench tests these components on dynamic browser tasks and static controls, with model-specific ablations showing that additional observations can also hurt.\nPlatforms: web\nTopics: evaluation, video, screen-understanding, memory, training-free\nPublished: 2026-06-28\nAuthors: Bojie Li, Noah Shi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.29472v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-29472/"}
{"id": "arxiv:2606.29537", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03035fcbc6cd3e6bf40f244870fb35694ba985ff7f809da66e2a376cbf99d66b", "submitted_at": "2026-06-28T17:59:17Z", "updated_at": "2026-07-13T10:30:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.29537v2"}, "authors": ["Mengqi Yuan", "Zilong Zhou", "Xinzhuang Xiong", "Weiming Wu", "Jiayang Sun", "Jiamin Song", "Kaiqian Cui", "Bowen Wang", "Haoyuan Wu", "Yitong Li", "Dunjie Lu", "Haikong Lu", "Qi Zhen", "Xinyuan Wang", "Jiaqi Deng", "Yuhao Yang", "Cheng Chen", "Boyuan Zheng", "Alex Su", "Xiao Yu", "Hao Zou", "Saaket Agashe", "Xing Han Lu", "Manpreet Kaur", "Zhengyang Qi", "Vincent Sunn Chen", "Frederic Sala", "Dayiheng Liu", "Junyang Lin", "Zhou Yu", "Yu Su", "Siva Reddy", "Xin Eric Wang", "Peng Qi", "Tianbao Xie", "Tao Yu"], "canonical_source_url": "https://arxiv.org/abs/2606.29537", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-29537/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03035fcbc6cd3e6bf40f244870fb35694ba985ff7f809da66e2a376cbf99d66b", "submitted_at": "2026-06-28T17:59:17Z", "updated_at": "2026-07-13T10:30:17Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.29537v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.29537"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29537"}], "name": "OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks", "publication": null, "published_at": "2026-06-28", "record_sha256": "f3e9e41680dc0e801b85320d9c24f9c1feaf008a190f635804b9f2298abd3859", "relations": [], "resource_id": "arxiv:2606.29537", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.29537v2", "topics": ["evaluation", "long-horizon", "workflow"], "updated_at": "2026-07-13"}, "text": "OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces extended computer-use workflows with realistic artifacts, changing information, hidden state, and partial-progress and safety assessments.\nPlatforms: desktop\nTopics: evaluation, long-horizon, workflow\nPublished: 2026-06-28\nAuthors: Mengqi Yuan, Zilong Zhou, Xinzhuang Xiong, Weiming Wu, Jiayang Sun, Jiamin Song, Kaiqian Cui, Bowen Wang, Haoyuan Wu, Yitong Li, Dunjie Lu, Haikong Lu, Qi Zhen, Xinyuan Wang, Jiaqi Deng, Yuhao Yang, Cheng Chen, Boyuan Zheng, Alex Su, Xiao Yu, Hao Zou, Saaket Agashe, Xing Han Lu, Manpreet Kaur, Zhengyang Qi, Vincent Sunn Chen, Frederic Sala, Dayiheng Liu, Junyang Lin, Zhou Yu, Yu Su, Siva Reddy, Xin Eric Wang, Peng Qi, Tianbao Xie, Tao Yu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.29537v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-29537/"}
{"id": "arxiv:2606.29705", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7beb70ac91463b77a8babaf2b077c46c5072e8f86d22f4d9afc9c1e77680f309", "submitted_at": "2026-06-29T02:16:21Z", "updated_at": "2026-06-29T02:16:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29705v1"}, "authors": ["Sunqi Fan", "Lingshan Chen", "Runqi Yin", "Qingle Liu", "Yongming Rao", "Meng-Hao Guo", "Shi-Min Hu"], "canonical_source_url": "https://arxiv.org/abs/2606.29705", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-29705/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7beb70ac91463b77a8babaf2b077c46c5072e8f86d22f4d9afc9c1e77680f309", "submitted_at": "2026-06-29T02:16:21Z", "updated_at": "2026-06-29T02:16:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.29705v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.29705"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.29705"}], "name": "GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots", "publication": null, "published_at": "2026-06-29", "record_sha256": "327ce86c28b22c8c8367c6585dc594b5f246aa46532d657092432fb097d02442", "relations": [], "resource_id": "arxiv:2606.29705", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.29705v1", "topics": ["grounding", "data-generation", "reinforcement-learning", "screen-understanding"], "updated_at": "2026-06-29"}, "text": "GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots\nResource type: paper\nReview status: curated\nScope: core\n\nGUICrafter first learns grounding from large collections of unannotated screenshots and webpages, then uses a smaller high-quality dataset for reinforcement-learning calibration. The curriculum targets reduced dependence on manually annotated GUI training data.\nPlatforms: unspecified\nTopics: grounding, data-generation, reinforcement-learning, screen-understanding\nPublished: 2026-06-29\nAuthors: Sunqi Fan, Lingshan Chen, Runqi Yin, Qingle Liu, Yongming Rao, Meng-Hao Guo, Shi-Min Hu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.29705v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-29705/"}
{"id": "arxiv:2606.30084", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50e44e08750d4152f9b5408885f8a6a016c5f789af9712e447ee147fb25adb8d", "submitted_at": "2026-06-29T10:20:39Z", "updated_at": "2026-06-29T10:20:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30084v1"}, "authors": ["Chen Liu", "Ling Chen", "Hanzhang Zhou", "Liangyu Chen", "Chenglin Cai", "Xin Yu", "Steven Hoi", "Yue Wang"], "canonical_source_url": "https://arxiv.org/abs/2606.30084", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-30084/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50e44e08750d4152f9b5408885f8a6a016c5f789af9712e447ee147fb25adb8d", "submitted_at": "2026-06-29T10:20:39Z", "updated_at": "2026-06-29T10:20:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30084v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.30084"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30084"}], "name": "One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding", "publication": null, "published_at": "2026-06-29", "record_sha256": "37bdeb3b77fecc94c1972d3ecc3c3d5a698d6b582c44b1f39dc8a80f6994a78c", "relations": [], "resource_id": "arxiv:2606.30084", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.30084v1", "topics": ["grounding", "efficiency", "screen-understanding", "evaluation"], "updated_at": "2026-06-29"}, "text": "One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nInnerZoom preserves target-region cues from intermediate decoder layers and reinjects a compact evidence state during coordinate decoding. Its single-forward GUI grounder is compared with crop-and-rerun zooming and matched supervised/RL baselines for accuracy and inference cost.\nPlatforms: unspecified\nTopics: grounding, efficiency, screen-understanding, evaluation\nPublished: 2026-06-29\nAuthors: Chen Liu, Ling Chen, Hanzhang Zhou, Liangyu Chen, Chenglin Cai, Xin Yu, Steven Hoi, Yue Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.30084v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-30084/"}
{"id": "arxiv:2606.30119", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "770c901e3490b2a43b02296136825e8e6bd035fb7494ea74b3a9af74be1afed1", "submitted_at": "2026-06-29T10:51:13Z", "updated_at": "2026-06-29T10:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30119v1"}, "authors": ["Iliana Fayolle", "Sihem Bouhenniche", "Samuel Pélissier", "Pierre Laperdrix", "Clémentine Maurice", "Walter Rudametkin"], "canonical_source_url": "https://arxiv.org/abs/2606.30119", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-30119/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "770c901e3490b2a43b02296136825e8e6bd035fb7494ea74b3a9af74be1afed1", "submitted_at": "2026-06-29T10:51:13Z", "updated_at": "2026-06-29T10:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30119v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.30119"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30119"}], "name": "On the Internet, Nobody Knows You're an LLM Bot: Unmasking Web Agents with Multi-Layer Fingerprinting", "publication": null, "published_at": "2026-06-29", "record_sha256": "caefb6eb9e26cb0fe2b5c077a9018c294a99a2143b89080b7a1802613d3a05d2", "relations": [], "resource_id": "arxiv:2606.30119", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.30119v1", "topics": ["safety-security", "observability", "evaluation"], "updated_at": "2026-06-29"}, "text": "On the Internet, Nobody Knows You're an LLM Bot: Unmasking Web Agents with Multi-Layer Fingerprinting\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates web agents against anti-bot mechanisms on instrumented honeysites and analyzes network, HTTP and browser fingerprints. The study compares resistance to blocking with the ability to distinguish agents, humans and agent implementations.\nPlatforms: web\nTopics: safety-security, observability, evaluation\nPublished: 2026-06-29\nAuthors: Iliana Fayolle, Sihem Bouhenniche, Samuel Pélissier, Pierre Laperdrix, Clémentine Maurice, Walter Rudametkin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.30119v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-30119/"}
{"id": "arxiv:2606.30697", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f3f2ae1438986c11146fec508dd61fd3e848ddcbf830f86a7dfa759280d730a", "submitted_at": "2026-06-29T03:19:40Z", "updated_at": "2026-06-29T03:19:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30697v1"}, "authors": ["Yogeswar Reddy Thota"], "canonical_source_url": "https://arxiv.org/abs/2606.30697", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-30697/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f3f2ae1438986c11146fec508dd61fd3e848ddcbf830f86a7dfa759280d730a", "submitted_at": "2026-06-29T03:19:40Z", "updated_at": "2026-06-29T03:19:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.30697v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.30697"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.30697"}], "name": "LUMOS: A Semantic Operating-System Layer for Accessibility-Grounded AI Agents", "publication": null, "published_at": "2026-06-29", "record_sha256": "6e5c17802e87a2492462430807bf54a5181b734df5943e8bee468a28b101789b", "relations": [], "resource_id": "arxiv:2606.30697", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.30697v1", "topics": ["grounding", "screen-understanding", "tool-integration", "automation"], "updated_at": "2026-06-29"}, "text": "LUMOS: A Semantic Operating-System Layer for Accessibility-Grounded AI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nLUMOS converts native accessibility metadata and browser UI structure into semantic interface blueprints with stable element identifiers and available actions. It combines live pointer-to-element grounding with an observe-act loop constrained to visible-UI primitives.\nPlatforms: web\nTopics: grounding, screen-understanding, tool-integration, automation\nPublished: 2026-06-29\nAuthors: Yogeswar Reddy Thota\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.30697v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-30697/"}
{"id": "arxiv:2606.31270", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8911abdc0c9c481d82abb213d368c16fba9fee909c354e0bc6272e14c59abfb7", "submitted_at": "2026-06-30T07:44:37Z", "updated_at": "2026-06-30T07:44:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.31270v1"}, "authors": ["Xueqiao Sun", "Xiaohan Wang", "Ludwig Schmidt", "Serena Yeung-Levy", "Yuhui Zhang"], "canonical_source_url": "https://arxiv.org/abs/2606.31270", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-31270/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8911abdc0c9c481d82abb213d368c16fba9fee909c354e0bc6272e14c59abfb7", "submitted_at": "2026-06-30T07:44:37Z", "updated_at": "2026-06-30T07:44:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2606.31270v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:36Z", "role": "primary", "sha256": "aeacc1a41bb90f3d296823356174fe77d1547ffbe7cd56652cf9d98fca09206c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.31270"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:36Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31270"}], "name": "Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents", "publication": null, "published_at": "2026-06-30", "record_sha256": "66ec1694a0bba860f8ae3905f0ac0d47f5aa4c4ae213c6fba060c0a2f9b9aef6", "relations": [], "resource_id": "arxiv:2606.31270", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.31270v1", "topics": ["self-reflection", "robustness", "training-free", "evaluation"], "updated_at": "2026-06-30"}, "text": "Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses failed computer-use trajectories to diagnose failure modes and generate inference-time agent code patches, with light human verification. An OSWorld study evaluates this failure-driven improvement loop without additional model training.\nPlatforms: unspecified\nTopics: self-reflection, robustness, training-free, evaluation\nPublished: 2026-06-30\nAuthors: Xueqiao Sun, Xiaohan Wang, Ludwig Schmidt, Serena Yeung-Levy, Yuhui Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.31270v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-31270/"}
{"id": "arxiv:2606.31410", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "974d471d25373502914bf0d146d5d4a54a5751281c04921a9d331d83bce64b33", "submitted_at": "2026-06-30T09:36:35Z", "updated_at": "2026-07-01T03:59:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31410v2"}, "authors": ["Wanxia Cao", "Chengzhen Duan", "Pei Fu", "Pengzhi Gao", "Niu Lian", "Fazhan Liu", "Hui Liu", "Heng Qu", "Qinzhuo Wu", "Zhehao Yu", "Tongbo Chen", "Shiqi Cui", "Anan Du", "Shukai Jia", "Yuanfa Li", "Wei Liu", "Yike Liu", "Wenchao Lu", "Zhenbo Luo", "Haoyuan Sun", "Jiatong Sun", "Cheng Tan", "Yajie Wang", "Changqiao Wu", "Tao Xiong", "Jiahui Yang", "Yuxuan Yuan", "Ruoceng Zhang", "Shaojie Zhang", "Jian Zhu", "Jian Luan", "Cong Zou"], "canonical_source_url": "https://arxiv.org/abs/2606.31410", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-31410/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "974d471d25373502914bf0d146d5d4a54a5751281c04921a9d331d83bce64b33", "submitted_at": "2026-06-30T09:36:35Z", "updated_at": "2026-07-01T03:59:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31410v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:05Z", "role": "primary", "sha256": "2e5722527f8da4b2cf06372d1e2daccd33ef7fdc7151704b411a0bddd734beae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2606.10423%2C2606.10522%2C2606.10742%2C2606.11078%2C2606.11520%2C2606.12485%2C2606.12666%2C2606.13239%2C2606.13385%2C2606.14106%2C2606.14832%2C2606.15017%2C2606.15034%2C2606.16659%2C2606.16748%2C2606.16802%2C2606.17321%2C2606.17645%2C2606.17929%2C2606.18448%2C2606.18671%2C2606.18890%2C2606.19388%2C2606.19926%2C2606.19930%2C2606.20363%2C2606.20636%2C2606.20717%2C2606.20724%2C2606.20785%2C2606.21654%2C2606.22864%2C2606.22948%2C2606.23449%2C2606.24515%2C2606.24525%2C2606.25705%2C2606.25760%2C2606.27330%2C2606.27595%2C2606.29445%2C2606.29537%2C2606.29705%2C2606.30697%2C2606.31154%2C2606.31270%2C2606.31410%2C2606.31612%2C2607.00502%2C2607.04425&max_results=50"}], "identifiers": {"arxiv": "2606.31410"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31410"}], "name": "Xiaomi-GUI-0 Technical Report", "publication": null, "published_at": "2026-06-30", "record_sha256": "60bc69f2278d976cd885950de2dd9a042ec80575bee404693ca13a8a487f30da", "relations": [], "resource_id": "arxiv:2606.31410", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2606.31410v2", "topics": ["mobile-navigation", "reinforcement-learning", "robustness"], "updated_at": "2026-07-01"}, "text": "Xiaomi-GUI-0 Technical Report\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a mobile GUI agent trained through real-device interaction, corrected failure trajectories, supervised learning, and reinforcement learning.\nPlatforms: android\nTopics: mobile-navigation, reinforcement-learning, robustness\nPublished: 2026-06-30\nAuthors: Wanxia Cao, Chengzhen Duan, Pei Fu, Pengzhi Gao, Niu Lian, Fazhan Liu, Hui Liu, Heng Qu, Qinzhuo Wu, Zhehao Yu, Tongbo Chen, Shiqi Cui, Anan Du, Shukai Jia, Yuanfa Li, Wei Liu, Yike Liu, Wenchao Lu, Zhenbo Luo, Haoyuan Sun, Jiatong Sun, Cheng Tan, Yajie Wang, Changqiao Wu, Tao Xiong, Jiahui Yang, Yuxuan Yuan, Ruoceng Zhang, Shaojie Zhang, Jian Zhu, Jian Luan, Cong Zou\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.31410v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-31410/"}
{"id": "arxiv:2606.31980", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0a3e641e0a71051f5fbd20f57af3db06f2b65c61025e78adcdd122f773d034b", "submitted_at": "2026-06-30T17:18:47Z", "updated_at": "2026-09-01T04:17:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31980v2"}, "authors": ["Meng Chen", "Anya Ji", "Tsung-Han Wu", "Tobias Maringgele", "David M. Chan", "Alane Suhr", "Amy Pavel"], "canonical_source_url": "https://arxiv.org/abs/2606.31980", "citation_url": "https://awesomecua.github.io/resources/arxiv-2606-31980/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0a3e641e0a71051f5fbd20f57af3db06f2b65c61025e78adcdd122f773d034b", "submitted_at": "2026-06-30T17:18:47Z", "updated_at": "2026-09-01T04:17:49Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2606.31980v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2606.31980"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2606.31980"}], "name": "DigitalCoach: Communication and Grounding Gaps in Human and Agentic Computer Use Coaching", "publication": null, "published_at": "2026-06-30", "record_sha256": "e0a9d6ad4f46617efb94ae9166d75ed19491af6cd47ad648c65a68f9060e9c5c", "relations": [], "resource_id": "arxiv:2606.31980", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2606.31980v2", "topics": ["human-in-the-loop", "video", "evaluation", "grounding"], "updated_at": "2026-09-01"}, "text": "DigitalCoach: Communication and Grounding Gaps in Human and Agentic Computer Use Coaching\nResource type: paper\nReview status: curated\nScope: adjacent\n\nDigitalCoach records expert–novice software coaching conversations aligned with screen and input events. Automated and interactive evaluations examine how model coaches differ from humans in explanation, error diagnosis and visual grounding when helping a learner operate software.\nPlatforms: unspecified\nTopics: human-in-the-loop, video, evaluation, grounding\nPublished: 2026-06-30\nAuthors: Meng Chen, Anya Ji, Tsung-Han Wu, Tobias Maringgele, David M. Chan, Alane Suhr, Amy Pavel\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2606.31980v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2606-31980/"}
{"id": "arxiv:2607.00333", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd4d31e477d5915626d471682e52e5cd6083d9b0c67196f6de5de2a29c6a3058", "submitted_at": "2026-07-01T02:17:19Z", "updated_at": "2026-07-14T03:01:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.00333v2"}, "authors": ["Zidong Zhang", "Zhentao Xie", "Wenrui Diao", "Jianliang Wu"], "canonical_source_url": "https://arxiv.org/abs/2607.00333", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-00333/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cd4d31e477d5915626d471682e52e5cd6083d9b0c67196f6de5de2a29c6a3058", "submitted_at": "2026-07-01T02:17:19Z", "updated_at": "2026-07-14T03:01:43Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.00333v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2607.00333"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.00333"}], "name": "(A)I Sees What You Don't: Exploiting New Attack Surfaces in Third-Party Mobile Agents", "publication": null, "published_at": "2026-07-01", "record_sha256": "0d6b55cd17ca34cf33f05c2c0dfa54dfbc7e086696b00afb61ae248181aeb3fa", "relations": [], "resource_id": "arxiv:2607.00333", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.00333v2", "topics": ["safety-security", "mobile-navigation"], "updated_at": "2026-07-14"}, "text": "(A)I Sees What You Don't: Exploiting New Attack Surfaces in Third-Party Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes cross-app privilege escalation and visual attack surfaces introduced by third-party VLM-powered mobile agents.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation\nPublished: 2026-07-01\nAuthors: Zidong Zhang, Zhentao Xie, Wenrui Diao, Jianliang Wu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.00333v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-00333/"}
{"id": "arxiv:2607.04334", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fdae5d6e5ea17639fe0e2b8e2d655df7201a438d84acdeb3e0d02a78051ae2", "submitted_at": "2026-07-05T14:33:29Z", "updated_at": "2026-07-18T16:56:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04334v2"}, "authors": ["Guijia Zhang", "Yuxun Chen", "Yuheng Qi", "Harry Yang"], "canonical_source_url": "https://arxiv.org/abs/2607.04334", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-04334/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e7fdae5d6e5ea17639fe0e2b8e2d655df7201a438d84acdeb3e0d02a78051ae2", "submitted_at": "2026-07-05T14:33:29Z", "updated_at": "2026-07-18T16:56:38Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04334v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.04334"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.04334"}], "name": "Do GUI Agents Believe Their Eyes? Diagnosing State-Belief Reliance on Pixels versus Structure", "publication": null, "published_at": "2026-07-05", "record_sha256": "4f9f08ea4fd352affbd450d094a82dad760835ea8aa70af18e1e9c5529d800b7", "relations": [], "resource_id": "arxiv:2607.04334", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.04334v2", "topics": ["screen-understanding", "state-tracking", "evaluation", "robustness", "safety-security"], "updated_at": "2026-07-18"}, "text": "Do GUI Agents Believe Their Eyes? Diagnosing State-Belief Reliance on Pixels versus Structure\nResource type: paper\nReview status: curated\nScope: core\n\nDefines visual state reliance and probes whether GUI agents resolve screenshot/structure conflicts using pixels, serialized interface structure or prior beliefs. Paired interventions and live multi-step tests connect misattributed state beliefs to failures and compare consistency-based mitigations.\nPlatforms: cross-platform\nTopics: screen-understanding, state-tracking, evaluation, robustness, safety-security\nPublished: 2026-07-05\nAuthors: Guijia Zhang, Yuxun Chen, Yuheng Qi, Harry Yang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.04334v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-04334/"}
{"id": "arxiv:2607.04425", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cb3020ccf25c31b805dc44218f5f9496cbc310e1df2906d8b15e5b60c31278", "submitted_at": "2026-07-05T17:37:12Z", "updated_at": "2026-08-10T13:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04425v2"}, "authors": ["Niu Lian", "Tongbo Chen", "Zhehao Yu", "Chengzhen Duan", "Fazhan Liu", "Hui Liu", "Pei Fu", "Jian Luan", "Heng Qu", "Shu-Tao Xia", "Jinpeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2607.04425", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-04425/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "32cb3020ccf25c31b805dc44218f5f9496cbc310e1df2906d8b15e5b60c31278", "submitted_at": "2026-07-05T17:37:12Z", "updated_at": "2026-08-10T13:11:37Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.04425v2"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.04425"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.04425"}], "name": "UI-MOPD: Multi-Platform On-Policy Distillation for Unified GUI Agents", "publication": null, "published_at": "2026-07-05", "record_sha256": "d8452145b3317149ada6974b9e334b6b766e9369e1ea26a8bdfaf6deb5d9a3f3", "relations": [], "resource_id": "arxiv:2607.04425", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.04425v2", "topics": ["data-generation", "imitation-learning", "action-prediction"], "updated_at": "2026-08-10"}, "text": "UI-MOPD: Multi-Platform On-Policy Distillation for Unified GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Uni-GUI, a collection of executable desktop and mobile trajectories, and UI-MOPD, which trains a shared policy from its own rollouts using platform-specific teachers. The distillation procedure routes each visited state to the appropriate teacher to preserve distinct interaction conventions.\nPlatforms: desktop, cross-platform\nTopics: data-generation, imitation-learning, action-prediction\nPublished: 2026-07-05\nAuthors: Niu Lian, Tongbo Chen, Zhehao Yu, Chengzhen Duan, Fazhan Liu, Hui Liu, Pei Fu, Jian Luan, Heng Qu, Shu-Tao Xia, Jinpeng Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.04425v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-04425/"}
{"id": "arxiv:2607.05120", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5fb30b8be2c6e452471e6398720291ec59a2c5ef96e64baca0735fb01b7a12e", "submitted_at": "2026-07-06T14:07:49Z", "updated_at": "2026-07-06T14:07:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05120v1"}, "authors": ["Woohyuk Choi", "Juhee Kim", "Taehyun Kang", "Jihyeon Jeong", "Luyi Xing", "Byoungyoung Lee"], "canonical_source_url": "https://arxiv.org/abs/2607.05120", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-05120/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5fb30b8be2c6e452471e6398720291ec59a2c5ef96e64baca0735fb01b7a12e", "submitted_at": "2026-07-06T14:07:49Z", "updated_at": "2026-07-06T14:07:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05120v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "identifiers": {"arxiv": "2607.05120"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.05120"}], "name": "Agent Data Injection Attacks are Realistic Threats to AI Agents", "publication": null, "published_at": "2026-07-06", "record_sha256": "5818a09ca5e97887ab0bdfd2bddcae189c0c0af66d1137419811efcee8bd3512", "relations": [], "resource_id": "arxiv:2607.05120", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.05120v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-07-06"}, "text": "Agent Data Injection Attacks are Realistic Threats to AI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study examines agent data injection, where untrusted content impersonates trusted metadata or tool-context data. It reports arbitrary-click vulnerabilities in web agents alongside separate coding-agent attacks, highlighting trust-boundary failures beyond instruction injection.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2026-07-06\nAuthors: Woohyuk Choi, Juhee Kim, Taehyun Kang, Jihyeon Jeong, Luyi Xing, Byoungyoung Lee\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2607.05120v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-05120/"}
{"id": "arxiv:2607.05277", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97d6c625b6ebc08cec7471df543812a91c4252f50d8d35f8d51fbef802dcb5a2", "submitted_at": "2026-07-06T16:22:37Z", "updated_at": "2026-07-06T16:22:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05277v1"}, "authors": ["Kristina Nikolić", "Egor Zverev", "Javier Rando", "Matthew Jagielski", "Edoardo Debenedetti", "Florian Tramèr"], "canonical_source_url": "https://arxiv.org/abs/2607.05277", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-05277/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97d6c625b6ebc08cec7471df543812a91c4252f50d8d35f8d51fbef802dcb5a2", "submitted_at": "2026-07-06T16:22:37Z", "updated_at": "2026-07-06T16:22:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.05277v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.05277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.05277"}], "name": "Untrusted Content Masking for Web Agents with Security Guarantees", "publication": null, "published_at": "2026-07-06", "record_sha256": "324647dbe52daf6bea2e5c7dc2539fdb5145d077f49e8bae6fe782163c253ecd", "relations": [], "resource_id": "arxiv:2607.05277", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.05277v1", "topics": ["safety-security", "sandbox", "web-navigation"], "updated_at": "2026-07-06"}, "text": "Untrusted Content Masking for Web Agents with Security Guarantees\nResource type: paper\nReview status: curated\nScope: core\n\nUntrusted Content Masking uses DOM structure to redact untrusted webpage regions before agent observation. A sandboxed interaction interface separates privileges so browser actions can proceed under the proposed trust-boundary model without exposing protected content to the planner.\nPlatforms: web\nTopics: safety-security, sandbox, web-navigation\nPublished: 2026-07-06\nAuthors: Kristina Nikolić, Egor Zverev, Javier Rando, Matthew Jagielski, Edoardo Debenedetti, Florian Tramèr\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.05277v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-05277/"}
{"id": "arxiv:2607.06118", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc6d846f43b9992388e94f3bcfdfc9a25d09b317fea62336aa22951c8fce9f8b", "submitted_at": "2026-07-07T10:27:31Z", "updated_at": "2026-07-07T10:27:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.06118v1"}, "authors": ["Wei Dong", "Tianyu Fu", "Zhe Yu", "Hanning Wang", "Anyang Su", "Zhizhou Fang", "Yuyang Chen", "Shuo Wang", "Minghui Wu", "Ping Jiang", "Zhen Lei", "Chenxu Zhao"], "canonical_source_url": "https://arxiv.org/abs/2607.06118", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-06118/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc6d846f43b9992388e94f3bcfdfc9a25d09b317fea62336aa22951c8fce9f8b", "submitted_at": "2026-07-07T10:27:31Z", "updated_at": "2026-07-07T10:27:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.06118v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.06118"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.06118"}], "name": "WebRetriever: A Large-Scale Comprehensive Benchmark for Efficient Web Agent Evaluation", "publication": null, "published_at": "2026-07-07", "record_sha256": "a8d53caacaa1de2313e046e9f9370e3493dc2d0ebee0de418baf525cfa706593", "relations": [], "resource_id": "arxiv:2607.06118", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.06118v1", "topics": ["evaluation", "web-navigation", "observability"], "updated_at": "2026-07-07"}, "text": "WebRetriever: A Large-Scale Comprehensive Benchmark for Efficient Web Agent Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nWebRetriever evaluates navigation, knowledge-assisted interaction and end-to-end information-extraction workflows under separate protocols. Its NavEval judge uses richer interaction context to diagnose cases where navigation success does not translate into useful task completion.\nPlatforms: web\nTopics: evaluation, web-navigation, observability\nPublished: 2026-07-07\nAuthors: Wei Dong, Tianyu Fu, Zhe Yu, Hanning Wang, Anyang Su, Zhizhou Fang, Yuyang Chen, Shuo Wang, Minghui Wu, Ping Jiang, Zhen Lei, Chenxu Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.06118v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-06118/"}
{"id": "arxiv:2607.09773", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53c254c53a834d4f28bb4b65317121f5c46f5dd6985d03a30f1cc1f0fa78707d", "submitted_at": "2026-07-07T16:36:25Z", "updated_at": "2026-09-04T15:47:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.09773v2"}, "authors": ["Mianqiu Huang", "Taofeng Xue", "Chong Peng", "Jinrui Ding", "Jie Yang", "Sicheng Fan", "Jiale Hong", "Yufei Gao", "Xiaocheng Zhang", "Linsen Guo", "Xin Yang", "Dengchang Zhao", "Yuchen Xie", "Peng Pei", "Xunliang Xie", "Xipeng Qiu"], "canonical_source_url": "https://arxiv.org/abs/2607.09773", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-09773/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53c254c53a834d4f28bb4b65317121f5c46f5dd6985d03a30f1cc1f0fa78707d", "submitted_at": "2026-07-07T16:36:25Z", "updated_at": "2026-09-04T15:47:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.09773v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.09773"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.09773"}], "name": "EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents", "publication": null, "published_at": "2026-07-07", "record_sha256": "ef4d253857e031f72b444de1bb8a5da79a7b7b4bad0282e22adbfae9c6e19e0f", "relations": [], "resource_id": "arxiv:2607.09773", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.09773v2", "topics": ["reinforcement-learning", "long-horizon", "sandbox", "data-generation"], "updated_at": "2026-09-04"}, "text": "EvoCUA-1.5: Online Reinforcement Learning for Multi-turn Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvoCUA-1.5 trains multi-turn desktop agents through online reinforcement learning in executable sandboxes. It combines step-level policy optimization, task filtering and calibration, an adaptive curriculum, and asynchronous rollout handling with staleness control.\nPlatforms: desktop\nTopics: reinforcement-learning, long-horizon, sandbox, data-generation\nPublished: 2026-07-07\nAuthors: Mianqiu Huang, Taofeng Xue, Chong Peng, Jinrui Ding, Jie Yang, Sicheng Fan, Jiale Hong, Yufei Gao, Xiaocheng Zhang, Linsen Guo, Xin Yang, Dengchang Zhao, Yuchen Xie, Peng Pei, Xunliang Xie, Xipeng Qiu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.09773v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-09773/"}
{"id": "arxiv:2607.10079", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cfa436d182ed8fbc8823213b9173fbfcf9fb421e8d2ae89a093a5393c374ba", "submitted_at": "2026-07-11T02:08:46Z", "updated_at": "2026-07-16T03:28:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.10079v3"}, "authors": ["Chengguang Gan", "Hanjun Wei", "Yunhao Liang", "Zhixi Cai", "Qinghao Zhang", "Shiwen Ni"], "canonical_source_url": "https://arxiv.org/abs/2607.10079", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-10079/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37cfa436d182ed8fbc8823213b9173fbfcf9fb421e8d2ae89a093a5393c374ba", "submitted_at": "2026-07-11T02:08:46Z", "updated_at": "2026-07-16T03:28:43Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.10079v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.10079"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.10079"}], "name": "MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation", "publication": null, "published_at": "2026-07-11", "record_sha256": "a2b6f649b9d8ec36ea269d8253a696dfb56991217ea77c2bd24528133e2fcad3", "relations": [], "resource_id": "arxiv:2607.10079", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.10079v3", "topics": ["evaluation", "grounding", "reinforcement-learning", "action-prediction"], "updated_at": "2026-07-16"}, "text": "MAG: A Web-Agent Benchmark and Harness for Multimodal Action and Guide Generation\nResource type: paper\nReview status: curated\nScope: core\n\nMAG jointly evaluates browser task actions and user-facing guides grounded in screenshots through element marks or pixel coordinates. Its harness covers annotation, live evaluation and joint metrics, with an expert-trajectory-augmented GRPO training study.\nPlatforms: web\nTopics: evaluation, grounding, reinforcement-learning, action-prediction\nPublished: 2026-07-11\nAuthors: Chengguang Gan, Hanjun Wei, Yunhao Liang, Zhixi Cai, Qinghao Zhang, Shiwen Ni\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.10079v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-10079/"}
{"id": "arxiv:2607.10608", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0473ced7881ed2007aa49c2767860c33719d5bf1c798dfe4771a4061b13d655e", "submitted_at": "2026-07-12T07:05:31Z", "updated_at": "2026-07-12T07:05:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.10608v1"}, "authors": ["Yixiong Chen", "Xinyi Bai", "Alan Yuille"], "canonical_source_url": "https://arxiv.org/abs/2607.10608", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-10608/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0473ced7881ed2007aa49c2767860c33719d5bf1c798dfe4771a4061b13d655e", "submitted_at": "2026-07-12T07:05:31Z", "updated_at": "2026-07-12T07:05:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.10608v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "identifiers": {"arxiv": "2607.10608"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.10608"}], "name": "The Compliance Trap: Diagnosing How AI Agents Consume Conflicting Memory", "publication": null, "published_at": "2026-07-12", "record_sha256": "8864f1745f8d98e726b1233744c2b95e1d2f6323b047e4b9f13771469f01efb6", "relations": [], "resource_id": "arxiv:2607.10608", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.10608v1", "topics": ["memory", "safety-security", "robustness", "evaluation", "long-horizon"], "updated_at": "2026-07-12"}, "text": "The Compliance Trap: Diagnosing How AI Agents Consume Conflicting Memory\nResource type: paper\nReview status: curated\nScope: core\n\nThe Entry–Propagation–Recovery framework traces where conflicting memory first changes an agent action, how the effect propagates and whether recovery occurs. WebArena and the controlled MemTrapBench separate memory-consumption failures across these stages.\nPlatforms: web\nTopics: memory, safety-security, robustness, evaluation, long-horizon\nPublished: 2026-07-12\nAuthors: Yixiong Chen, Xinyi Bai, Alan Yuille\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2607.10608v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-10608/"}
{"id": "arxiv:2607.11185", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "554844f23cddbdaadc3c52e93d07d06c756faadbd0136ffef0984b25423f84f8", "submitted_at": "2026-07-13T07:32:35Z", "updated_at": "2026-07-13T07:32:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11185v1"}, "authors": ["Bowen Lv", "Xiao Liu", "Yanyu Ren", "Hanyu Lai", "Bohao Jing", "Hanchen Zhang", "Yanxiao Zhao", "Shuntian Yao", "Jie Tang", "Yuxiao Dong"], "canonical_source_url": "https://arxiv.org/abs/2607.11185", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-11185/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "554844f23cddbdaadc3c52e93d07d06c756faadbd0136ffef0984b25423f84f8", "submitted_at": "2026-07-13T07:32:35Z", "updated_at": "2026-07-13T07:32:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11185v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.11185"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.11185"}], "name": "SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL", "publication": null, "published_at": "2026-07-13", "record_sha256": "768008dd640036c0af69e3010e3ab4d394202e973d1198f0b3e4c2ce9b603fa5", "relations": [], "resource_id": "arxiv:2607.11185", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.11185v1", "topics": ["data-generation", "reinforcement-learning", "efficiency"], "updated_at": "2026-07-13"}, "text": "SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL\nResource type: paper\nReview status: curated\nScope: core\n\nScales computer-agent reinforcement learning through verifiable task generation, capability-aware rollout allocation, and segmented visual context.\nPlatforms: desktop\nTopics: data-generation, reinforcement-learning, efficiency\nPublished: 2026-07-13\nAuthors: Bowen Lv, Xiao Liu, Yanyu Ren, Hanyu Lai, Bohao Jing, Hanchen Zhang, Yanxiao Zhao, Shuntian Yao, Jie Tang, Yuxiao Dong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.11185v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-11185/"}
{"id": "arxiv:2607.11388", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40daa1b180a0c2eaba7b907a0f61b8dd8819cb3033b9da118162330d97cfecf7", "submitted_at": "2026-07-13T10:48:25Z", "updated_at": "2026-07-13T10:48:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11388v1"}, "authors": ["Wenyi Wu", "Sibo Zhu", "Kun Zhou", "Aayush Salvi", "Zixuan Song", "Biwei Huang"], "canonical_source_url": "https://arxiv.org/abs/2607.11388", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-11388/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "40daa1b180a0c2eaba7b907a0f61b8dd8819cb3033b9da118162330d97cfecf7", "submitted_at": "2026-07-13T10:48:25Z", "updated_at": "2026-07-13T10:48:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.11388v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "identifiers": {"arxiv": "2607.11388"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.11388"}], "name": "StructAgent: Harness Long-horizon Digital Agents with Unified Causal Structure", "publication": null, "published_at": "2026-07-13", "record_sha256": "7bbc0dc5a154da926e371e9c5f29a3f33d610114be967567cc740d4d910056e6", "relations": [], "resource_id": "arxiv:2607.11388", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.11388v1", "topics": ["state-tracking", "planning", "long-horizon", "robustness"], "updated_at": "2026-07-13"}, "text": "StructAgent: Harness Long-horizon Digital Agents with Unified Causal Structure\nResource type: paper\nReview status: curated\nScope: core\n\nStructAgent maintains compact task state and regulates long-horizon execution through verified transitions and progress checkpoints. Its workflow supports evidence-based completion and targeted recovery, with desktop computer-use evaluation and a separate Minecraft extension.\nPlatforms: desktop\nTopics: state-tracking, planning, long-horizon, robustness\nPublished: 2026-07-13\nAuthors: Wenyi Wu, Sibo Zhu, Kun Zhou, Aayush Salvi, Zixuan Song, Biwei Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.11388v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-11388/"}
{"id": "arxiv:2607.12625", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "281de2bb6e45a58efd047735accc2049deab0f8815931c86abbd38b86b6d23bf", "submitted_at": "2026-07-14T11:04:01Z", "updated_at": "2026-07-15T04:41:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.12625v2"}, "authors": ["Yunxin Li", "Jinchao Li", "Shibo Su", "Zhenran Xu", "Chenrui Zhao", "Tongshu Bian", "Xiaoman Liang", "Meishan Zhang", "Baotian Hu", "Min Zhang"], "canonical_source_url": "https://arxiv.org/abs/2607.12625", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-12625/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "281de2bb6e45a58efd047735accc2049deab0f8815931c86abbd38b86b6d23bf", "submitted_at": "2026-07-14T11:04:01Z", "updated_at": "2026-07-15T04:41:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.12625v2"}}, "environments": ["desktop", "android", "ios", "harmonyos", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "identifiers": {"arxiv": "2607.12625"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.12625"}], "name": "KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill", "publication": null, "published_at": "2026-07-14", "record_sha256": "329c515f7b01cba59aa7d2c4314eda42aa03c53da0febf15c95983ef5f7d3f13", "relations": [], "resource_id": "arxiv:2607.12625", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.12625v2", "topics": ["memory", "planning", "interaction-history", "mobile-navigation"], "updated_at": "2026-07-15"}, "text": "KnowAct-GUIClaw: Know Deeply, Act Perfectly, Personal GUI Assistant with Self-Evolving Memory and Skill\nResource type: paper\nReview status: curated\nScope: core\n\nKnowAct-GUIClaw combines a host planner with pluggable GUI subagents, attributable experience memory and an evolving skill library. It uses user profiles and feedback for task decomposition and evaluates transfer across Android, iOS, HarmonyOS and Windows.\nPlatforms: desktop, android, ios, harmonyos, windows\nTopics: memory, planning, interaction-history, mobile-navigation\nPublished: 2026-07-14\nAuthors: Yunxin Li, Jinchao Li, Shibo Su, Zhenran Xu, Chenrui Zhao, Tongshu Bian, Xiaoman Liang, Meishan Zhang, Baotian Hu, Min Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.12625v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-12625/"}
{"id": "arxiv:2607.12640", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e96ea0cf71ff705c064db1bcd1d2cbf8b8cd712bb4e8030931cafbb31ff29b10", "submitted_at": "2026-07-14T11:17:35Z", "updated_at": "2026-07-14T11:17:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.12640v1"}, "authors": ["Chengguang Gan", "Zhixi Cai", "Yunhao Liang", "Hanjun Wei", "Shiwen Ni", "Qinghao Zhang"], "canonical_source_url": "https://arxiv.org/abs/2607.12640", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-12640/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e96ea0cf71ff705c064db1bcd1d2cbf8b8cd712bb4e8030931cafbb31ff29b10", "submitted_at": "2026-07-14T11:17:35Z", "updated_at": "2026-07-14T11:17:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.12640v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.12640"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.12640"}], "name": "A Learning-Rate-Gated Failure of GRPO in a Small Language and Vision-Language Model Web Agent: A Controlled Null and Its Mechanism", "publication": null, "published_at": "2026-07-14", "record_sha256": "c0c29410b98b899014898998b7e4f6a0fc6c16a769a51e740f5ba22074fd0610", "relations": [], "resource_id": "arxiv:2607.12640", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.12640v1", "topics": ["reinforcement-learning", "evaluation", "robustness"], "updated_at": "2026-07-14"}, "text": "A Learning-Rate-Gated Failure of GRPO in a Small Language and Vision-Language Model Web Agent: A Controlled Null and Its Mechanism\nResource type: paper\nReview status: curated\nScope: core\n\nStudies a controlled setting where GRPO fails to improve a strong small web agent and can reduce text-track performance. Matched reachable-reward tasks and component grafting distinguish limited optimization headroom from learning-rate-dependent failure mechanisms.\nPlatforms: web\nTopics: reinforcement-learning, evaluation, robustness\nPublished: 2026-07-14\nAuthors: Chengguang Gan, Zhixi Cai, Yunhao Liang, Hanjun Wei, Shiwen Ni, Qinghao Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.12640v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-12640/"}
{"id": "arxiv:2607.14443", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a779b2c1ff9ba3b531d399204b0986e3affcea81c67fdd7951e3b7da343b0e17", "submitted_at": "2026-07-16T00:35:25Z", "updated_at": "2026-07-16T00:35:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14443v1"}, "authors": ["Yong Liu", "Zhenyi Zhong", "Zhanpeng Shi"], "canonical_source_url": "https://arxiv.org/abs/2607.14443", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-14443/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a779b2c1ff9ba3b531d399204b0986e3affcea81c67fdd7951e3b7da343b0e17", "submitted_at": "2026-07-16T00:35:25Z", "updated_at": "2026-07-16T00:35:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14443v1"}}, "environments": ["desktop", "macos"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.14443"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.14443"}], "name": "Tactile: Giving Computer-Using Agents Hands and Feet", "publication": null, "published_at": "2026-07-16", "record_sha256": "d2f89dba167d95153bd04412e8d6f3af6cb67f68b629cbd0a70036e70d861b12", "relations": [], "resource_id": "arxiv:2607.14443", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.14443v1", "topics": ["grounding", "automation", "observability", "tool-integration", "robustness", "evaluation"], "updated_at": "2026-07-16"}, "text": "Tactile: Giving Computer-Using Agents Hands and Feet\nResource type: paper\nReview status: curated\nScope: core\n\nTactile turns accessibility, OCR and visual evidence into explicit desktop targets with executable actions and verification cues. Its execution loop chooses semantic actions or coordinate fallbacks while retaining provenance for replay and diagnosis.\nPlatforms: desktop, macos\nTopics: grounding, automation, observability, tool-integration, robustness, evaluation\nPublished: 2026-07-16\nAuthors: Yong Liu, Zhenyi Zhong, Zhanpeng Shi\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.14443v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-14443/"}
{"id": "arxiv:2607.14548", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "151d4882fecc835dfffb9651bd860ebe8c85919a63ff5730e191f9146fbba272", "submitted_at": "2026-07-16T04:12:42Z", "updated_at": "2026-07-16T04:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14548v1"}, "authors": ["Hy Vision Team", "Huawen Shen", "Zhengyang Tang", "Shangpin Peng", "Liang Wu", "Anran Zhang", "Weinong Wang", "Yiduo Guo", "Chenxin Li", "Zhengyao Fang", "Yang Ding", "Junyi Li", "Fei Tang", "Zheng Ruan", "Yi Zhang", "Xingran Zhou", "Dingchen Yang", "Sunqi Fan", "Zhiyi Wan", "Han Hu", "Xin Lai", "Pengyuan Lyu", "Chengquan Zhang"], "canonical_source_url": "https://arxiv.org/abs/2607.14548", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-14548/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "151d4882fecc835dfffb9651bd860ebe8c85919a63ff5730e191f9146fbba272", "submitted_at": "2026-07-16T04:12:42Z", "updated_at": "2026-07-16T04:12:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.14548v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.14548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.14548"}], "name": "HyMobileAgent: Data-Environment Co-Scaling for Efficient GUI Agents", "publication": null, "published_at": "2026-07-16", "record_sha256": "0151250e6948b7c014829c08e471f57256fa866ef0eaf5887ad6c8072f2ec0de", "relations": [], "resource_id": "arxiv:2607.14548", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.14548v1", "topics": ["mobile-navigation", "data-generation", "planning"], "updated_at": "2026-07-16"}, "text": "HyMobileAgent: Data-Environment Co-Scaling for Efficient GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines GUI perception data, instructional videos, real and simulated devices, and planning with reflection to develop a mobile agent.\nPlatforms: unspecified\nTopics: mobile-navigation, data-generation, planning\nPublished: 2026-07-16\nAuthors: Hy Vision Team, Huawen Shen, Zhengyang Tang, Shangpin Peng, Liang Wu, Anran Zhang, Weinong Wang, Yiduo Guo, Chenxin Li, Zhengyao Fang, Yang Ding, Junyi Li, Fei Tang, Zheng Ruan, Yi Zhang, Xingran Zhou, Dingchen Yang, Sunqi Fan, Zhiyi Wan, Han Hu, Xin Lai, Pengyuan Lyu, Chengquan Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.14548v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-14548/"}
{"id": "arxiv:2607.15193", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1179c657434761187f3fa3bed99f46530a69878578b7a74c68febae20e64d6b3", "submitted_at": "2026-07-16T16:48:39Z", "updated_at": "2026-07-16T16:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.15193v1"}, "authors": ["Madhumitha Venkatesan", "Shicheng Wen", "Jiajing Guo", "Jorge Piazentin Ono", "Liu Ren", "Dongyu Liu"], "canonical_source_url": "https://arxiv.org/abs/2607.15193", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-15193/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1179c657434761187f3fa3bed99f46530a69878578b7a74c68febae20e64d6b3", "submitted_at": "2026-07-16T16:48:39Z", "updated_at": "2026-07-16T16:48:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.15193v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.15193"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.15193"}], "name": "Plover: Steering GUI Agents through Plan-Centric Interaction", "publication": null, "published_at": "2026-07-16", "record_sha256": "f0dbb37e83ef218317c5fa852a65229cfd626011a39d30e456f8a43646873bca", "relations": [], "resource_id": "arxiv:2607.15193", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.15193v1", "topics": ["planning", "human-in-the-loop", "observability", "workflow"], "updated_at": "2026-07-16"}, "text": "Plover: Steering GUI Agents through Plan-Centric Interaction\nResource type: paper\nReview status: curated\nScope: core\n\nPlover exposes GUI-agent plans and replanning as persistent artifacts that users can inspect and edit during execution. Its planner–executor design supports localized correction through plan edits, language guidance and screenshot-grounded interventions, studied through failure repair and workflow scenarios.\nPlatforms: unspecified\nTopics: planning, human-in-the-loop, observability, workflow\nPublished: 2026-07-16\nAuthors: Madhumitha Venkatesan, Shicheng Wen, Jiajing Guo, Jorge Piazentin Ono, Liu Ren, Dongyu Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.15193v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-15193/"}
{"id": "arxiv:2607.15550", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a15963a7fb4dfd488f0e79732d9ea0c9a03d43c1e3d5e50dd1ee8e00f1072be", "submitted_at": "2026-07-17T01:45:54Z", "updated_at": "2026-09-02T01:49:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.15550v3"}, "authors": ["Xue Yu", "Bo Yuan", "Kailin Zhao", "Pengshuai Yang", "Hong Hu", "Junlan Feng"], "canonical_source_url": "https://arxiv.org/abs/2607.15550", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-15550/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a15963a7fb4dfd488f0e79732d9ea0c9a03d43c1e3d5e50dd1ee8e00f1072be", "submitted_at": "2026-07-17T01:45:54Z", "updated_at": "2026-09-02T01:49:21Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.15550v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.15550"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.15550"}], "name": "SeerGuard: A Safety Framework for Mobile GUI Agents via World Model Prediction", "publication": null, "published_at": "2026-07-17", "record_sha256": "a3afe3778fc358e6539318e9d1db002bf180b8c3a9f33614dbbadf944f0aa815", "relations": [], "resource_id": "arxiv:2607.15550", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.15550v3", "topics": ["safety-security", "world-models"], "updated_at": "2026-09-02"}, "text": "SeerGuard: A Safety Framework for Mobile GUI Agents via World Model Prediction\nResource type: paper\nReview status: curated\nScope: core\n\nUses instruction screening and a safety-augmented world model to assess likely consequences of mobile GUI actions before execution.\nPlatforms: unspecified\nTopics: safety-security, world-models\nPublished: 2026-07-17\nAuthors: Xue Yu, Bo Yuan, Kailin Zhao, Pengshuai Yang, Hong Hu, Junlan Feng\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.15550v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-15550/"}
{"id": "arxiv:2607.17050", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6488798626abbec89ca739a0c5d188ae440484eb59b3be11f2b319ce742b22ec", "submitted_at": "2026-07-19T03:29:08Z", "updated_at": "2026-07-19T03:29:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17050v1"}, "authors": ["Yaohan Yang", "Minglei Shi", "Borui Zhang", "Jie Zhou", "Jiwen Lu"], "canonical_source_url": "https://arxiv.org/abs/2607.17050", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-17050/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6488798626abbec89ca739a0c5d188ae440484eb59b3be11f2b319ce742b22ec", "submitted_at": "2026-07-19T03:29:08Z", "updated_at": "2026-07-19T03:29:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17050v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "identifiers": {"arxiv": "2607.17050"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.17050"}], "name": "EvoGUI: An Evolution-Aware Benchmark for GUI State-Transition Understanding", "publication": null, "published_at": "2026-07-19", "record_sha256": "5494bf6567521986d4bd3672fa0136fa8528180688a8db7afe2846324aa52b7c", "relations": [], "resource_id": "arxiv:2607.17050", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.17050v1", "topics": ["evaluation", "world-models", "action-prediction"], "updated_at": "2026-07-19"}, "text": "EvoGUI: An Evolution-Aware Benchmark for GUI State-Transition Understanding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces EvoGUI, which derives temporal ordering, inverse action/value prediction, and successor-discrimination probes from recorded GUI trajectories. Its web-derived benchmark isolates transition understanding from the other abilities involved in end-to-end task success.\nPlatforms: web\nTopics: evaluation, world-models, action-prediction\nPublished: 2026-07-19\nAuthors: Yaohan Yang, Minglei Shi, Borui Zhang, Jie Zhou, Jiwen Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.17050v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-17050/"}
{"id": "arxiv:2607.17527", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2069acf693e0dced49d0e10e21bf9fe28b99f50f1b932837f16137e0d3170a80", "submitted_at": "2026-07-20T04:03:38Z", "updated_at": "2026-07-20T04:03:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17527v1"}, "authors": ["Ruei-Che Chang", "Wenqian Xu", "Dingzeyu Li", "Bryan Wang", "Anhong Guo"], "canonical_source_url": "https://arxiv.org/abs/2607.17527", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-17527/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2069acf693e0dced49d0e10e21bf9fe28b99f50f1b932837f16137e0d3170a80", "submitted_at": "2026-07-20T04:03:38Z", "updated_at": "2026-07-20T04:03:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.17527v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.17527"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.17527"}], "name": "Sidekick: Designing Communication for Effective Multitasking with Computer Use Agents", "publication": null, "published_at": "2026-07-20", "record_sha256": "d89c7c293b365a899aef2629210aed7a7a77dbf4729a9b505f265a191c5c99c9", "relations": [], "resource_id": "arxiv:2607.17527", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.17527v1", "topics": ["human-in-the-loop", "observability", "long-horizon"], "updated_at": "2026-07-20"}, "text": "Sidekick: Designing Communication for Effective Multitasking with Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Sidekick, a communication interface that uses ambient status cues, multimodal action summaries, and visual and verbal explanations of reasoning to help people monitor computer-use agents. A user study examines progress awareness, error traceability, and multitasking across background and foreground interaction.\nPlatforms: unspecified\nTopics: human-in-the-loop, observability, long-horizon\nPublished: 2026-07-20\nAuthors: Ruei-Che Chang, Wenqian Xu, Dingzeyu Li, Bryan Wang, Anhong Guo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.17527v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-17527/"}
{"id": "arxiv:2607.18046", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b85b453a688543477d31a940f9d1da2857b2cb7a8cb9eea91a39a3a2b8a1e1c7", "submitted_at": "2026-07-20T15:16:59Z", "updated_at": "2026-08-18T02:32:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.18046v2"}, "authors": ["Zhuohang Fan", "Beichen Zhang", "Yuanfa Li", "Changqiao Wu", "Wei Liu", "Jian Luan", "Weigang Zhang"], "canonical_source_url": "https://arxiv.org/abs/2607.18046", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-18046/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b85b453a688543477d31a940f9d1da2857b2cb7a8cb9eea91a39a3a2b8a1e1c7", "submitted_at": "2026-07-20T15:16:59Z", "updated_at": "2026-08-18T02:32:56Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.18046v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.18046"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.18046"}], "name": "SEE: Structure-aware Exploring & Exploiting for Long-horizon GUI Agent Trajectory Synthesis", "publication": null, "published_at": "2026-07-20", "record_sha256": "0fa63251eb58a1dd2dab45b25c726a97bf3a960417300440e13568a42e2f45f7", "relations": [], "resource_id": "arxiv:2607.18046", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.18046v2", "topics": ["data-generation", "long-horizon", "mobile-navigation"], "updated_at": "2026-08-18"}, "text": "SEE: Structure-aware Exploring & Exploiting for Long-horizon GUI Agent Trajectory Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SEE, which first explores mobile interfaces to build a screen-and-element transition graph, then composes task trajectories through graph planning and controlled sampling. It studies coverage, loop avoidance, and transfer after training agents on the generated sequences.\nPlatforms: unspecified\nTopics: data-generation, long-horizon, mobile-navigation\nPublished: 2026-07-20\nAuthors: Zhuohang Fan, Beichen Zhang, Yuanfa Li, Changqiao Wu, Wei Liu, Jian Luan, Weigang Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.18046v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-18046/"}
{"id": "arxiv:2607.18659", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e382b15a9d25e99aa187a72c7d2a33bbeca82b2b347ba580ca0a1d73a8e01d19", "submitted_at": "2026-07-21T03:02:57Z", "updated_at": "2026-07-21T03:02:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.18659v1"}, "authors": ["Behzad Ousat", "Nikita Turkmen", "Lalchandra Rampersaud", "Dillan Bailey", "Amin Kharraz"], "canonical_source_url": "https://arxiv.org/abs/2607.18659", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-18659/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e382b15a9d25e99aa187a72c7d2a33bbeca82b2b347ba580ca0a1d73a8e01d19", "submitted_at": "2026-07-21T03:02:57Z", "updated_at": "2026-07-21T03:02:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.18659v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2607.18659"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.18659"}], "name": "Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents", "publication": null, "published_at": "2026-07-21", "record_sha256": "9dd9a89ea13a3e7ad0d8a8b3d8442994286387e3325a0ad237006528b6e573e9", "relations": [], "resource_id": "arxiv:2607.18659", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.18659v1", "topics": ["safety-security", "web-navigation", "evaluation"], "updated_at": "2026-07-21"}, "text": "Broken Gates: Re-evaluating Web Bot Defenses in the Age of LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMeasures how autonomous LLM browser agents interact with commercial bot-management and anti-automation defenses on the web.\nPlatforms: web\nTopics: safety-security, web-navigation, evaluation\nPublished: 2026-07-21\nAuthors: Behzad Ousat, Nikita Turkmen, Lalchandra Rampersaud, Dillan Bailey, Amin Kharraz\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.18659v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-18659/"}
{"id": "arxiv:2607.21557", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "965fe5aabfb44b4c82290fc347a463eb58fb2be7a3761496efd38167d0f7d5b6", "submitted_at": "2026-07-23T17:38:30Z", "updated_at": "2026-08-07T17:40:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.21557v3"}, "authors": ["Xiao Yu", "Baolin Peng", "Ruize Xu", "Hao Zou", "Qianhui Wu", "Hao Cheng", "Wenlin Yao", "Nikhil Singh", "Zhou Yu", "Jianfeng Gao"], "canonical_source_url": "https://arxiv.org/abs/2607.21557", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-21557/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "965fe5aabfb44b4c82290fc347a463eb58fb2be7a3761496efd38167d0f7d5b6", "submitted_at": "2026-07-23T17:38:30Z", "updated_at": "2026-08-07T17:40:50Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2607.21557v3"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "identifiers": {"arxiv": "2607.21557"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.21557"}], "name": "OpenForgeRL: Train Harness-native Agents in Any Environment", "publication": null, "published_at": "2026-07-23", "record_sha256": "bd0f1f1cd5ac520c9d780da0669994d5cac70e5a1b41e0db48d37e07734afb71", "relations": [], "resource_id": "arxiv:2607.21557", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.21557v3", "topics": ["reinforcement-learning", "sandbox", "tool-integration", "robustness"], "updated_at": "2026-08-07"}, "text": "OpenForgeRL: Train Harness-native Agents in Any Environment\nResource type: paper\nReview status: curated\nScope: core\n\nOpenForgeRL connects stateful agent harnesses to reinforcement-learning infrastructure through a model-call recording proxy and isolated rollout containers. Its validation includes browser and computer-use agents, alongside tool-oriented agents, and examines how harness choice affects learned reliability.\nPlatforms: web\nTopics: reinforcement-learning, sandbox, tool-integration, robustness\nPublished: 2026-07-23\nAuthors: Xiao Yu, Baolin Peng, Ruize Xu, Hao Zou, Qianhui Wu, Hao Cheng, Wenlin Yao, Nikhil Singh, Zhou Yu, Jianfeng Gao\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.21557v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-21557/"}
{"id": "arxiv:2607.22689", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "078ff9a56b769923d15a98581685a3e2f47aac7d898ea731999cabaec3918fa7", "submitted_at": "2026-07-17T12:48:03Z", "updated_at": "2026-07-17T12:48:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22689v1"}, "authors": ["Zedong Yu", "Qianxing Li", "Zhi Gao", "Liuyu Xiang", "Chenrui Shi", "Yang Liu", "Huiming Wu", "Yujie Wei", "Yuhao Fei", "Yubo Fu", "Zhaofeng He"], "canonical_source_url": "https://arxiv.org/abs/2607.22689", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-22689/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "078ff9a56b769923d15a98581685a3e2f47aac7d898ea731999cabaec3918fa7", "submitted_at": "2026-07-17T12:48:03Z", "updated_at": "2026-07-17T12:48:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22689v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.22689"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.22689"}], "name": "Beyond Sequential Interaction: Benchmarking Parallel Execution and Coordination for GUI Agents", "publication": null, "published_at": "2026-07-17", "record_sha256": "670e9fab6c6a0584a72555644ca58d8b4a0c987573556ea7e9d240e122b8073c", "relations": [], "resource_id": "arxiv:2607.22689", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.22689v1", "topics": ["evaluation", "planning", "efficiency", "long-horizon"], "updated_at": "2026-07-17"}, "text": "Beyond Sequential Interaction: Benchmarking Parallel Execution and Coordination for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ParaGUIBench, with tasks and efficiency measurements for coordinated GUI agents operating separate desktop instances over a shared file system. Its ParaGUI planner-worker system decomposes tasks and dispatches concurrent subtasks for comparison with serial execution.\nPlatforms: desktop\nTopics: evaluation, planning, efficiency, long-horizon\nPublished: 2026-07-17\nAuthors: Zedong Yu, Qianxing Li, Zhi Gao, Liuyu Xiang, Chenrui Shi, Yang Liu, Huiming Wu, Yujie Wei, Yuhao Fei, Yubo Fu, Zhaofeng He\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.22689v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-22689/"}
{"id": "arxiv:2607.22798", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a010fbef09d8359550cf156bf0dcba873f25a0d6fb8f118d53bf6e23715cf203", "submitted_at": "2026-07-24T14:17:03Z", "updated_at": "2026-07-24T14:17:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22798v1"}, "authors": ["Yan Yang", "Xiangru Jian", "Ziyang Luo", "Zirui Zhao", "Yutong Dai", "Ziji Shi", "Hanshu Yan", "Jun Hao Liew", "Silvio Savarese", "Junnan Li"], "canonical_source_url": "https://arxiv.org/abs/2607.22798", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-22798/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a010fbef09d8359550cf156bf0dcba873f25a0d6fb8f118d53bf6e23715cf203", "submitted_at": "2026-07-24T14:17:03Z", "updated_at": "2026-07-24T14:17:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.22798v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.22798"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.22798"}], "name": "StateAct: Program State, before Pixels, for Long-Horizon Computer-Use Agents", "publication": null, "published_at": "2026-07-24", "record_sha256": "8ac0981be3aee93b70e915639cce697196ada4b9a864dc0eb1c0c132e519fbf2", "relations": [], "resource_id": "arxiv:2607.22798", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.22798v1", "topics": ["tool-integration", "state-tracking", "long-horizon"], "updated_at": "2026-07-24"}, "text": "StateAct: Program State, before Pixels, for Long-Horizon Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces StateAct, a code-first computer-use harness that manipulates program state directly and delegates screenshot-and-click subgoals to a GUI agent. An independent completion gate checks saved outputs, while fresh subagents handle subgoals to limit accumulated context.\nPlatforms: desktop\nTopics: tool-integration, state-tracking, long-horizon\nPublished: 2026-07-24\nAuthors: Yan Yang, Xiangru Jian, Ziyang Luo, Zirui Zhao, Yutong Dai, Ziji Shi, Hanshu Yan, Jun Hao Liew, Silvio Savarese, Junnan Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.22798v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-22798/"}
{"id": "arxiv:2607.23263", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2de35aa3c7f4457d43d485eb136d8b380ea61870cc7253e807880ae731c51316", "submitted_at": "2026-07-25T16:00:45Z", "updated_at": "2026-09-26T12:15:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.23263v2"}, "authors": ["Yang Wan", "Zhenhao Zhang", "Jierui Wang", "Xiangxin Zhou", "Linchao Zhu"], "canonical_source_url": "https://arxiv.org/abs/2607.23263", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-23263/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2de35aa3c7f4457d43d485eb136d8b380ea61870cc7253e807880ae731c51316", "submitted_at": "2026-07-25T16:00:45Z", "updated_at": "2026-09-26T12:15:44Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.23263v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.23263"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.23263"}], "name": "SeekJudge: A Practical Reward Framework for Reinforcement Learning in Computer-Use Agents", "publication": null, "published_at": "2026-07-25", "record_sha256": "170031606adbe730acdec2d746439ba539aae033008e78345385a320ff129d6b", "relations": [], "resource_id": "arxiv:2607.23263", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.23263v2", "topics": ["reward-models", "evaluation", "data-generation"], "updated_at": "2026-09-26"}, "text": "SeekJudge: A Practical Reward Framework for Reinforcement Learning in Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SeekJudge, which iteratively searches and analyzes GUI trajectories to assess instruction fulfillment. The work combines dense judgment-data synthesis, a step-label benchmark, and reward-server scheduling to study accurate and practical reward feedback for agent training.\nPlatforms: unspecified\nTopics: reward-models, evaluation, data-generation\nPublished: 2026-07-25\nAuthors: Yang Wan, Zhenhao Zhang, Jierui Wang, Xiangxin Zhou, Linchao Zhu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.23263v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-23263/"}
{"id": "arxiv:2607.24112", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4651964cbe45e4d740ed8f789753f82ae808e442d82c6598e7b3c997c09453c3", "submitted_at": "2026-07-27T07:54:08Z", "updated_at": "2026-07-27T07:54:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24112v1"}, "authors": ["Xiangyan Liu", "Kaixin Li", "Haonan Wang", "Biao Wu", "Meng Fang", "Longxu Dou", "Chao Du", "Michael Qizhe Shieh", "Tianyu Pang"], "canonical_source_url": "https://arxiv.org/abs/2607.24112", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-24112/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4651964cbe45e4d740ed8f789753f82ae808e442d82c6598e7b3c997c09453c3", "submitted_at": "2026-07-27T07:54:08Z", "updated_at": "2026-07-27T07:54:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24112v1"}}, "environments": ["desktop", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.24112"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.24112"}], "name": "Scaling GUI Agents with Visual State Transitions", "publication": null, "published_at": "2026-07-27", "record_sha256": "73c96fe08d05224fca45dd105e42467ad55a08313fe1ac5ee4a902ddeed6d470", "relations": [], "resource_id": "arxiv:2607.24112", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.24112v1", "topics": ["world-models", "action-prediction", "data-generation"], "updated_at": "2026-07-27"}, "text": "Scaling GUI Agents with Visual State Transitions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces State Transition Pretraining, which jointly learns to infer actions from GUI changes and predict future states from current screens and actions. Subsequent instruction-trajectory fine-tuning tests whether these representations transfer to desktop and mobile agent tasks.\nPlatforms: desktop, android\nTopics: world-models, action-prediction, data-generation\nPublished: 2026-07-27\nAuthors: Xiangyan Liu, Kaixin Li, Haonan Wang, Biao Wu, Meng Fang, Longxu Dou, Chao Du, Michael Qizhe Shieh, Tianyu Pang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.24112v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-24112/"}
{"id": "arxiv:2607.24167", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57fa0a3eaad7525991656989ed3011101026362acec43895f32cf6e24be2f4ee", "submitted_at": "2026-07-27T08:48:01Z", "updated_at": "2026-07-27T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24167v1"}, "authors": ["Guangyi Liu", "Huan Zhao", "Quanming Yao"], "canonical_source_url": "https://arxiv.org/abs/2607.24167", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-24167/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "57fa0a3eaad7525991656989ed3011101026362acec43895f32cf6e24be2f4ee", "submitted_at": "2026-07-27T08:48:01Z", "updated_at": "2026-07-27T08:48:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.24167v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.24167"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.24167"}], "name": "Falsifiable Commitment Planning for Self-Correcting Web Agents", "publication": null, "published_at": "2026-07-27", "record_sha256": "d7f97c285414433f985d78b85a63c23e65d5f6cdd188f53d60c4441806b47d8e", "relations": [], "resource_id": "arxiv:2607.24167", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.24167v1", "topics": ["planning", "robustness", "self-reflection"], "updated_at": "2026-07-27"}, "text": "Falsifiable Commitment Planning for Self-Correcting Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces FCPAgent, which attaches confirming and falsifying evidence to each skill-grounded plan step. A plan–test–repair loop checks actions and observations, then localizes contradictions so execution, skills, or plans can be revised at the appropriate level.\nPlatforms: web\nTopics: planning, robustness, self-reflection\nPublished: 2026-07-27\nAuthors: Guangyi Liu, Huan Zhao, Quanming Yao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.24167v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-24167/"}
{"id": "arxiv:2607.25904", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1688a528d35a8213a8027385b1b1327536567f6ec307aa6562eea543ee71d7f", "submitted_at": "2026-07-28T16:01:38Z", "updated_at": "2026-07-29T05:13:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.25904v2"}, "authors": ["Chenrui Shi", "Yuwei Wu", "Yang Liu", "Ruining Feng", "Zirui Shang", "Zhi Gao", "Lifeng Fan", "Che Sun"], "canonical_source_url": "https://arxiv.org/abs/2607.25904", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-25904/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d1688a528d35a8213a8027385b1b1327536567f6ec307aa6562eea543ee71d7f", "submitted_at": "2026-07-28T16:01:38Z", "updated_at": "2026-07-29T05:13:57Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.25904v2"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}], "identifiers": {"arxiv": "2607.25904"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.25904"}], "name": "Interactive Reward Agent: GUI Task Evaluation via Environment-State Verification", "publication": null, "published_at": "2026-07-28", "record_sha256": "8e60208180a3afe2e93f3021a3ace297c9b5be58970d5bc38ac97a9d6824b9bc", "relations": [], "resource_id": "arxiv:2607.25904", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.25904v2", "topics": ["reward-models", "evaluation", "tool-integration"], "updated_at": "2026-07-29"}, "text": "Interactive Reward Agent: GUI Task Evaluation via Environment-State Verification\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces an Interactive Reward Agent that proposes completion conditions and verifies them using system, application, and GUI tools after execution. GUI-RewardBench provides annotated desktop trajectories for evaluating these evidence-based judgments and their use in reinforcement learning.\nPlatforms: desktop, linux\nTopics: reward-models, evaluation, tool-integration\nPublished: 2026-07-28\nAuthors: Chenrui Shi, Yuwei Wu, Yang Liu, Ruining Feng, Zirui Shang, Zhi Gao, Lifeng Fan, Che Sun\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.25904v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-25904/"}
{"id": "arxiv:2607.26041", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352c892a52692301e14125e126a52508feb39ef7cd034d7cb0c48a27fb54b8ac", "submitted_at": "2026-07-28T17:49:51Z", "updated_at": "2026-07-29T19:07:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.26041v2"}, "authors": ["Abhishek Pillai", "Samir Kumar Nayak", "Yuan Chen"], "canonical_source_url": "https://arxiv.org/abs/2607.26041", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-26041/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352c892a52692301e14125e126a52508feb39ef7cd034d7cb0c48a27fb54b8ac", "submitted_at": "2026-07-28T17:49:51Z", "updated_at": "2026-07-29T19:07:50Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.26041v2"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.26041"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.26041"}], "name": "Desktop-Delta Bench: Do Computer-Use Models Understand Desktop GUI Transitions?", "publication": null, "published_at": "2026-07-28", "record_sha256": "f9abd0545716f483a6b8771c566c0c0691a7d37f067fac403380562ef49097c5", "relations": [], "resource_id": "arxiv:2607.26041", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.26041v2", "topics": ["evaluation", "screen-understanding", "state-tracking"], "updated_at": "2026-07-29"}, "text": "Desktop-Delta Bench: Do Computer-Use Models Understand Desktop GUI Transitions?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Desktop-Delta Bench, an offline diagnostic of GUI transition understanding using temporal-ordering tasks and before/after action inference. It includes decoy observations to test whether models track the source and task relevance of screen changes.\nPlatforms: desktop, linux\nTopics: evaluation, screen-understanding, state-tracking\nPublished: 2026-07-28\nAuthors: Abhishek Pillai, Samir Kumar Nayak, Yuan Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.26041v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-26041/"}
{"id": "arxiv:2607.28074", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f28e1ee8e101000f3cfe9215555251656fcb40b921b5c41c895fbc958f867e3", "submitted_at": "2026-07-30T11:48:22Z", "updated_at": "2026-07-30T11:48:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28074v1"}, "authors": ["Yash Pandya", "Sahil Gupta", "Sarthak Harne", "Archana Yadav", "Kavyansh Chourasia", "Hussein Mozannar", "Vibhav Vineet", "Sara Abdali", "Corby Rosset", "Yash Lara", "Ahmed Awadallah", "Ece Kamar", "Akshay Nambi"], "canonical_source_url": "https://arxiv.org/abs/2607.28074", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28074/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f28e1ee8e101000f3cfe9215555251656fcb40b921b5c41c895fbc958f867e3", "submitted_at": "2026-07-30T11:48:22Z", "updated_at": "2026-07-30T11:48:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28074v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28074"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28074"}], "name": "Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale", "publication": null, "published_at": "2026-07-30", "record_sha256": "9e3c6b8d3a7f126af062eeb15c4a6b262aaa320a7c85069d272af78fddbba38a", "relations": [], "resource_id": "arxiv:2607.28074", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28074v1", "topics": ["sandbox", "data-generation", "reinforcement-learning"], "updated_at": "2026-07-30"}, "text": "Echoverse: Deep, Evolving Environments for Training Computer-Use Agents at Scale\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Echoverse, which compiles specifications into stateful applications with database-grounded graders and uses rollouts to improve both the environments and agent policies. It examines behavioral depth, targeted interface practice, and environment repair as sources of training value.\nPlatforms: web\nTopics: sandbox, data-generation, reinforcement-learning\nPublished: 2026-07-30\nAuthors: Yash Pandya, Sahil Gupta, Sarthak Harne, Archana Yadav, Kavyansh Chourasia, Hussein Mozannar, Vibhav Vineet, Sara Abdali, Corby Rosset, Yash Lara, Ahmed Awadallah, Ece Kamar, Akshay Nambi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28074v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28074/"}
{"id": "arxiv:2607.28227", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52fb1b49501132f1dbe3564daa02cca22a93fc8d7c28dec8ee6070a23889574a", "submitted_at": "2026-07-30T13:58:41Z", "updated_at": "2026-07-30T13:58:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28227v1"}, "authors": ["Hanzhang Zhou", "Panrong Tong", "Xu Zhang", "Quyu Kong", "Chenglin Cai", "Tianyu Xia", "Gongjie Zhang", "Jianan Zhang", "Long Li", "Long Chen", "Lei Wang", "Gaole Dai", "Pengxiang Li", "Liangyu Chen", "Yue Wang", "Steven Hoi"], "canonical_source_url": "https://arxiv.org/abs/2607.28227", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28227/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52fb1b49501132f1dbe3564daa02cca22a93fc8d7c28dec8ee6070a23889574a", "submitted_at": "2026-07-30T13:58:41Z", "updated_at": "2026-07-30T13:58:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28227v1"}}, "environments": ["cross-platform", "web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28227"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28227"}], "name": "Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents", "publication": null, "published_at": "2026-07-30", "record_sha256": "05a83adcef1b249d5889e474ee8dc20be2fd04d726e54fdabe2cd56c28721676", "relations": [], "resource_id": "arxiv:2607.28227", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28227v1", "topics": ["reinforcement-learning", "long-horizon", "tool-integration"], "updated_at": "2026-07-30"}, "text": "Qwen-UI-Agent Technical Report: Toward Next-Generation Real-World Centric Foundation GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a foundation GUI agent with combined GUI and command-line actions, large-scale interaction infrastructure, and iterative environment and task generation.\nPlatforms: cross-platform, web, desktop\nTopics: reinforcement-learning, long-horizon, tool-integration\nPublished: 2026-07-30\nAuthors: Hanzhang Zhou, Panrong Tong, Xu Zhang, Quyu Kong, Chenglin Cai, Tianyu Xia, Gongjie Zhang, Jianan Zhang, Long Li, Long Chen, Lei Wang, Gaole Dai, Pengxiang Li, Liangyu Chen, Yue Wang, Steven Hoi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28227v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28227/"}
{"id": "arxiv:2607.28367", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75f6c942d6854a259556b7ef1060fa0f142755162da329580c6177cd1adef15", "submitted_at": "2026-07-30T15:29:42Z", "updated_at": "2026-07-30T15:29:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28367v1"}, "authors": ["Zihan Dong", "Zhiyuan Ma", "Zekun Wang", "Yunqing Li", "Zirou Liu", "Ruixuan Deng", "Qishi Zhan", "Rui Qian"], "canonical_source_url": "https://arxiv.org/abs/2607.28367", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28367/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d75f6c942d6854a259556b7ef1060fa0f142755162da329580c6177cd1adef15", "submitted_at": "2026-07-30T15:29:42Z", "updated_at": "2026-07-30T15:29:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28367v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28367"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28367"}], "name": "How Benchmarks Mis-Score Computer-Use Agents", "publication": null, "published_at": "2026-07-30", "record_sha256": "ff4b587064a45ab8a82fab2c6cbe19451ff51f17827a7b0cab5774e57c808485", "relations": [], "resource_id": "arxiv:2607.28367", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28367v1", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-07-30"}, "text": "How Benchmarks Mis-Score Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAudits failure-scored computer-use trajectories to distinguish agent errors from broken tasks and evaluator false negatives. It organizes reliability issues across task construction, trajectory evidence, scoring, and reporting, then derives corresponding evaluation design practices.\nPlatforms: web, desktop\nTopics: evaluation, observability, robustness\nPublished: 2026-07-30\nAuthors: Zihan Dong, Zhiyuan Ma, Zekun Wang, Yunqing Li, Zirou Liu, Ruixuan Deng, Qishi Zhan, Rui Qian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28367v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28367/"}
{"id": "arxiv:2607.28399", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "463cf35eb57b3d38dda6ce768e6f08d93b1c16d6abba0d790760d846cd56fa2c", "submitted_at": "2026-07-30T15:50:10Z", "updated_at": "2026-07-30T15:50:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28399v1"}, "authors": ["Zihan Dong", "Rui Qian", "Qishi Zhan", "Dongshen Peng", "Kaixin Li", "Yu Li"], "canonical_source_url": "https://arxiv.org/abs/2607.28399", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28399/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "463cf35eb57b3d38dda6ce768e6f08d93b1c16d6abba0d790760d846cd56fa2c", "submitted_at": "2026-07-30T15:50:10Z", "updated_at": "2026-07-30T15:50:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28399v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28399"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28399"}], "name": "Why Are GUI Agents Correct but Late? Decode on the Decision-Time Critical Path, Tested with Pre-Compiled Policy Trees", "publication": null, "published_at": "2026-07-30", "record_sha256": "54385a5992a3c8de591a81c206e3192e555e5c22fefd021828a860e0dc9e5df7", "relations": [], "resource_id": "arxiv:2607.28399", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28399v1", "topics": ["efficiency", "planning", "robustness"], "updated_at": "2026-07-30"}, "text": "Why Are GUI Agents Correct but Late? Decode on the Decision-Time Critical Path, Tested with Pre-Compiled Policy Trees\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Adaptive Anticipatory Policy Trees, which precompute guarded actions and deadlines during idle periods so transient GUI events can be handled without new text generation. Paired experiments examine when this prepared execution helps and when reactive control remains preferable.\nPlatforms: unspecified\nTopics: efficiency, planning, robustness\nPublished: 2026-07-30\nAuthors: Zihan Dong, Rui Qian, Qishi Zhan, Dongshen Peng, Kaixin Li, Yu Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28399v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28399/"}
{"id": "arxiv:2607.28573", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bacd1c5df5bf3163c83fae546ae4ced36996f33115568dfbc9e44b17fe507a31", "submitted_at": "2026-07-30T17:36:36Z", "updated_at": "2026-07-30T17:36:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28573v1"}, "authors": ["Woongkyu Lee", "Jungwook Choi"], "canonical_source_url": "https://arxiv.org/abs/2607.28573", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28573/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bacd1c5df5bf3163c83fae546ae4ced36996f33115568dfbc9e44b17fe507a31", "submitted_at": "2026-07-30T17:36:36Z", "updated_at": "2026-07-30T17:36:36Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.28573v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28573"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28573"}], "name": "Rethinking Inference-Time Scaling in Local Computer-Use Agents: Failure Modes and Compute Tradeoffs", "publication": null, "published_at": "2026-07-30", "record_sha256": "78acf1369aa96b784584acf2f6f9ad5687538584cbe689c1153b5fac55886a80", "relations": [], "resource_id": "arxiv:2607.28573", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28573v1", "topics": ["efficiency", "evaluation", "robustness"], "updated_at": "2026-07-30"}, "text": "Rethinking Inference-Time Scaling in Local Computer-Use Agents: Failure Modes and Compute Tradeoffs\nResource type: paper\nReview status: curated\nScope: core\n\nExamines inference-time scaling for local computer-use models by varying context, step budgets, decomposition, and parallel execution. OSWorld experiments track both task performance and changing failure modes, exposing cases where extra computation extends errors or adds coordination overhead.\nPlatforms: desktop\nTopics: efficiency, evaluation, robustness\nPublished: 2026-07-30\nAuthors: Woongkyu Lee, Jungwook Choi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28573v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28573/"}
{"id": "arxiv:2607.28609", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3e82df7abffbd0c8f7e969ed8edb77423263a2b1d826a3e72529b0a51903d58", "submitted_at": "2026-07-30T17:57:41Z", "updated_at": "2026-08-06T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.28609v2"}, "authors": ["Qiushi Sun", "Kanzhi Cheng", "Yian Wang", "Bowen Yang", "Hang Yan", "Liheng Chen", "Fangzhi Xu", "Zichen Ding", "Nuo Chen", "Jialin Cao", "Xingdong Gong", "Zehao Li", "Kaiming Jin", "Xinfeng Yuan", "Zhoumianze Liu", "Jingyang Gong", "Zhangyue Yin", "Jiahui Gao", "Zhiyong Wu", "Tianbao Xie", "Jianbing Zhang", "Ben Kao", "Lingpeng Kong"], "canonical_source_url": "https://arxiv.org/abs/2607.28609", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-28609/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3e82df7abffbd0c8f7e969ed8edb77423263a2b1d826a3e72529b0a51903d58", "submitted_at": "2026-07-30T17:57:41Z", "updated_at": "2026-08-06T17:55:59Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2607.28609v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.28609"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.28609"}], "name": "OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models", "publication": null, "published_at": "2026-07-30", "record_sha256": "f90bba85a6d13f94adc9b7c5339d6fdb6522712af938598f9510572b0bd19683", "relations": [], "resource_id": "arxiv:2607.28609", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.28609v2", "topics": ["reward-models", "evaluation", "data-generation"], "updated_at": "2026-08-06"}, "text": "OSReward: Instituting Standardized Evaluation for Cross-Platform Computer-Use Reward Models\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OSReward and its harder and multidimensional variants for assessing judgments of computer-use trajectories. It also releases a reasoning-annotated judgment corpus and trains reward models, examining reliability, leniency bias, and evaluation cost.\nPlatforms: cross-platform\nTopics: reward-models, evaluation, data-generation\nPublished: 2026-07-30\nAuthors: Qiushi Sun, Kanzhi Cheng, Yian Wang, Bowen Yang, Hang Yan, Liheng Chen, Fangzhi Xu, Zichen Ding, Nuo Chen, Jialin Cao, Xingdong Gong, Zehao Li, Kaiming Jin, Xinfeng Yuan, Zhoumianze Liu, Jingyang Gong, Zhangyue Yin, Jiahui Gao, Zhiyong Wu, Tianbao Xie, Jianbing Zhang, Ben Kao, Lingpeng Kong\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.28609v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-28609/"}
{"id": "arxiv:2607.29320", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfffaffc68695489e2a705515378112b232e66b6f9e8e3a9bd5df04be95bd2c9", "submitted_at": "2026-07-31T11:51:04Z", "updated_at": "2026-07-31T11:51:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.29320v1"}, "authors": ["Hang Yan", "Zhangxuan GU", "Beitong Zhou", "Jiaxuan Chen", "Runze Li", "Yusong Hu", "Shuheng Shen", "Changhua Meng"], "canonical_source_url": "https://arxiv.org/abs/2607.29320", "citation_url": "https://awesomecua.github.io/resources/arxiv-2607-29320/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dfffaffc68695489e2a705515378112b232e66b6f9e8e3a9bd5df04be95bd2c9", "submitted_at": "2026-07-31T11:51:04Z", "updated_at": "2026-07-31T11:51:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2607.29320v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2607.29320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2607.29320"}], "name": "MAGA: Multi-Platform Self-Fusion of GUI Agents via Structured Action Distillation", "publication": null, "published_at": "2026-07-31", "record_sha256": "3f5b7dacc2b799a8bc0d8c9ebee0d6b3c079b9488f25f0d5238e10bca0e302d5", "relations": [], "resource_id": "arxiv:2607.29320", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2607.29320v1", "topics": ["imitation-learning", "grounding", "automation"], "updated_at": "2026-07-31"}, "text": "MAGA: Multi-Platform Self-Fusion of GUI Agents via Structured Action Distillation\nResource type: paper\nReview status: curated\nScope: core\n\nMAGA distills specialized GUI agents into one policy by allocating supervision according to structured-action correctness rather than treating all response tokens equally. A training-only teacher hint supports fusion across mobile, web and desktop settings without changing student inputs.\nPlatforms: cross-platform\nTopics: imitation-learning, grounding, automation\nPublished: 2026-07-31\nAuthors: Hang Yan, Zhangxuan GU, Beitong Zhou, Jiaxuan Chen, Runze Li, Yusong Hu, Shuheng Shen, Changhua Meng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2607.29320v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2607-29320/"}
{"id": "arxiv:2608.00202", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c93c90e6d4026452810a866a0c49d57afdec15ad5ced30606c925a8e83c8f7", "submitted_at": "2026-07-31T18:27:12Z", "updated_at": "2026-07-31T18:27:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.00202v1"}, "authors": ["Yashaswi Malla", "Sandra Siby"], "canonical_source_url": "https://arxiv.org/abs/2608.00202", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-00202/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "60c93c90e6d4026452810a866a0c49d57afdec15ad5ced30606c925a8e83c8f7", "submitted_at": "2026-07-31T18:27:12Z", "updated_at": "2026-07-31T18:27:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.00202v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2608.00202"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.00202"}], "name": "From Monoliths to Swarms: A Study of Attack Surface Evolution in the Transition to Multi-Agent Web Systems", "publication": null, "published_at": "2026-07-31", "record_sha256": "2a5b7b62a101f9b5dcade4f61684fad82503d87e784a0c4c70030472487ee9f4", "relations": [], "resource_id": "arxiv:2608.00202", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.00202v1", "topics": ["safety-security", "web-navigation", "planning"], "updated_at": "2026-07-31"}, "text": "From Monoliths to Swarms: A Study of Attack Surface Evolution in the Transition to Multi-Agent Web Systems\nResource type: paper\nReview status: curated\nScope: core\n\nCategorizes structural attack surfaces that emerge when single-agent browser systems are decomposed into multi-agent web workflows.\nPlatforms: web\nTopics: safety-security, web-navigation, planning\nPublished: 2026-07-31\nAuthors: Yashaswi Malla, Sandra Siby\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.00202v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-00202/"}
{"id": "arxiv:2608.02352", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26a61ccfdddb5cdc2da2b14d29ae8d4098544fea4e9458126f2a49c28418c4f0", "submitted_at": "2026-08-03T15:04:20Z", "updated_at": "2026-08-03T15:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.02352v1"}, "authors": ["Dunjie Lu", "Shuai Bai", "Tianyi Bai", "Sicheng Fan", "Chang Gao", "Jian Guan", "Feng Hu", "Mianqiu Huang", "Xingyang Huang", "Yizhen Jiang", "Yuheng Jing", "Dehui Kong", "Ning Li", "Dayiheng Liu", "Shixuan Liu", "Zheng Liu", "Que Shen", "Bowen Wang", "Junli Wang", "Chencan Wu", "Rui Xie", "Tianbao Xie", "Zhihui Xie", "Haiyang Xu", "An Yang", "Tao Yu", "Wenzhen Yuan", "Xi Zhang", "Zhenru Zhang", "Mingkang Zhu", "Zhaoqing Zhu", "Yizhong Cao", "Kai Dang", "Binyuan Hui", "Kaixin Li", "Junyang Lin", "Haiquan Wang", "Zekun Wang", "Yiheng Xu", "Fan Yan", "Mengqi Yuan", "Danyang Zhang", "Jiajun Zhang", "Zhipeng Zhang", "Fan Zhou", "Fan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2608.02352", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-02352/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "26a61ccfdddb5cdc2da2b14d29ae8d4098544fea4e9458126f2a49c28418c4f0", "submitted_at": "2026-08-03T15:04:20Z", "updated_at": "2026-08-03T15:04:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.02352v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.02352"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.02352"}], "name": "Qwen-CUA: Native Computer Use for (almost) Everything", "publication": null, "published_at": "2026-08-03", "record_sha256": "36c4502e7f0254be8aca9a1b6b8708f9c460ff58fc0b4f94763d4867650c53f3", "relations": [], "resource_id": "arxiv:2608.02352", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.02352v1", "topics": ["reinforcement-learning", "imitation-learning", "data-generation", "interaction-history", "long-horizon", "evaluation"], "updated_at": "2026-08-03"}, "text": "Qwen-CUA: Native Computer Use for (almost) Everything\nResource type: paper\nReview status: curated\nScope: core\n\nQwen-CUA operates through screenshots and keyboard/mouse events while compressing older visual history into reusable context blocks. Its report combines verifiable task generation, trajectory-based training and separate browser and Bash-augmented studies.\nPlatforms: desktop, web\nTopics: reinforcement-learning, imitation-learning, data-generation, interaction-history, long-horizon, evaluation\nPublished: 2026-08-03\nAuthors: Dunjie Lu, Shuai Bai, Tianyi Bai, Sicheng Fan, Chang Gao, Jian Guan, Feng Hu, Mianqiu Huang, Xingyang Huang, Yizhen Jiang, Yuheng Jing, Dehui Kong, Ning Li, Dayiheng Liu, Shixuan Liu, Zheng Liu, Que Shen, Bowen Wang, Junli Wang, Chencan Wu, Rui Xie, Tianbao Xie, Zhihui Xie, Haiyang Xu, An Yang, Tao Yu, Wenzhen Yuan, Xi Zhang, Zhenru Zhang, Mingkang Zhu, Zhaoqing Zhu, Yizhong Cao, Kai Dang, Binyuan Hui, Kaixin Li, Junyang Lin, Haiquan Wang, Zekun Wang, Yiheng Xu, Fan Yan, Mengqi Yuan, Danyang Zhang, Jiajun Zhang, Zhipeng Zhang, Fan Zhou, Fan Zhou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.02352v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-02352/"}
{"id": "arxiv:2608.02643", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a11e065bef4cf860c36ca2f682801ca21b552fbd40bd1b6f5d9ae39b56e7d77a", "submitted_at": "2026-07-31T13:59:45Z", "updated_at": "2026-09-06T04:43:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.02643v2"}, "authors": ["Weijia Zhang", "Kunlun Zhu", "Zeyi Liu", "Yinting Chen", "Tianyi Ma", "Jiateng Liu", "Jiaxun Zhang", "Bingxuan Li", "Xiangru Tang", "Heng Ji", "Jiaxuan You"], "canonical_source_url": "https://arxiv.org/abs/2608.02643", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-02643/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a11e065bef4cf860c36ca2f682801ca21b552fbd40bd1b6f5d9ae39b56e7d77a", "submitted_at": "2026-07-31T13:59:45Z", "updated_at": "2026-09-06T04:43:23Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.02643v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.02643"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.02643"}], "name": "CUADebug: Diagnosing and Repairing Computer-Use Agent Failures", "publication": null, "published_at": "2026-07-31", "record_sha256": "0edc0e5fcb07193ebf32f6f9a19438c4367b058a59231c0b69249e359edd01dc", "relations": [], "resource_id": "arxiv:2608.02643", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.02643v2", "topics": ["evaluation", "robustness", "self-reflection", "state-tracking"], "updated_at": "2026-09-06"}, "text": "CUADebug: Diagnosing and Repairing Computer-Use Agent Failures\nResource type: paper\nReview status: curated\nScope: core\n\nCUADebug combines a failure taxonomy and human-annotated OSWorld failure trajectories with an agent that locates causal steps and proposes corrections from screenshots and action traces. It evaluates root-cause diagnosis separately from controlled task re-execution and failure recovery.\nPlatforms: unspecified\nTopics: evaluation, robustness, self-reflection, state-tracking\nPublished: 2026-07-31\nAuthors: Weijia Zhang, Kunlun Zhu, Zeyi Liu, Yinting Chen, Tianyi Ma, Jiateng Liu, Jiaxun Zhang, Bingxuan Li, Xiangru Tang, Heng Ji, Jiaxuan You\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.02643v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-02643/"}
{"id": "arxiv:2608.03270", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb64bb5faa37e68b35a2398ad896fc02181a4d2749a3067391e6b3bf023a0f48", "submitted_at": "2026-08-04T07:47:37Z", "updated_at": "2026-08-04T07:47:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03270v1"}, "authors": ["Zichuan Fu", "Shirong Wang", "Wenlin Zhang", "Guojing Li", "Yimin Deng", "Jingtong Gao", "Junjia Qi", "Hanyu Yan", "Yefeng Zheng", "Xiaopeng Li", "Wanyu Wang", "Xian Wu", "Xiangyu Zhao"], "canonical_source_url": "https://arxiv.org/abs/2608.03270", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-03270/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb64bb5faa37e68b35a2398ad896fc02181a4d2749a3067391e6b3bf023a0f48", "submitted_at": "2026-08-04T07:47:37Z", "updated_at": "2026-08-04T07:47:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.03270v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.03270"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03270"}], "name": "GUI-Lens: Coarse-to-Fine Cropping for GUI Grounding with General-Purpose VLMs", "publication": null, "published_at": "2026-08-04", "record_sha256": "0ad5eb99e1ac879cc2984fa1e8ddac3d9b4708ed8c1823898604136f6179acb1", "relations": [], "resource_id": "arxiv:2608.03270", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.03270v1", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-04"}, "text": "GUI-Lens: Coarse-to-Fine Cropping for GUI Grounding with General-Purpose VLMs\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Lens uses OCR and detected interface components as coordinate references while a general-purpose VLM selects progressively enlarged screenshot crops. Instruction-conditioned checks constrain the crop and click proposals before mapping the final location back to the original screen.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, evaluation\nPublished: 2026-08-04\nAuthors: Zichuan Fu, Shirong Wang, Wenlin Zhang, Guojing Li, Yimin Deng, Jingtong Gao, Junjia Qi, Hanyu Yan, Yefeng Zheng, Xiaopeng Li, Wanyu Wang, Xian Wu, Xiangyu Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.03270v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-03270/"}
{"id": "arxiv:2608.03327", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c8692712194d0ff822091d13324e6bb540339103a824f1030f7a971699d0a0b", "submitted_at": "2026-08-04T08:35:51Z", "updated_at": "2026-08-06T07:26:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03327v2"}, "authors": ["Siqi Fan", "Minghao Li", "Xiaoqian Ma", "Wenhui Tan", "Xiusheng Huang", "Juntong Wu", "Liujie Zhang", "Shuo Shang", "Weihang Chen"], "canonical_source_url": "https://arxiv.org/abs/2608.03327", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-03327/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2c8692712194d0ff822091d13324e6bb540339103a824f1030f7a971699d0a0b", "submitted_at": "2026-08-04T08:35:51Z", "updated_at": "2026-08-06T07:26:24Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03327v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.03327"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03327"}], "name": "Screenshots or Tools? Eliciting Tool Use and Managing Multimodal Context in Hybrid GUI-MCP Computer-Use Agents", "publication": null, "published_at": "2026-08-04", "record_sha256": "923a8ae9c985a529315955cfbdb1cb62de32e23cb8fe98e841dda10c6d1ed1d5", "relations": [], "resource_id": "arxiv:2608.03327", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.03327v2", "topics": ["tool-integration", "reinforcement-learning", "efficiency", "interaction-history", "evaluation"], "updated_at": "2026-08-06"}, "text": "Screenshots or Tools? Eliciting Tool Use and Managing Multimodal Context in Hybrid GUI-MCP Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis study separates tool-adoption behavior from task competence in hybrid GUI–MCP agents. It tests tool-use incentives and screenshot-history compression, showing why changed action preferences alone need not improve held-out accuracy.\nPlatforms: desktop\nTopics: tool-integration, reinforcement-learning, efficiency, interaction-history, evaluation\nPublished: 2026-08-04\nAuthors: Siqi Fan, Minghao Li, Xiaoqian Ma, Wenhui Tan, Xiusheng Huang, Juntong Wu, Liujie Zhang, Shuo Shang, Weihang Chen\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.03327v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-03327/"}
{"id": "arxiv:2608.03740", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "152e21cd4f1f290aafadfdcfaa10e15a03378404540a46b5f1a5464a2d432d23", "submitted_at": "2026-08-04T14:34:41Z", "updated_at": "2026-09-28T13:07:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03740v2"}, "authors": ["Yu Ran", "Wentao Zhao"], "canonical_source_url": "https://arxiv.org/abs/2608.03740", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-03740/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "152e21cd4f1f290aafadfdcfaa10e15a03378404540a46b5f1a5464a2d432d23", "submitted_at": "2026-08-04T14:34:41Z", "updated_at": "2026-09-28T13:07:55Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.03740v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.03740"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.03740"}], "name": "MissClick: Execution-Aware Adversarial Attacks on Coordinate Generation in GUI Grounding Models", "publication": null, "published_at": "2026-08-04", "record_sha256": "b4be089b3c4264fdd4790dd2f817add0c86af925c7a626785b78aced9ff82b55", "relations": [], "resource_id": "arxiv:2608.03740", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.03740v2", "topics": ["safety-security", "grounding", "robustness", "evaluation"], "updated_at": "2026-09-28"}, "text": "MissClick: Execution-Aware Adversarial Attacks on Coordinate Generation in GUI Grounding Models\nResource type: paper\nReview status: curated\nScope: core\n\nMissClick designs white-box attacks around the numerical effect of coordinate digits on executable GUI clicks. Separate untargeted displacement and targeted place-weighted objectives are evaluated across desktop, web and mobile grounding settings.\nPlatforms: cross-platform\nTopics: safety-security, grounding, robustness, evaluation\nPublished: 2026-08-04\nAuthors: Yu Ran, Wentao Zhao\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.03740v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-03740/"}
{"id": "arxiv:2608.04741", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30e5ce3d86881041491ae507839b2f3d302691aea1d2fdfc92cfc00f35efefe7", "submitted_at": "2026-08-05T12:08:16Z", "updated_at": "2026-08-05T12:08:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04741v1"}, "authors": ["Longtao Guo", "Zelin Zhang", "Kaifeng Huang", "Yang Shi"], "canonical_source_url": "https://arxiv.org/abs/2608.04741", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-04741/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "30e5ce3d86881041491ae507839b2f3d302691aea1d2fdfc92cfc00f35efefe7", "submitted_at": "2026-08-05T12:08:16Z", "updated_at": "2026-08-05T12:08:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04741v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.04741"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.04741"}], "name": "LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents", "publication": null, "published_at": "2026-08-05", "record_sha256": "0d8d477ed11845ca9937b93a938e34b6d19d1bacb76da92ffa845d1e41f4eda6", "relations": [], "resource_id": "arxiv:2608.04741", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.04741v1", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": "2026-08-05"}, "text": "LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces LoginTrap, a black-box indirect-injection evaluation in which attacker-controlled page context and login flows induce unauthorized authentication. It measures end-to-end private-data leakage across web-agent backbones, architectures and defenses.\nPlatforms: web\nTopics: safety-security, robustness, evaluation\nPublished: 2026-08-05\nAuthors: Longtao Guo, Zelin Zhang, Kaifeng Huang, Yang Shi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.04741v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-04741/"}
{"id": "arxiv:2608.04755", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f8b84135f72558eea1fae428df7cb30f62232631de49e690e73010aeca22d4c", "submitted_at": "2026-08-05T12:23:16Z", "updated_at": "2026-08-05T12:23:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04755v1"}, "authors": ["Dongsheng Chen", "Yuxuan Li", "Guanhua Chen", "Jiaxin Zhang", "Xiangyu Zhao", "Lei Ma", "Xin Yao", "Xuetao Wei"], "canonical_source_url": "https://arxiv.org/abs/2608.04755", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-04755/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f8b84135f72558eea1fae428df7cb30f62232631de49e690e73010aeca22d4c", "submitted_at": "2026-08-05T12:23:16Z", "updated_at": "2026-08-05T12:23:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.04755v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.04755"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.04755"}], "name": "\"Allow\" to Achieve, Over-Privileged Inadvertently: The Unintended Cost of Task-Completion-Driven Pop-up Decisions in Mobile GUI Agents", "publication": null, "published_at": "2026-08-05", "record_sha256": "11324d702c79359011f13b569d46b1a49b9b3857ad38feb95dcc014350a81865", "relations": [], "resource_id": "arxiv:2608.04755", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.04755v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-08-05"}, "text": "\"Allow\" to Achieve, Over-Privileged Inadvertently: The Unintended Cost of Task-Completion-Driven Pop-up Decisions in Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nThis permission-literacy study tests whether mobile GUI agents grant only task-necessary permissions. Controlled changes to task context and requester identity expose authorization biases and the trade-off between suppressing unnecessary and legitimate grants.\nPlatforms: unspecified\nTopics: safety-security, evaluation, robustness\nPublished: 2026-08-05\nAuthors: Dongsheng Chen, Yuxuan Li, Guanhua Chen, Jiaxin Zhang, Xiangyu Zhao, Lei Ma, Xin Yao, Xuetao Wei\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.04755v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-04755/"}
{"id": "arxiv:2608.05587", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7934927c82a43e6ae6a246ade1d50c4551cc49d263ca67cfb32906cbec9ea4", "submitted_at": "2026-08-06T04:17:27Z", "updated_at": "2026-08-06T04:17:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05587v1"}, "authors": ["Linqiang Guo", "Wei Liu", "Li Gu", "Yang Wang", "Tse-Hsun", "Chen"], "canonical_source_url": "https://arxiv.org/abs/2608.05587", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-05587/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dc7934927c82a43e6ae6a246ade1d50c4551cc49d263ca67cfb32906cbec9ea4", "submitted_at": "2026-08-06T04:17:27Z", "updated_at": "2026-08-06T04:17:27Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05587v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.05587"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.05587"}], "name": "StepReflect: Structured UI Transition Reflection for Mobile GUI Agents", "publication": null, "published_at": "2026-08-06", "record_sha256": "a70b6949d862194f0c844aa0c81368ff765983d286e2dd326cca146c1cfb2bde", "relations": [], "resource_id": "arxiv:2608.05587", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.05587v1", "topics": ["self-reflection", "state-tracking", "imitation-learning", "reward-models", "evaluation", "efficiency"], "updated_at": "2026-08-06"}, "text": "StepReflect: Structured UI Transition Reflection for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStepReflect predicts structured per-action reflections from transition specifications and paired visual evidence. The paper distinguishes offline transition-label accuracy from online task success when the reflection model is integrated into mobile agents.\nPlatforms: android\nTopics: self-reflection, state-tracking, imitation-learning, reward-models, evaluation, efficiency\nPublished: 2026-08-06\nAuthors: Linqiang Guo, Wei Liu, Li Gu, Yang Wang, Tse-Hsun, Chen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.05587v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-05587/"}
{"id": "arxiv:2608.05891", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce37dd8217afbaa07ddea25c0daf78cdbfb449851ca6ab17b7f3ecacd451eb2f", "submitted_at": "2026-08-06T11:15:41Z", "updated_at": "2026-08-06T11:15:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05891v1"}, "authors": ["Weikai Xu", "Yunren Feng", "Haoxiang Lei", "Kun Huang", "Yuxuan Liu", "Kang Zhao", "Xiaolin Hu", "Shuo Shang", "Bo An"], "canonical_source_url": "https://arxiv.org/abs/2608.05891", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-05891/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ce37dd8217afbaa07ddea25c0daf78cdbfb449851ca6ab17b7f3ecacd451eb2f", "submitted_at": "2026-08-06T11:15:41Z", "updated_at": "2026-08-06T11:15:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.05891v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.05891"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.05891"}], "name": "AppDeltaWorld: Transition-Grounded Delta Code World Model for Mobile GUI Agents", "publication": null, "published_at": "2026-08-06", "record_sha256": "277d2a00e1000943141943d2afb0c1c3d2bf305a0a14574be6f69b29e3ba5d5e", "relations": [], "resource_id": "arxiv:2608.05891", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.05891v1", "topics": ["world-models", "data-generation", "imitation-learning", "reinforcement-learning", "mobile-navigation"], "updated_at": "2026-08-06"}, "text": "AppDeltaWorld: Transition-Grounded Delta Code World Model for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAppDeltaWorld predicts executable code updates for the next mobile UI state, using transition-constrained retrieval and rendered visual assets. It evaluates reconstruction fidelity separately from policy training and adaptation using generated environments.\nPlatforms: android\nTopics: world-models, data-generation, imitation-learning, reinforcement-learning, mobile-navigation\nPublished: 2026-08-06\nAuthors: Weikai Xu, Yunren Feng, Haoxiang Lei, Kun Huang, Yuxuan Liu, Kang Zhao, Xiaolin Hu, Shuo Shang, Bo An\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.05891v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-05891/"}
{"id": "arxiv:2608.06065", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb90627131765746e3a138ab0862dd68ea963106db4055760d78041c856acc8b", "submitted_at": "2026-08-06T14:10:23Z", "updated_at": "2026-08-06T14:10:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06065v1"}, "authors": ["Weiwei Li", "Junzhuo Liu", "Tong Chu", "Hengfu Yu", "Wen Li"], "canonical_source_url": "https://arxiv.org/abs/2608.06065", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-06065/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb90627131765746e3a138ab0862dd68ea963106db4055760d78041c856acc8b", "submitted_at": "2026-08-06T14:10:23Z", "updated_at": "2026-08-06T14:10:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06065v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.06065"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.06065"}], "name": "The Next Screenshot Knows: Gated Hindsight Distillation for Mobile GUI Agents", "publication": null, "published_at": "2026-08-06", "record_sha256": "95bb8b7e805255cd157ce18cb5dc6f7f0d584b3cb25272e10b9e3bc317014570", "relations": [], "resource_id": "arxiv:2608.06065", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.06065v1", "topics": ["imitation-learning", "interaction-history", "mobile-navigation"], "updated_at": "2026-08-06"}, "text": "The Next Screenshot Knows: Gated Hindsight Distillation for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGated Hindsight Distillation gives a teacher access to the next screenshot during training while the student sees only the available trajectory prefix. Distillation is applied when hindsight repairs a student failure, and the resulting policies are evaluated on Android tasks.\nPlatforms: android\nTopics: imitation-learning, interaction-history, mobile-navigation\nPublished: 2026-08-06\nAuthors: Weiwei Li, Junzhuo Liu, Tong Chu, Hengfu Yu, Wen Li\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.06065v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-06065/"}
{"id": "arxiv:2608.06171", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9055ec9c70777fc94179983d82bf841bac12332ebed7f49342883c8d842d0fbc", "submitted_at": "2026-08-06T15:37:04Z", "updated_at": "2026-08-06T15:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06171v1"}, "authors": ["Jiaming Wei", "Zekun Wu", "Adriano Koshiyama", "Maria Perez-Ortiz"], "canonical_source_url": "https://arxiv.org/abs/2608.06171", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-06171/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9055ec9c70777fc94179983d82bf841bac12332ebed7f49342883c8d842d0fbc", "submitted_at": "2026-08-06T15:37:04Z", "updated_at": "2026-08-06T15:37:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.06171v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.06171"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.06171"}], "name": "Routing Is Least Learnable Where It Is Most Valuable: Bounds on Representation Routing for Web Agents", "publication": null, "published_at": "2026-08-06", "record_sha256": "8f2837b0e48a232c5b3436f3764f9da292e8e7827dcdcf006345009bce6f159f", "relations": [], "resource_id": "arxiv:2608.06171", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.06171v1", "topics": ["screen-understanding", "efficiency", "evaluation", "robustness"], "updated_at": "2026-08-06"}, "text": "Routing Is Least Learnable Where It Is Most Valuable: Bounds on Representation Routing for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMeasures the complementarity, cost and rerun variability of six browser observation modes on WebArena and VisualWebArena. Its routing experiments distinguish an oracle cost opportunity from policies that fail to robustly outperform a fixed observation mode.\nPlatforms: web\nTopics: screen-understanding, efficiency, evaluation, robustness\nPublished: 2026-08-06\nAuthors: Jiaming Wei, Zekun Wu, Adriano Koshiyama, Maria Perez-Ortiz\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.06171v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-06171/"}
{"id": "arxiv:2608.07775", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d99685a75d2023db5c9ad943e103ec779eebd037fe7a113848d8a31e08fd68b", "submitted_at": "2026-08-07T21:47:43Z", "updated_at": "2026-08-07T21:47:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.07775v1"}, "authors": ["Xiaoou Liu", "Longchao Da", "Hanyang Chen", "Yuan Ling", "Hua Wei"], "canonical_source_url": "https://arxiv.org/abs/2608.07775", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-07775/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4d99685a75d2023db5c9ad943e103ec779eebd037fe7a113848d8a31e08fd68b", "submitted_at": "2026-08-07T21:47:43Z", "updated_at": "2026-08-07T21:47:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.07775v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:44.898646+00:00", "role": "primary", "sha256": "d4afe27efcce2a3b9401f9677304db3cdedfa66f537293bb742bd1abfb16ee59", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2608.07775v1"}], "identifiers": {"arxiv": "2608.07775"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:40.283922+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.07775"}], "name": "AndroidReality: How Far Are Mobile Agents from the Real World?", "publication": null, "published_at": "2026-08-07", "record_sha256": "c33f1d6f16f58fe3fd2f17477db523aa6050939ccb6697f0872acc9188214038", "relations": [], "resource_id": "arxiv:2608.07775", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.07775v1", "topics": ["evaluation", "robustness", "self-reflection", "mobile-navigation"], "updated_at": "2026-08-07"}, "text": "AndroidReality: How Far Are Mobile Agents from the Real World?\nResource type: paper\nReview status: curated\nScope: core\n\nExtends AndroidWorld with controlled state, transition, and action perturbations to test mobile-agent robustness, and studies a training-free introspective recovery mechanism.\nPlatforms: android\nTopics: evaluation, robustness, self-reflection, mobile-navigation\nPublished: 2026-08-07\nAuthors: Xiaoou Liu, Longchao Da, Hanyang Chen, Yuan Ling, Hua Wei\nRecorded arXiv version: v1\n\nThe reviewed protocol normally activates one perturbation type at a time while preserving task semantics and reward. Results do not establish robustness to arbitrary combined deployment disturbances; experiments were not independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Perturbation formulation; 3.2 Injection; 3.3 Evaluation protocol; 4.3 Error analysis); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2608.07775v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-07775/"}
{"id": "arxiv:2608.08392", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0e469cd7d84080e11df46d826460bea79eae71b8fa5786cacd96aac02b8b0a8", "submitted_at": "2026-08-09T01:08:19Z", "updated_at": "2026-08-09T01:08:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08392v1"}, "authors": ["Zejun Xu", "Taiyi Chen", "Jin Li", "Yongtong Gu", "Qi Cheng", "Aixuan Lv", "Shuai Zhu", "Pengfei Zhu", "Kaichen Yang", "Boyu Sun", "Yixian Yang", "Mulong Xie", "Xin Liu", "Dagang Li", "Xiaoteng Ma", "Hongru Wang"], "canonical_source_url": "https://arxiv.org/abs/2608.08392", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-08392/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d0e469cd7d84080e11df46d826460bea79eae71b8fa5786cacd96aac02b8b0a8", "submitted_at": "2026-08-09T01:08:19Z", "updated_at": "2026-08-09T01:08:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08392v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.08392"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.08392"}], "name": "CAP: A Scalable Benchmark for Evaluating Cross-Site Browser Agents with Complex Actions and Perception", "publication": null, "published_at": "2026-08-09", "record_sha256": "775d5d9559f51c8fdb6c079aed5db3b1465fdd7e0570ddaee8ba64fa58a76a0f", "relations": [], "resource_id": "arxiv:2608.08392", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.08392v1", "topics": ["evaluation", "web-navigation", "cross-app", "screen-understanding"], "updated_at": "2026-08-09"}, "text": "CAP: A Scalable Benchmark for Evaluating Cross-Site Browser Agents with Complex Actions and Perception\nResource type: paper\nReview status: curated\nScope: core\n\nCAP builds cross-site browser workflows by decomposing websites into operation and perception requirements and recomposing them into tasks. Its judge-based evaluation diagnoses complex interactions and visual-perception failures across real websites.\nPlatforms: web\nTopics: evaluation, web-navigation, cross-app, screen-understanding\nPublished: 2026-08-09\nAuthors: Zejun Xu, Taiyi Chen, Jin Li, Yongtong Gu, Qi Cheng, Aixuan Lv, Shuai Zhu, Pengfei Zhu, Kaichen Yang, Boyu Sun, Yixian Yang, Mulong Xie, Xin Liu, Dagang Li, Xiaoteng Ma, Hongru Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.08392v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-08392/"}
{"id": "arxiv:2608.08939", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e4845ce85dd67b7790cf5de357f12deb72b71eb72031298e1064809099f063", "submitted_at": "2026-08-09T22:13:55Z", "updated_at": "2026-08-09T22:13:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08939v1"}, "authors": ["Rahul Deivasigamani", "Sayeda Faatin Alvi", "Derqui Andrea", "Kaushal Punjabi", "Stjepan Picek"], "canonical_source_url": "https://arxiv.org/abs/2608.08939", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-08939/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64e4845ce85dd67b7790cf5de357f12deb72b71eb72031298e1064809099f063", "submitted_at": "2026-08-09T22:13:55Z", "updated_at": "2026-08-09T22:13:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.08939v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2608.08939"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.08939"}], "name": "Not an A11y: How Android Accessibility Exposes Mobile AI Agents to Indirect Prompt Injection", "publication": null, "published_at": "2026-08-09", "record_sha256": "40effabc31ff43eccc927b0705eaa68ca3a6827aca5fa75fe8bb832df3018afc", "relations": [], "resource_id": "arxiv:2608.08939", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.08939v1", "topics": ["safety-security", "mobile-navigation", "screen-understanding"], "updated_at": "2026-08-09"}, "text": "Not an A11y: How Android Accessibility Exposes Mobile AI Agents to Indirect Prompt Injection\nResource type: paper\nReview status: curated\nScope: core\n\nDemonstrates that unsanitized Android accessibility tree metadata exposes mobile GUI agents to stealthy indirect prompt injection attacks.\nPlatforms: android\nTopics: safety-security, mobile-navigation, screen-understanding\nPublished: 2026-08-09\nAuthors: Rahul Deivasigamani, Sayeda Faatin Alvi, Derqui Andrea, Kaushal Punjabi, Stjepan Picek\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.08939v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-08939/"}
{"id": "arxiv:2608.09654", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f05f128cccb514987b7defb52c8f43b45d917e0e0f9570db4e75c4fe23b38533", "submitted_at": "2026-08-10T14:29:07Z", "updated_at": "2026-08-10T14:29:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09654v1"}, "authors": ["Yuke Li", "Xuehan Hou"], "canonical_source_url": "https://arxiv.org/abs/2608.09654", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-09654/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f05f128cccb514987b7defb52c8f43b45d917e0e0f9570db4e75c4fe23b38533", "submitted_at": "2026-08-10T14:29:07Z", "updated_at": "2026-08-10T14:29:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09654v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.09654"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09654"}], "name": "Hallucination-Free GUI Grounding via Regression-Free Layout-Aware Matching", "publication": null, "published_at": "2026-08-10", "record_sha256": "c5ee5a660c374a9db3e3f8a5c62c02f7975827db38c8dd3a2068640116a2a734", "relations": [], "resource_id": "arxiv:2608.09654", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.09654v1", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-08-10"}, "text": "Hallucination-Free GUI Grounding via Regression-Free Layout-Aware Matching\nResource type: paper\nReview status: curated\nScope: core\n\nThis framework separates frozen-model instruction expansion from a trained layout-aware matcher that localizes GUI elements without coordinate regression. Binary text/icon supervision supports the matcher, with evaluations on GUI grounding and Mind2Web element-selection tasks.\nPlatforms: web\nTopics: grounding, screen-understanding, evaluation\nPublished: 2026-08-10\nAuthors: Yuke Li, Xuehan Hou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.09654v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-09654/"}
{"id": "arxiv:2608.09723", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87e9b3463232f606aea60df7d86626d5aa188e774b0fe515af43e1be49b1c66b", "submitted_at": "2026-08-10T15:23:52Z", "updated_at": "2026-08-10T15:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09723v1"}, "authors": ["Renshan Zhang", "Haoyang Meng", "Yixiao He", "Rui Shao", "April Hua Liu", "Liqiang Nie"], "canonical_source_url": "https://arxiv.org/abs/2608.09723", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-09723/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87e9b3463232f606aea60df7d86626d5aa188e774b0fe515af43e1be49b1c66b", "submitted_at": "2026-08-10T15:23:52Z", "updated_at": "2026-08-10T15:23:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09723v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.09723"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09723"}], "name": "LookAgain: Closed-Loop GUI Grounding with Visually Grounded Reflection", "publication": null, "published_at": "2026-08-10", "record_sha256": "928dbe30b8e9ef94646ab74556e350f8083f339c5aabe7053590f22d3cfe20e0", "relations": [], "resource_id": "arxiv:2608.09723", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.09723v1", "topics": ["grounding", "self-reflection", "reinforcement-learning", "imitation-learning"], "updated_at": "2026-08-10"}, "text": "LookAgain: Closed-Loop GUI Grounding with Visually Grounded Reflection\nResource type: paper\nReview status: curated\nScope: core\n\nLookAgain treats a GUI coordinate as a revisable hypothesis, rendering its marker and a local screenshot patch before confirmation. Reflective supervised trajectories provide a cold start, followed by GRPO using terminal grounding correctness.\nPlatforms: unspecified\nTopics: grounding, self-reflection, reinforcement-learning, imitation-learning\nPublished: 2026-08-10\nAuthors: Renshan Zhang, Haoyang Meng, Yixiao He, Rui Shao, April Hua Liu, Liqiang Nie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.09723v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-09723/"}
{"id": "arxiv:2608.09944", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5ccb0d41798a1a3236e6f9b4361814cccdd40facf9b43bf0bf52932337484dc", "submitted_at": "2026-07-01T16:24:07Z", "updated_at": "2026-07-01T16:24:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09944v1"}, "authors": ["Santosh Patapati"], "canonical_source_url": "https://arxiv.org/abs/2608.09944", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-09944/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5ccb0d41798a1a3236e6f9b4361814cccdd40facf9b43bf0bf52932337484dc", "submitted_at": "2026-07-01T16:24:07Z", "updated_at": "2026-07-01T16:24:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.09944v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:45.064380+00:00", "role": "primary", "sha256": "f26fba818a44c6a7f541ab3b0341ca8f4798d837a425a2ac9fb0fc47a553b7bc", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:45.064380+00:00", "role": "primary", "sha256": "f26fba818a44c6a7f541ab3b0341ca8f4798d837a425a2ac9fb0fc47a553b7bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=200&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:26.923383+00:00", "role": "primary", "sha256": "777197864eeaa34d180bf6b91f32408b7b3d8cc5c9faecc95eca276f5bd426de", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.09944v1"}], "identifiers": {"arxiv": "2608.09944"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:45.064380+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.09944"}], "name": "Navigation Alone Is Not Enough: Evaluating Explanatory Assistive UI Agents", "publication": null, "published_at": "2026-07-01", "record_sha256": "4e3e0275c56f6e1e4a195c8035c208536470e1ad6d081ddb694887a57609c9a4", "relations": [], "resource_id": "arxiv:2608.09944", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.09944v1", "topics": ["evaluation", "human-in-the-loop", "web-navigation"], "updated_at": "2026-07-01"}, "text": "Navigation Alone Is Not Enough: Evaluating Explanatory Assistive UI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces NeXUI, an assistive UI-agent benchmark that evaluates navigation together with interface-grounded explanations and user oversight in recorded interface states.\nPlatforms: web\nTopics: evaluation, human-in-the-loop, web-navigation\nPublished: 2026-07-01\nAuthors: Santosh Patapati\nRecorded arXiv version: v1\n\nPackaged tasks use recorded linear transition graphs. Reported model experiments use a text-input validation subset, so their results are not evidence of screenshot-based navigation performance. No results were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.3 Interface states and task annotations; 3.4 Evaluation; 4.1 Models and experimental setup; 5 Discussion); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2608.09944v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-09944/"}
{"id": "arxiv:2608.10366", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1089e57e0f0146f29a4a411c5c6f72da3c9f45be223fbbc5c9630315be6a680", "submitted_at": "2026-08-11T01:45:56Z", "updated_at": "2026-08-11T01:45:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10366v1"}, "authors": ["Mizanur Rahman", "Mohammed Saidul Islam", "Ridwan Mahbub", "Md Tahmid Rahman Laskar", "Shafiq Joty", "Enamul Hoque Prince"], "canonical_source_url": "https://arxiv.org/abs/2608.10366", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-10366/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1089e57e0f0146f29a4a411c5c6f72da3c9f45be223fbbc5c9630315be6a680", "submitted_at": "2026-08-11T01:45:56Z", "updated_at": "2026-08-11T01:45:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.10366v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:09Z", "role": "primary", "sha256": "65cc3856c70a0fb267f9996872e077de10e9f54821ab1ffb1d82f0a92fa11dbf", "supports": ["name", "kind", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2607.05277%2C2607.06118%2C2607.06306%2C2607.07820%2C2607.11185%2C2607.11388%2C2607.12625%2C2607.12640%2C2607.13027%2C2607.14277%2C2607.14443%2C2607.14548%2C2607.15193%2C2607.15550%2C2607.17050%2C2607.18046%2C2607.19865%2C2607.21557%2C2607.22798%2C2607.23263%2C2607.24112%2C2607.24167%2C2607.25904%2C2607.26041%2C2607.27155%2C2607.28074%2C2607.28227%2C2607.28609%2C2607.29320%2C2608.01964%2C2608.02352%2C2608.02643%2C2608.03270%2C2608.03327%2C2608.04530%2C2608.04741%2C2608.05587%2C2608.05784%2C2608.05891%2C2608.06065%2C2608.06171%2C2608.06477%2C2608.09278%2C2608.09654%2C2608.10366%2C2608.10692%2C2608.11588%2C2608.12429%2C2608.15930%2C2608.17310&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:26Z", "role": "primary", "sha256": "2158cabc9a6211e0b75e8e3307f80ebc0d42c42fccc03d0143dcf574c2f663db", "supports": ["summary", "environments", "scope"], "url": "https://github.com/vis-nlp/DSAgentBench"}], "identifiers": {"arxiv": "2608.10366"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:09Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.10366"}], "name": "DSAgentBench: Can Agents Automate End-to-End Data-Science Workflows in Real Computer Environments?", "publication": null, "published_at": "2026-08-11", "record_sha256": "f4ae1d40b3c0c9a1e6cd9fabb570da0c57dbb91e4f6b462a00d0846c92c2b499", "relations": [], "resource_id": "arxiv:2608.10366", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.10366v1", "topics": ["evaluation", "workflow", "science", "long-horizon"], "updated_at": "2026-08-11"}, "text": "DSAgentBench: Can Agents Automate End-to-End Data-Science Workflows in Real Computer Environments?\nResource type: paper\nReview status: curated\nScope: core\n\nDSAgentBench evaluates end-to-end data-science workflows spanning analysis, modeling and visualization inside a real desktop environment. Task evaluators check produced analytical and visual artifacts, while the released interface supports screenshot or accessibility observations and GUI actions.\nPlatforms: desktop, linux\nTopics: evaluation, workflow, science, long-horizon\nPublished: 2026-08-11\nAuthors: Mizanur Rahman, Mohammed Saidul Islam, Ridwan Mahbub, Md Tahmid Rahman Laskar, Shafiq Joty, Enamul Hoque Prince\nRecorded arXiv version: v1\n\nThe paper abstract and official project README were reviewed. The README documents an Ubuntu desktop and GUI action interface and includes LLM-assisted evaluators; this annotation does not assert that every evaluator is deterministic. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.10366v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-10366/"}
{"id": "arxiv:2608.11434", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2606ae96ca3ea99b5c7278485601f9bde3b64eefdd3bfecc2bbfc678c39b4fc", "submitted_at": "2026-08-11T21:00:46Z", "updated_at": "2026-08-11T21:00:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.11434v1"}, "authors": ["Ziqiang Wang", "Li Gu", "Zhixiang Chi", "Zhi Liu", "Seyed Mehdi Ayyoubzadeh", "Yuanhao Yu", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2608.11434", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-11434/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a2606ae96ca3ea99b5c7278485601f9bde3b64eefdd3bfecc2bbfc678c39b4fc", "submitted_at": "2026-08-11T21:00:46Z", "updated_at": "2026-08-11T21:00:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.11434v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:32Z", "role": "primary", "sha256": "5128028eea81e3ae80715cb038513012598c0c81a7cbfe8c0d4d6adb6c00e518", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2607.03726%2C2607.05120%2C2607.05465%2C2607.05743%2C2607.06008%2C2607.06595%2C2607.07403%2C2607.07744%2C2607.08252%2C2607.08964%2C2607.10455%2C2607.10608%2C2607.12406%2C2607.12756%2C2607.13085%2C2607.13594%2C2607.13884%2C2607.13988%2C2607.15899%2C2607.16352%2C2607.16636%2C2607.16900%2C2607.17044%2C2607.20064%2C2607.20121%2C2607.21522%2C2607.22738%2C2607.23722%2C2607.24720%2C2607.26300%2C2607.26769%2C2607.27881%2C2607.28229%2C2607.28272%2C2607.28430%2C2608.00548%2C2608.02163%2C2608.03223%2C2608.03451%2C2608.03499%2C2608.05013%2C2608.05430%2C2608.06811%2C2608.07873%2C2608.09268%2C2608.10357%2C2608.10669%2C2608.11274%2C2608.11434%2C2608.15089&max_results=50"}], "identifiers": {"arxiv": "2608.11434"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.11434"}], "name": "Benchmarking LLM Judges for Mobile Agent Evaluation", "publication": null, "published_at": "2026-08-11", "record_sha256": "dbeadd9f41ba67c4061cebacf577726e55fbba00356134810f41edfd96cd60b5", "relations": [], "resource_id": "arxiv:2608.11434", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.11434v1", "topics": ["evaluation", "reward-models", "robustness"], "updated_at": "2026-08-11"}, "text": "Benchmarking LLM Judges for Mobile Agent Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nMobileJudgeBench compares completion judges against human labels on mobile-agent trajectories. It studies screenshot-sampling baselines, model-dependent error profiles and how judge quality relates to agent rankings and reinforcement-learning rewards.\nPlatforms: unspecified\nTopics: evaluation, reward-models, robustness\nPublished: 2026-08-11\nAuthors: Ziqiang Wang, Li Gu, Zhixiang Chi, Zhi Liu, Seyed Mehdi Ayyoubzadeh, Yuanhao Yu, Yang Wang\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2608.11434v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-11434/"}
{"id": "arxiv:2608.12429", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf543ef0b77eb6fe430c8cef5382932f13e4e4136aaba9c14c44b5a591d6612a", "submitted_at": "2026-08-12T12:03:20Z", "updated_at": "2026-08-12T12:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.12429v1"}, "authors": ["Ruitao Wang", "Yuwen Hao", "Menglin Yang"], "canonical_source_url": "https://arxiv.org/abs/2608.12429", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-12429/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bf543ef0b77eb6fe430c8cef5382932f13e4e4136aaba9c14c44b5a591d6612a", "submitted_at": "2026-08-12T12:03:20Z", "updated_at": "2026-08-12T12:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.12429v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.12429"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.12429"}], "name": "SynWeaver: Website-Prior Task and Trajectory Co-Synthesis for Web Agents", "publication": null, "published_at": "2026-08-12", "record_sha256": "6a03c07accc8e2fce292e8172790edf0378ba6cded0e216cd58f21a64b5e9d8a", "relations": [], "resource_id": "arxiv:2608.12429", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.12429v1", "topics": ["data-generation", "web-navigation", "state-tracking", "robustness"], "updated_at": "2026-08-12"}, "text": "SynWeaver: Website-Prior Task and Trajectory Co-Synthesis for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSynWeaver explores a website to build a map of page states and executable interactions, then learns website priors for task proposals. Joint task/trajectory synthesis revises inconsistent pairs and verifies or repairs the resulting supervision.\nPlatforms: web\nTopics: data-generation, web-navigation, state-tracking, robustness\nPublished: 2026-08-12\nAuthors: Ruitao Wang, Yuwen Hao, Menglin Yang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.12429v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-12429/"}
{"id": "arxiv:2608.14131", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8a6bd69d421bdfb9a21c22e715a59d0afb0b0cab75510f28ff569903ecffae5", "submitted_at": "2026-08-14T09:36:33Z", "updated_at": "2026-08-14T09:36:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14131v1"}, "authors": ["Thilo Reintjes", "Sivajeet Chand", "Derui Zhu", "Sushant Kumar Pandey", "Alexander Pretschner"], "canonical_source_url": "https://arxiv.org/abs/2608.14131", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-14131/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f8a6bd69d421bdfb9a21c22e715a59d0afb0b0cab75510f28ff569903ecffae5", "submitted_at": "2026-08-14T09:36:33Z", "updated_at": "2026-08-14T09:36:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14131v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.14131"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.14131"}], "name": "LegacyWorld: Atomicity-Aware Evaluation of GUI Agents for Legacy Workflows", "publication": null, "published_at": "2026-08-14", "record_sha256": "6e743330a8b696ac5e59c35e85ec09ef3d9ee806e59278c416b5a30dafb4fdb4", "relations": [], "resource_id": "arxiv:2608.14131", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.14131v1", "topics": ["evaluation", "safety-security", "workflow", "recording"], "updated_at": "2026-08-14"}, "text": "LegacyWorld: Atomicity-Aware Evaluation of GUI Agents for Legacy Workflows\nResource type: paper\nReview status: curated\nScope: core\n\nLegacyWorld evaluates Windows GUI workflows under an atomicity criterion: finish correctly or fail without persistent unintended changes. It compares expert prompts with prompts derived from recorded golden paths and separates useful completion, safe failure and non-atomic side effects.\nPlatforms: desktop, windows\nTopics: evaluation, safety-security, workflow, recording\nPublished: 2026-08-14\nAuthors: Thilo Reintjes, Sivajeet Chand, Derui Zhu, Sushant Kumar Pandey, Alexander Pretschner\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.14131v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-14131/"}
{"id": "arxiv:2608.14132", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2660e01921a78bed7280b34f35a1b587e62339c2e280d6294bbb6d6e666786e", "submitted_at": "2026-08-14T09:36:59Z", "updated_at": "2026-08-14T09:36:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14132v1"}, "authors": ["Xiaokai Yan", "Jingtao Ding", "Yong Li", "Zhiwen Yu"], "canonical_source_url": "https://arxiv.org/abs/2608.14132", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-14132/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e2660e01921a78bed7280b34f35a1b587e62339c2e280d6294bbb6d6e666786e", "submitted_at": "2026-08-14T09:36:59Z", "updated_at": "2026-08-14T09:36:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.14132v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:40.283922+00:00", "role": "primary", "sha256": "ddd5bd966c7430bd9d4fd0269e7d3d46d8839fa5bfa57d601b97b4dc735cfc9a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:05:22.211464+00:00", "role": "primary", "sha256": "8d6aa8484ff4a7cbaac589c2c220407ebf6ecd1e216015a248125abdb0f2ed2d", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.14132v1"}], "identifiers": {"arxiv": "2608.14132"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:40.283922+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.14132"}], "name": "Act2Intention: A Benchmark For Developing Active Mobile Agents Through Inferring User Intention from GUI Actions", "publication": null, "published_at": "2026-08-14", "record_sha256": "7c01278dcdc4fc8536d9c5d6f77d5ebcd981778b70a42c2873f89d011dc37c13", "relations": [], "resource_id": "arxiv:2608.14132", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.14132v1", "topics": ["human-in-the-loop", "planning", "mobile-navigation", "evaluation"], "updated_at": "2026-08-14"}, "text": "Act2Intention: A Benchmark For Developing Active Mobile Agents Through Inferring User Intention from GUI Actions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces intention understanding, prediction, and confirmed execution tasks over mobile GUI trajectories, with a proactive agent combining personalized intention models and experience-guided actions.\nPlatforms: android\nTopics: human-in-the-loop, planning, mobile-navigation, evaluation\nPublished: 2026-08-14\nAuthors: Xiaokai Yan, Jingtao Ding, Yong Li, Zhiwen Yu\nRecorded arXiv version: v1\n\nThe benchmark combines real and generated trajectory material. Opportune intervention timing and real-user acceptance are stated limitations. Section 4.2.2 specifies Android emulator instances for generated action trajectories. No results were independently reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3 Task definition; 4.1 Raw data collection; 4.2 Trajectory construction; 7 Limitations and future directions); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2608.14132v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-14132/"}
{"id": "arxiv:2608.15930", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efe635792c696bca73c9a436be67d8528582fb239d4d4f8d3de531219b1f4f8e", "submitted_at": "2026-08-16T20:59:59Z", "updated_at": "2026-08-16T20:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.15930v1"}, "authors": ["Zihan Ding", "Longxu Dou", "Qi Gao", "Xiangwu Guo", "Shengchao Hu", "Zilong Huang", "Zihang Jiang", "Lei Ke", "Mengcheng Lan", "Weixian Lei", "Hanxuan Li", "Honglin Li", "Xiyun Li", "Zaitang Li", "Leowei Liang", "Xin Luo", "Haozhe Ma", "Jiayi Mao", "Zhoujie Pan", "Can Qin", "Tianyuan Qu", "Weiqi Wang", "Wenkai Wang", "Yonglin Wang", "Yuxin Wang", "Chenxu Wu", "Yingchen Yu", "Chenyu Zhang", "Yuhao Zheng"], "canonical_source_url": "https://arxiv.org/abs/2608.15930", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-15930/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "efe635792c696bca73c9a436be67d8528582fb239d4d4f8d3de531219b1f4f8e", "submitted_at": "2026-08-16T20:59:59Z", "updated_at": "2026-08-16T20:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.15930v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.15930"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.15930"}], "name": "UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations", "publication": null, "published_at": "2026-08-16", "record_sha256": "f8cc76f793eaaa6d9f711e71da76741e048fc9acfe0cfb2ffe464cd32b217ed5", "relations": [], "resource_id": "arxiv:2608.15930", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.15930v1", "topics": ["workflow", "data-generation", "long-horizon"], "updated_at": "2026-08-16"}, "text": "UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nCombines environment-grounded training with demonstration-conditioned workflows, and introduces OSWorkerBench for extended office-task evaluation.\nPlatforms: desktop, windows\nTopics: workflow, data-generation, long-horizon\nPublished: 2026-08-16\nAuthors: Zihan Ding, Longxu Dou, Qi Gao, Xiangwu Guo, Shengchao Hu, Zilong Huang, Zihang Jiang, Lei Ke, Mengcheng Lan, Weixian Lei, Hanxuan Li, Honglin Li, Xiyun Li, Zaitang Li, Leowei Liang, Xin Luo, Haozhe Ma, Jiayi Mao, Zhoujie Pan, Can Qin, Tianyuan Qu, Weiqi Wang, Wenkai Wang, Yonglin Wang, Yuxin Wang, Chenxu Wu, Yingchen Yu, Chenyu Zhang, Yuhao Zheng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.15930v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-15930/"}
{"id": "arxiv:2608.17319", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "474187aaa79165dc8652eec1e2af6b1733b2ee37b3cd125c89ff0defcaf2e193", "submitted_at": "2026-08-18T03:23:31Z", "updated_at": "2026-08-18T03:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17319v1"}, "authors": ["AIMAE Team", "Tianxiang Chen", "Yan Cheng", "Zhangye Han", "Xiaowei Li", "Chang Liu", "Cheng Liu", "Zhongqiang Ma", "Long Peng", "Xiaobing Tu", "Yinggui Wang", "Hongliang Wei", "Chen Wu", "Daiping Xin", "Kunyu Zhou", "Pengyang Zhou", "Peiyuan Chen", "Ziyuan Chen", "Yutao Deng", "Chunyu Dong", "Xiangyu Fu", "Yicheng Feng", "Ruian He", "Haochen Li", "Miancan Liu", "Zhengqin Liu", "Wei Peng", "Jinkui Ren", "Haoyu Tan", "Dong Xiao", "Rongkun Xue", "Shujian Yang", "Xianhang Ye", "Ziqi Yuan", "Ziyang Yu", "Linghan Zhang", "Xiantao Zhang", "Xuanpu Zhao", "Yinan Zhao", "Zhenghui Zhao", "Bin Zhu", "Likai Zou"], "canonical_source_url": "https://arxiv.org/abs/2608.17319", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-17319/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "474187aaa79165dc8652eec1e2af6b1733b2ee37b3cd125c89ff0defcaf2e193", "submitted_at": "2026-08-18T03:23:31Z", "updated_at": "2026-08-18T03:23:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17319v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.17319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.17319"}], "name": "Wuying-Browser-Agent: Real-World Centric Fundamental Long-Horizon Browser Agents", "publication": null, "published_at": "2026-08-18", "record_sha256": "036ab326f1f365d730eb3fcc663b18a971a53a0f9a05ad33c8d27600d5ef726f", "relations": [], "resource_id": "arxiv:2608.17319", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.17319v1", "topics": ["web-navigation", "long-horizon", "reinforcement-learning"], "updated_at": "2026-08-18"}, "text": "Wuying-Browser-Agent: Real-World Centric Fundamental Long-Horizon Browser Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCombines a browser execution framework, recovery-oriented supervision, and online reinforcement learning, with a bilingual live-web task benchmark.\nPlatforms: web\nTopics: web-navigation, long-horizon, reinforcement-learning\nPublished: 2026-08-18\nAuthors: AIMAE Team, Tianxiang Chen, Yan Cheng, Zhangye Han, Xiaowei Li, Chang Liu, Cheng Liu, Zhongqiang Ma, Long Peng, Xiaobing Tu, Yinggui Wang, Hongliang Wei, Chen Wu, Daiping Xin, Kunyu Zhou, Pengyang Zhou, Peiyuan Chen, Ziyuan Chen, Yutao Deng, Chunyu Dong, Xiangyu Fu, Yicheng Feng, Ruian He, Haochen Li, Miancan Liu, Zhengqin Liu, Wei Peng, Jinkui Ren, Haoyu Tan, Dong Xiao, Rongkun Xue, Shujian Yang, Xianhang Ye, Ziqi Yuan, Ziyang Yu, Linghan Zhang, Xiantao Zhang, Xuanpu Zhao, Yinan Zhao, Zhenghui Zhao, Bin Zhu, Likai Zou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.17319v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-17319/"}
{"id": "arxiv:2608.17659", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c49a67b38314152b83e793b9ade7ce4102a44aa27c331504a8758b94df906c67", "submitted_at": "2026-08-18T11:33:22Z", "updated_at": "2026-08-18T11:33:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17659v1"}, "authors": ["Sujin Chen", "Lijun Li", "Tianyi Du", "Jing Shao"], "canonical_source_url": "https://arxiv.org/abs/2608.17659", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-17659/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c49a67b38314152b83e793b9ade7ce4102a44aa27c331504a8758b94df906c67", "submitted_at": "2026-08-18T11:33:22Z", "updated_at": "2026-08-18T11:33:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.17659v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.17659"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.17659"}], "name": "MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps", "publication": null, "published_at": "2026-08-18", "record_sha256": "6359e1cf9ed90c25704534fc92acd07999a57c4a9856e0e5f1ebf078df14c0ee", "relations": [], "resource_id": "arxiv:2608.17659", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.17659v1", "topics": ["safety-security", "evaluation"], "updated_at": "2026-08-18"}, "text": "MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates environmental injection risks in Android tasks using programmatic risk checks and model-based adjudication of ambiguous cases.\nPlatforms: android\nTopics: safety-security, evaluation\nPublished: 2026-08-18\nAuthors: Sujin Chen, Lijun Li, Tianyi Du, Jing Shao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.17659v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-17659/"}
{"id": "arxiv:2608.18307", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de4889c021f35c169092e17801fec414cce21b37c74a4ff72af6a6c67a1a0a25", "submitted_at": "2026-08-18T20:38:26Z", "updated_at": "2026-08-18T20:38:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.18307v1"}, "authors": ["Tianchen Guan", "Xinlei Lin", "Royce Cheng-Yue", "Xiangjun Wang", "Shuyan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2608.18307", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-18307/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "de4889c021f35c169092e17801fec414cce21b37c74a4ff72af6a6c67a1a0a25", "submitted_at": "2026-08-18T20:38:26Z", "updated_at": "2026-08-18T20:38:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.18307v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.18307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.18307"}], "name": "ComponentBench: Diagnosing Component-Level Failures in Computer-Use Agents", "publication": null, "published_at": "2026-08-18", "record_sha256": "56d73cb486651fa3d2cd98aec760bdccd0b6fba87e17380cb3dc1e9481768fa1", "relations": [], "resource_id": "arxiv:2608.18307", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.18307v1", "topics": ["evaluation", "screen-understanding", "efficiency", "robustness"], "updated_at": "2026-08-18"}, "text": "ComponentBench: Diagnosing Component-Level Failures in Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nComponentBench evaluates short interactions with canonical web UI components using verified tasks and matched human reference trajectories. Its diagnostic pipeline compares observation/action representations and analyzes structural difficulty, failure patterns and interaction efficiency.\nPlatforms: web\nTopics: evaluation, screen-understanding, efficiency, robustness\nPublished: 2026-08-18\nAuthors: Tianchen Guan, Xinlei Lin, Royce Cheng-Yue, Xiangjun Wang, Shuyan Zhou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.18307v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-18307/"}
{"id": "arxiv:2608.20319", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85b3a99d551fcdffdd306d839b39cdfbe7f0f7654aa9adbf81f939aa874e3b9", "submitted_at": "2026-08-20T17:57:00Z", "updated_at": "2026-08-20T17:57:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.20319v1"}, "authors": ["Yucheng Jiang", "Zora Zhiruo Wang", "Ruishi Chen", "Diyi Yang"], "canonical_source_url": "https://arxiv.org/abs/2608.20319", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-20319/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a85b3a99d551fcdffdd306d839b39cdfbe7f0f7654aa9adbf81f939aa874e3b9", "submitted_at": "2026-08-20T17:57:00Z", "updated_at": "2026-08-20T17:57:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.20319v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.20319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.20319"}], "name": "Inducing Task Models from Computer-Use Traces", "publication": null, "published_at": "2026-08-20", "record_sha256": "620656547a75991984cacfca794f54b8d7496a3213a9517511439385ad326876", "relations": [], "resource_id": "arxiv:2608.20319", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.20319v1", "topics": ["recording", "workflow", "memory", "interaction-history"], "updated_at": "2026-08-20"}, "text": "Inducing Task Models from Computer-Use Traces\nResource type: paper\nReview status: curated\nScope: core\n\nTask Model Induction disentangles concurrent activities in screenshot and mouse/keyboard traces, then constructs hierarchical goals and procedural control flow for each task. Evaluation tests reconstruction of observed execution and reuse of derived skills on held-out tasks.\nPlatforms: unspecified\nTopics: recording, workflow, memory, interaction-history\nPublished: 2026-08-20\nAuthors: Yucheng Jiang, Zora Zhiruo Wang, Ruishi Chen, Diyi Yang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.20319v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-20319/"}
{"id": "arxiv:2608.21794", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cabd094270d01b20df7b029be29c5d460d10beb84e43e3e783f4f11997b74114", "submitted_at": "2026-08-22T06:30:03Z", "updated_at": "2026-08-22T06:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21794v1"}, "authors": ["Qijia Chen", "Giulio Jacucci"], "canonical_source_url": "https://arxiv.org/abs/2608.21794", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-21794/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cabd094270d01b20df7b029be29c5d460d10beb84e43e3e783f4f11997b74114", "submitted_at": "2026-08-22T06:30:03Z", "updated_at": "2026-08-22T06:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21794v1"}}, "environments": ["web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.21794"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21794"}], "name": "Lexical Coupling in GUI Element Grounding: Sentence Embeddings Track Labels across Mobile and Web", "publication": null, "published_at": "2026-08-22", "record_sha256": "61fb4f832f96b80e8f7d593da8ce8f5af04e4a9bfce9dfb84aecc7f5c4025ae4", "relations": [], "resource_id": "arxiv:2608.21794", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.21794v1", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-22"}, "text": "Lexical Coupling in GUI Element Grounding: Sentence Embeddings Track Labels across Mobile and Web\nResource type: paper\nReview status: curated\nScope: core\n\nStudies whether text-embedding GUI grounding scores reflect semantic matching or recovery of visible labels. Same-screen ranking comparisons with lexical baselines expose differences between deployable fusion, oracle selection, and performance on label-poor targets.\nPlatforms: web, cross-platform\nTopics: grounding, evaluation, screen-understanding\nPublished: 2026-08-22\nAuthors: Qijia Chen, Giulio Jacucci\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.21794v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-21794/"}
{"id": "arxiv:2608.21830", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9434f8da41679177f6978c039bd42fcacce543de2bd0e4ce05b62ab4d61ed761", "submitted_at": "2026-08-22T08:01:47Z", "updated_at": "2026-08-22T08:01:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21830v1"}, "authors": ["Chengyang Gu", "Le Zhang", "Jingbo Zhou", "Yize Chen", "Yu Shi", "Siqi Bao", "Zheng-Fan Wu", "Hua Wu", "Hui Xiong"], "canonical_source_url": "https://arxiv.org/abs/2608.21830", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-21830/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9434f8da41679177f6978c039bd42fcacce543de2bd0e4ce05b62ab4d61ed761", "submitted_at": "2026-08-22T08:01:47Z", "updated_at": "2026-08-22T08:01:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.21830v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.21830"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21830"}], "name": "Beyond Success and Failure: Length-Aware Contrastive Learning for GUI Agents", "publication": null, "published_at": "2026-08-22", "record_sha256": "e7792f1f0188ff02c748cfeaecda96ef93d4b490527553246c8c732f2e5e642f", "relations": [], "resource_id": "arxiv:2608.21830", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.21830v1", "topics": ["reinforcement-learning", "efficiency", "action-prediction"], "updated_at": "2026-08-22"}, "text": "Beyond Success and Failure: Length-Aware Contrastive Learning for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces LACL-GUI, a contrastive learning objective that distinguishes trajectory quality within both successful and failed GUI executions. It favors concise successes and uses divergence from successful trajectories to differentiate failures during policy optimization.\nPlatforms: unspecified\nTopics: reinforcement-learning, efficiency, action-prediction\nPublished: 2026-08-22\nAuthors: Chengyang Gu, Le Zhang, Jingbo Zhou, Yize Chen, Yu Shi, Siqi Bao, Zheng-Fan Wu, Hua Wu, Hui Xiong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.21830v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-21830/"}
{"id": "arxiv:2608.21832", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "970441704314fabffc46cb4a340dd6e0bde0e84ab9b7fcd2db2c9a8be9dc61d8", "submitted_at": "2026-08-22T08:04:59Z", "updated_at": "2026-08-27T07:03:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21832v2"}, "authors": ["Md Abrar Jahin", "Md Rizwan Parvez"], "canonical_source_url": "https://arxiv.org/abs/2608.21832", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-21832/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "970441704314fabffc46cb4a340dd6e0bde0e84ab9b7fcd2db2c9a8be9dc61d8", "submitted_at": "2026-08-22T08:04:59Z", "updated_at": "2026-08-27T07:03:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21832v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.21832"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21832"}], "name": "GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding", "publication": null, "published_at": "2026-08-22", "record_sha256": "0bbff2e9232f5bda60316538df23a12d0f7ef86170a78f43078fc96d5844a785", "relations": [], "resource_id": "arxiv:2608.21832", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.21832v2", "topics": ["grounding", "evaluation", "screen-understanding"], "updated_at": "2026-08-27"}, "text": "GUI-Primitives: Diagnosing Spatial Reasoning Failures in Vision-Language GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-Primitives, a paired-instruction benchmark that changes a spatial relation while holding the screenshot and anchor fixed. Its diagnostics separate failure to localize either candidate from failure to choose the candidate satisfying the relation.\nPlatforms: unspecified\nTopics: grounding, evaluation, screen-understanding\nPublished: 2026-08-22\nAuthors: Md Abrar Jahin, Md Rizwan Parvez\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.21832v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-21832/"}
{"id": "arxiv:2608.21898", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d862afa969c0b4d1c6edc8d126017749844158f786b29c3b0710afecdec2008", "submitted_at": "2026-08-22T10:18:45Z", "updated_at": "2026-09-28T15:54:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21898v2"}, "authors": ["Chenghao Zhang", "Yuxi Cheng", "Saisai Hu", "Canran Xiao", "Yikai Mao", "Dan Roth"], "canonical_source_url": "https://arxiv.org/abs/2608.21898", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-21898/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7d862afa969c0b4d1c6edc8d126017749844158f786b29c3b0710afecdec2008", "submitted_at": "2026-08-22T10:18:45Z", "updated_at": "2026-09-28T15:54:31Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.21898v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.21898"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.21898"}], "name": "Training Needs Trustworthy Worlds: Verified Synthetic Web Environments for Agent Learning", "publication": null, "published_at": "2026-08-22", "record_sha256": "6f684b48357a56d300e402aa39beaadc75435589cf12ba3cbd5db1a09ae04875", "relations": [], "resource_id": "arxiv:2608.21898", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.21898v2", "topics": ["data-generation", "sandbox", "reinforcement-learning"], "updated_at": "2026-09-28"}, "text": "Training Needs Trustworthy Worlds: Verified Synthetic Web Environments for Agent Learning\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds synthetic websites with verified navigation, backend state, and feasible tasks to provide reliable supervision for web-agent training.\nPlatforms: web\nTopics: data-generation, sandbox, reinforcement-learning\nPublished: 2026-08-22\nAuthors: Chenghao Zhang, Yuxi Cheng, Saisai Hu, Canran Xiao, Yikai Mao, Dan Roth\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.21898v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-21898/"}
{"id": "arxiv:2608.22577", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ea33a82ab7543a64b60ee3024c6f3e39f88b4e3f06eff94f8c790c56cf76715", "submitted_at": "2026-08-23T20:04:21Z", "updated_at": "2026-08-25T04:24:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.22577v2"}, "authors": ["Jiaxuan Luo", "Zhanfeng Liao", "Jiayao Teng", "Yuan Wang"], "canonical_source_url": "https://arxiv.org/abs/2608.22577", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-22577/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2ea33a82ab7543a64b60ee3024c6f3e39f88b4e3f06eff94f8c790c56cf76715", "submitted_at": "2026-08-23T20:04:21Z", "updated_at": "2026-08-25T04:24:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.22577v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.22577"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.22577"}], "name": "CausalCache: Conditional High-Fidelity Restoration for Long-Horizon GUI Agents", "publication": null, "published_at": "2026-08-23", "record_sha256": "5b943f9cc1922bca6ab5f9148a34687e9c1b30538364043fcd58d71af84b0905", "relations": [], "resource_id": "arxiv:2608.22577", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.22577v2", "topics": ["memory", "long-horizon", "efficiency"], "updated_at": "2026-08-25"}, "text": "CausalCache: Conditional High-Fidelity Restoration for Long-Horizon GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CausalCache, which selects archived screenshots to restore alongside text summaries under a fixed visual-memory budget. A history-gated adapter processes restored images, with evaluation distinguishing standard step limits from longer-horizon and mobile-transfer settings.\nPlatforms: desktop\nTopics: memory, long-horizon, efficiency\nPublished: 2026-08-23\nAuthors: Jiaxuan Luo, Zhanfeng Liao, Jiayao Teng, Yuan Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.22577v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-22577/"}
{"id": "arxiv:2608.22847", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c1198ce65d5a2fcffc51b37bc4553b35b9e0d809757dc4181747c57c7ae443d", "submitted_at": "2026-08-24T06:27:47Z", "updated_at": "2026-08-24T06:27:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.22847v1"}, "authors": ["Long Zhang", "Yuhan Chen", "Chaoran Zhang", "Wanxia Cao", "Kun Huang", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Chenliang Li", "Lixin Zou"], "canonical_source_url": "https://arxiv.org/abs/2608.22847", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-22847/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4c1198ce65d5a2fcffc51b37bc4553b35b9e0d809757dc4181747c57c7ae443d", "submitted_at": "2026-08-24T06:27:47Z", "updated_at": "2026-08-24T06:27:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.22847v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.22847"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.22847"}], "name": "GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis", "publication": null, "published_at": "2026-08-24", "record_sha256": "f0eb38d3796eb33bb749bece2170183a100e4e15d77cadf47b91e6154425f6be", "relations": [], "resource_id": "arxiv:2608.22847", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.22847v1", "topics": ["data-generation", "reward-models", "reinforcement-learning"], "updated_at": "2026-08-24"}, "text": "GSAR: Goal-State-Anchor Rewards for Mobile GUI Agents with Self-Evolving Data Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GSAR, which synthesizes GUI tasks and successful goal states, then annotates task-relevant interface elements as reward anchors. These anchors support offline trajectory verification and online policy training.\nPlatforms: android\nTopics: data-generation, reward-models, reinforcement-learning\nPublished: 2026-08-24\nAuthors: Long Zhang, Yuhan Chen, Chaoran Zhang, Wanxia Cao, Kun Huang, Pengzhi Gao, Wei Liu, Jian Luan, Chenliang Li, Lixin Zou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.22847v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-22847/"}
{"id": "arxiv:2608.24015", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e4fe2cb74124b2897d86a4fee32113164b86a8f5abae040bf77db4900e0bc4", "submitted_at": "2026-08-25T03:10:01Z", "updated_at": "2026-08-25T03:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24015v1"}, "authors": ["Yijie Ma", "Chaoyue Niu", "Fan Wu", "Guihai Chen"], "canonical_source_url": "https://arxiv.org/abs/2608.24015", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24015/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a7e4fe2cb74124b2897d86a4fee32113164b86a8f5abae040bf77db4900e0bc4", "submitted_at": "2026-08-25T03:10:01Z", "updated_at": "2026-08-25T03:10:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24015v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.24015"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24015"}], "name": "Reflection with Action-Induced Visual Differences for Desktop GUI Agents", "publication": null, "published_at": "2026-08-25", "record_sha256": "6e0c986f7d0e8416dc17c1099ee4b1990ac5f7a3d45f1c8be7255ceea88fff97", "relations": [], "resource_id": "arxiv:2608.24015", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.24015v1", "topics": ["self-reflection", "screen-understanding", "robustness"], "updated_at": "2026-08-25"}, "text": "Reflection with Action-Induced Visual Differences for Desktop GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Evidence-First Reflection, which separates extraction of action-induced screen changes from judgment of whether an operation achieved its objective. It filters visual differences into relevant evidence before a desktop agent’s reflector decides how execution should continue.\nPlatforms: desktop, windows\nTopics: self-reflection, screen-understanding, robustness\nPublished: 2026-08-25\nAuthors: Yijie Ma, Chaoyue Niu, Fan Wu, Guihai Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24015v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24015/"}
{"id": "arxiv:2608.24017", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bddfe5b5b2312e9b71d01dbb6bb23a8b0ad1c8cd787e4e98449b71a2ac2d678", "submitted_at": "2026-08-25T03:16:45Z", "updated_at": "2026-08-25T03:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24017v1"}, "authors": ["Lin-Fa Lee", "YI-YU Chang", "Kuo-Hui Yeh"], "canonical_source_url": "https://arxiv.org/abs/2608.24017", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24017/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1bddfe5b5b2312e9b71d01dbb6bb23a8b0ad1c8cd787e4e98449b71a2ac2d678", "submitted_at": "2026-08-25T03:16:45Z", "updated_at": "2026-08-25T03:16:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24017v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.24017"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24017"}], "name": "WebMCP-Phalanx: Enforcing and Characterizing Trust Boundaries for Browser-Integrated LLM Agents", "publication": null, "published_at": "2026-08-25", "record_sha256": "4064a539cec1baec2fe66a26f17b9028164eacd8e7c0e65a94ab60fe1768fead", "relations": [], "resource_id": "arxiv:2608.24017", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2608.24017v1", "topics": ["safety-security", "tool-integration"], "updated_at": "2026-08-25"}, "text": "WebMCP-Phalanx: Enforcing and Characterizing Trust Boundaries for Browser-Integrated LLM Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nProposes WebMCP-Phalanx, a browser-integrated runtime that binds page-exposed tools to their registering principals and separates content inspection from privileged invocation. It studies lifecycle and prompt-injection attacks, including a timing bypass that motivates validating all visible metadata before tool use.\nPlatforms: web\nTopics: safety-security, tool-integration\nPublished: 2026-08-25\nAuthors: Lin-Fa Lee, YI-YU Chang, Kuo-Hui Yeh\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24017v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24017/"}
{"id": "arxiv:2608.24099", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51901009de5898c52a1b9d4958b698808a8f8b850d19afa32dd7a3f32d42a851", "submitted_at": "2026-08-25T05:57:25Z", "updated_at": "2026-08-25T05:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24099v1"}, "authors": ["Guo Gan", "Yilun Zhao", "Cong Chen", "Jinbiao Wei", "Tingyu Song", "Zheyuan Yang", "Lin Fu", "Hong Zhou"], "canonical_source_url": "https://arxiv.org/abs/2608.24099", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24099/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "51901009de5898c52a1b9d4958b698808a8f8b850d19afa32dd7a3f32d42a851", "submitted_at": "2026-08-25T05:57:25Z", "updated_at": "2026-08-25T05:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24099v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.24099"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24099"}], "name": "Are Android GUI Agents Robust Against Runtime Anomalies? AnTrap: Evaluating Agents in Dynamic Adversarial Environments", "publication": null, "published_at": "2026-08-25", "record_sha256": "99554ed3705c76ebf1e3b15dc6a6e572e91554128ed494a428e4116430b5f9da", "relations": [], "resource_id": "arxiv:2608.24099", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.24099v1", "topics": ["evaluation", "robustness", "reinforcement-learning"], "updated_at": "2026-08-25"}, "text": "Are Android GUI Agents Robust Against Runtime Anomalies? AnTrap: Evaluating Agents in Dynamic Adversarial Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AnTrap, which injects task-solvable runtime anomalies into Android GUI-agent trajectories using a layered failure taxonomy. It compares model robustness and adversarial reinforcement learning to distinguish recoverable interface traps from persistent contextual failures.\nPlatforms: android\nTopics: evaluation, robustness, reinforcement-learning\nPublished: 2026-08-25\nAuthors: Guo Gan, Yilun Zhao, Cong Chen, Jinbiao Wei, Tingyu Song, Zheyuan Yang, Lin Fu, Hong Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24099v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24099/"}
{"id": "arxiv:2608.24174", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76cd09f70bfec0d2a49cddae8adcf32ca37fb39bcade7c1d36d5f3dc3fa6bfa7", "submitted_at": "2026-08-25T07:40:12Z", "updated_at": "2026-08-25T07:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24174v1"}, "authors": ["Tao Xiong", "Xavier Hu", "Wenkai Wang", "Qinzhuo Wu", "Changqiao Wu", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Shengyu Zhang"], "canonical_source_url": "https://arxiv.org/abs/2608.24174", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24174/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "76cd09f70bfec0d2a49cddae8adcf32ca37fb39bcade7c1d36d5f3dc3fa6bfa7", "submitted_at": "2026-08-25T07:40:12Z", "updated_at": "2026-08-25T07:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24174v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.24174"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24174"}], "name": "Task-Adaptive Rubrics for GUI Reward Modeling", "publication": null, "published_at": "2026-08-25", "record_sha256": "ebcab7d0fdd9d271556bc8c572559b7d741749762aff4cf9f7ae41fdfc6ce1b5", "relations": [], "resource_id": "arxiv:2608.24174", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.24174v1", "topics": ["reward-models", "evaluation", "reinforcement-learning"], "updated_at": "2026-08-25"}, "text": "Task-Adaptive Rubrics for GUI Reward Modeling\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AdaptRubric, which retrieves reusable criteria for a GUI task family and then generates instruction-specific checks for concrete values and constraints. It evaluates these adaptive rubrics in offline reward assessment and online agent optimization.\nPlatforms: unspecified\nTopics: reward-models, evaluation, reinforcement-learning\nPublished: 2026-08-25\nAuthors: Tao Xiong, Xavier Hu, Wenkai Wang, Qinzhuo Wu, Changqiao Wu, Pengzhi Gao, Wei Liu, Jian Luan, Shengyu Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24174v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24174/"}
{"id": "arxiv:2608.24848", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d2180e9b1b0c3ae20e32b0b597dc209e363551aaa5031c98504ef08314979a6", "submitted_at": "2026-08-25T17:35:42Z", "updated_at": "2026-08-25T17:35:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24848v1"}, "authors": ["Fei Tang", "Huawen Shen", "Zhiqiong Lu", "Zhengxi Lu", "Pengyuan Lyu", "Chengquan Zhang", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2608.24848", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24848/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d2180e9b1b0c3ae20e32b0b597dc209e363551aaa5031c98504ef08314979a6", "submitted_at": "2026-08-25T17:35:42Z", "updated_at": "2026-08-25T17:35:42Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24848v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.24848"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24848"}], "name": "BrowserForge: Scaling Web Episode via Parallel Browser Sandboxes", "publication": null, "published_at": "2026-08-25", "record_sha256": "cb37414b9c4a604e728fd5446c8f4b45aa9418ce44e547272f881e44920d7e91", "relations": [], "resource_id": "arxiv:2608.24848", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.24848v1", "topics": ["data-generation", "sandbox", "web-navigation"], "updated_at": "2026-08-25"}, "text": "BrowserForge: Scaling Web Episode via Parallel Browser Sandboxes\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces BrowserForge, which combines broad website sourcing, parallel browser sandboxes, and task-proposer/solver interaction to collect verified web trajectories. The pipeline uses page structure during synthesis but trains a policy that acts from screenshots.\nPlatforms: web\nTopics: data-generation, sandbox, web-navigation\nPublished: 2026-08-25\nAuthors: Fei Tang, Huawen Shen, Zhiqiong Lu, Zhengxi Lu, Pengyuan Lyu, Chengquan Zhang, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24848v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24848/"}
{"id": "arxiv:2608.24913", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "046d93a43dd5a011809a780c7d998f98ee0abb3e931d733b8ad7807b86050ec7", "submitted_at": "2026-07-26T17:21:24Z", "updated_at": "2026-07-26T17:21:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24913v1"}, "authors": ["Lily Bundgaard Wanscher", "Markus Heidemann Lorensen", "Mohammed Ammad Shafiq", "Mahyar Tourchi Moghaddam", "Mina Alipour"], "canonical_source_url": "https://arxiv.org/abs/2608.24913", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-24913/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "046d93a43dd5a011809a780c7d998f98ee0abb3e931d733b8ad7807b86050ec7", "submitted_at": "2026-07-26T17:21:24Z", "updated_at": "2026-07-26T17:21:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.24913v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}], "identifiers": {"arxiv": "2608.24913"}, "kind": "paper", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.24913"}], "name": "From Blind Edits to Verified Repair: Building Trustworthy User-Side LLM Agents for Web Accessibility", "publication": null, "published_at": "2026-07-26", "record_sha256": "866fd271a81f714dc0e334c5dc0673544a2c85ee41db6269e16ab4c68b0f29e1", "relations": [], "resource_id": "arxiv:2608.24913", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.24913v1", "topics": ["ui-design", "web-navigation", "safety-security"], "updated_at": "2026-07-26"}, "text": "From Blind Edits to Verified Repair: Building Trustworthy User-Side LLM Agents for Web Accessibility\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds a local browser-extension agent that generates and verifies additive CSS repairs for WCAG web accessibility violations.\nPlatforms: web\nTopics: ui-design, web-navigation, safety-security\nPublished: 2026-07-26\nAuthors: Lily Bundgaard Wanscher, Markus Heidemann Lorensen, Mohammed Ammad Shafiq, Mahyar Tourchi Moghaddam, Mina Alipour\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.24913v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-24913/"}
{"id": "arxiv:2608.25777", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af6b60c6a39a990e950cf4b98f905f2d3aa5ae62d4055630c25a650508ec8bf8", "submitted_at": "2026-08-26T13:20:27Z", "updated_at": "2026-09-25T23:54:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.25777v2"}, "authors": ["Weiming Li", "Helen Paik", "Yulei Sui"], "canonical_source_url": "https://arxiv.org/abs/2608.25777", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-25777/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "af6b60c6a39a990e950cf4b98f905f2d3aa5ae62d4055630c25a650508ec8bf8", "submitted_at": "2026-08-26T13:20:27Z", "updated_at": "2026-09-25T23:54:04Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.25777v2"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.25777"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.25777"}], "name": "LocalLSTC: A Long Short-Term Control Architecture for Locally Deployed GUI Agents", "publication": null, "published_at": "2026-08-26", "record_sha256": "37772c97a8a3fadd57cfead8e49d8534792e1095bc0742675386fc99218fa416", "relations": [], "resource_id": "arxiv:2608.25777", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.25777v2", "topics": ["state-tracking", "planning", "training-free"], "updated_at": "2026-09-25"}, "text": "LocalLSTC: A Long Short-Term Control Architecture for Locally Deployed GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces LocalLSTC, a training-free control architecture that maintains persistent task state and uses it to form short execution commitments. Outcomes feed back into that state for progress checks, recovery, and termination in locally deployed desktop agents.\nPlatforms: desktop, windows\nTopics: state-tracking, planning, training-free\nPublished: 2026-08-26\nAuthors: Weiming Li, Helen Paik, Yulei Sui\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.25777v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-25777/"}
{"id": "arxiv:2608.26204", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50c26a7ab741ecb3cb855305ee4f7d5b5729296248d650c469600aeb73101781", "submitted_at": "2026-08-25T23:21:57Z", "updated_at": "2026-09-15T04:43:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.26204v2"}, "authors": ["Joy Chen", "Alejandro Castillejo Munoz", "Pierluca D'Oro", "Yuxuan Sun", "Chloe Evans", "Joseph Tighe"], "canonical_source_url": "https://arxiv.org/abs/2608.26204", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-26204/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50c26a7ab741ecb3cb855305ee4f7d5b5729296248d650c469600aeb73101781", "submitted_at": "2026-08-25T23:21:57Z", "updated_at": "2026-09-15T04:43:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.26204v2"}}, "environments": ["desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.26204"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.26204"}], "name": "ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices", "publication": null, "published_at": "2026-08-25", "record_sha256": "b11793c95a12c179de6de2d6840e55e9cec46a749b8b3ae0ffd01aadfd5bba74", "relations": [], "resource_id": "arxiv:2608.26204", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.26204v2", "topics": ["safety-security", "evaluation", "human-in-the-loop"], "updated_at": "2026-09-15"}, "text": "ADeptS-Bench: Measuring the Trustworthiness of Computer Use Agents Across Devices\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates interface-embedded threats and whether computer-use agents seek clarification when mobile or desktop instructions are ambiguous.\nPlatforms: desktop, cross-platform\nTopics: safety-security, evaluation, human-in-the-loop\nPublished: 2026-08-25\nAuthors: Joy Chen, Alejandro Castillejo Munoz, Pierluca D'Oro, Yuxuan Sun, Chloe Evans, Joseph Tighe\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.26204v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-26204/"}
{"id": "arxiv:2608.26991", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1626d5190f145c668c9b7f363b7f84904c12daae485e63b858e440ded5ea6df", "submitted_at": "2026-08-27T11:38:19Z", "updated_at": "2026-08-27T11:38:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.26991v1"}, "authors": ["Rui Xie", "Lu Chen"], "canonical_source_url": "https://arxiv.org/abs/2608.26991", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-26991/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e1626d5190f145c668c9b7f363b7f84904c12daae485e63b858e440ded5ea6df", "submitted_at": "2026-08-27T11:38:19Z", "updated_at": "2026-08-27T11:38:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.26991v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:31:35Z", "role": "primary", "sha256": "6499b4f4917eec0765d16d6e26a08d7edc5146a30753b19cf8218517656bb070", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2608.16211%2C2608.16391%2C2608.17067%2C2608.17975%2C2608.19741%2C2608.21500%2C2608.22301%2C2608.22510%2C2608.23035%2C2608.23320%2C2608.26238%2C2608.26546%2C2608.26991%2C2608.27260%2C2608.29814%2C2608.31076%2C2609.02292%2C2609.04173%2C2609.04298%2C2609.05776%2C2609.06059%2C2609.07529%2C2609.08149%2C2609.08418%2C2609.09187%2C2609.11561%2C2609.12533%2C2609.15096%2C2609.15544%2C2609.20732%2C2609.21293%2C2609.23986%2C2609.24971%2C2609.24997%2C2609.27455%2C2609.29647%2C2609.29964%2C2609.30798%2C2609.32835%2C2609.34428%2C2609.35002%2C2609.35232%2C2609.35394%2C2609.35581%2C2609.36071%2C2609.36352%2C2609.36777%2C2609.37017%2C2609.38923%2C2609.39222&max_results=50"}], "identifiers": {"arxiv": "2608.26991"}, "kind": "paper", "last_verified_at": "2026-10-17T12:31:35Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.26991"}], "name": "ASIL: Replacing Screenshot-and-Click with Structured State and Semantic Actions", "publication": null, "published_at": "2026-08-27", "record_sha256": "688773b718f69952a869012adc67ee8aec6c7b15f8c44f399e7918b4d3bcbfa1", "relations": [], "resource_id": "arxiv:2608.26991", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.26991v1", "topics": ["automation", "tool-integration", "evaluation", "reinforcement-learning", "imitation-learning"], "updated_at": "2026-08-27"}, "text": "ASIL: Replacing Screenshot-and-Click with Structured State and Semantic Actions\nResource type: paper\nReview status: curated\nScope: core\n\nASIL exposes application state as structured JSON and provides code-executable semantic actions. Its application benchmark compares this interface with screenshot control and application-native APIs, and studies supervised and reinforcement learning on the structured interface.\nPlatforms: unspecified\nTopics: automation, tool-integration, evaluation, reinforcement-learning, imitation-learning\nPublished: 2026-08-27\nAuthors: Rui Xie, Lu Chen\nRecorded arXiv version: v1\n\nReviewed against the complete retained official abstract. Experimental findings and release availability have not been independently reproduced or verified.\n\nPrimary source: https://arxiv.org/abs/2608.26991v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-26991/"}
{"id": "arxiv:2608.27508", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "576702fae164cbcb4adc86a81115487aae7e1c40cf07a1ecf3ef7c4c7e822c3b", "submitted_at": "2026-08-27T07:09:28Z", "updated_at": "2026-08-27T07:09:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27508v1"}, "authors": ["Yu Han", "Tianwen Qian"], "canonical_source_url": "https://arxiv.org/abs/2608.27508", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-27508/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "576702fae164cbcb4adc86a81115487aae7e1c40cf07a1ecf3ef7c4c7e822c3b", "submitted_at": "2026-08-27T07:09:28Z", "updated_at": "2026-08-27T07:09:28Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27508v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.27508"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27508"}], "name": "WM-R1: Training GUI Agents to Reason and leverage World Models with Reinforcement Learning", "publication": null, "published_at": "2026-08-27", "record_sha256": "d3c38ca31a729386b2b910e7a02abe550b2271e22ffb6c1a639762d206ef6d7a", "relations": [], "resource_id": "arxiv:2608.27508", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.27508v1", "topics": ["world-models", "reinforcement-learning"], "updated_at": "2026-08-27"}, "text": "WM-R1: Training GUI Agents to Reason and leverage World Models with Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nUses learned world models for mobile-agent training rollouts and for reasoning about the consequences of candidate GUI actions.\nPlatforms: android\nTopics: world-models, reinforcement-learning\nPublished: 2026-08-27\nAuthors: Yu Han, Tianwen Qian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.27508v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-27508/"}
{"id": "arxiv:2608.27808", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14864aae8f977d474ff3e7dd3e38951374862abfb8077395f2dc5a9e1cc5a489", "submitted_at": "2026-08-28T00:50:47Z", "updated_at": "2026-08-28T00:50:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27808v1"}, "authors": ["Divake Kumar", "Sina Tayebati", "Devashri Naik", "Amanda Sofie Rios", "Nilesh Ahuja", "Omesh Tickoo", "Ranganath Krishnan", "Amit Ranjan Trivedi"], "canonical_source_url": "https://arxiv.org/abs/2608.27808", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-27808/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14864aae8f977d474ff3e7dd3e38951374862abfb8077395f2dc5a9e1cc5a489", "submitted_at": "2026-08-28T00:50:47Z", "updated_at": "2026-08-28T00:50:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27808v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.27808"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27808"}], "name": "CURA: Certified Runtime Alarms for Computer-Use Agents", "publication": null, "published_at": "2026-08-28", "record_sha256": "d05d98d3cfc536e75c90f11fe37f3e22435b6bd3d8766ea748c4f33b2d47eae6", "relations": [], "resource_id": "arxiv:2608.27808", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.27808v1", "topics": ["observability", "safety-security", "evaluation"], "updated_at": "2026-08-28"}, "text": "CURA: Certified Runtime Alarms for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CURA, a telemetry-based runtime monitor that turns computer-use trajectories into sequential failure alarms and can trigger mid-task oversight. Its certificate controls false alarms, while the study separately reports detection performance and limits of retrospective monitoring.\nPlatforms: desktop\nTopics: observability, safety-security, evaluation\nPublished: 2026-08-28\nAuthors: Divake Kumar, Sina Tayebati, Devashri Naik, Amanda Sofie Rios, Nilesh Ahuja, Omesh Tickoo, Ranganath Krishnan, Amit Ranjan Trivedi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.27808v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-27808/"}
{"id": "arxiv:2608.27866", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352699b04f3a9406d7cab95ce9b5fb000e3e98a28996777e75b7399a6290357a", "submitted_at": "2026-08-28T03:15:11Z", "updated_at": "2026-08-28T03:15:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27866v1"}, "authors": ["Jiahe Ying", "Wendong Bu", "Kaihang Pan", "Bingchen Miao", "Siyu Chen", "Wen Wang", "Xueming Jiang", "Juncheng Li", "Siliang Tang"], "canonical_source_url": "https://arxiv.org/abs/2608.27866", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-27866/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "352699b04f3a9406d7cab95ce9b5fb000e3e98a28996777e75b7399a6290357a", "submitted_at": "2026-08-28T03:15:11Z", "updated_at": "2026-08-28T03:15:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.27866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.27866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.27866"}], "name": "Iron: Intent-Aligned and Retrospective Dual Learning Framework for Enhancing Generalist Virtual Agents", "publication": null, "published_at": "2026-08-28", "record_sha256": "cb0ee67028c3e0720dd2f86a7b23e5902dd2ce3c55dfeb72fb11918d1f080b88", "relations": [], "resource_id": "arxiv:2608.27866", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.27866v1", "topics": ["reinforcement-learning", "data-generation", "grounding"], "updated_at": "2026-08-28"}, "text": "Iron: Intent-Aligned and Retrospective Dual Learning Framework for Enhancing Generalist Virtual Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Iron, a GUI-agent training framework that aligns low-level actions with high-level intent using cycle-consistency rewards. It also repurposes failed trajectories through hindsight reconstruction and evaluates transfer to other digital tasks.\nPlatforms: web\nTopics: reinforcement-learning, data-generation, grounding\nPublished: 2026-08-28\nAuthors: Jiahe Ying, Wendong Bu, Kaihang Pan, Bingchen Miao, Siyu Chen, Wen Wang, Xueming Jiang, Juncheng Li, Siliang Tang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.27866v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-27866/"}
{"id": "arxiv:2608.29305", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbf657b2228bad25e668e3c9d7ec20c0ecab403d34b7e0b5608d350c46b8e50f", "submitted_at": "2026-08-29T14:45:09Z", "updated_at": "2026-08-29T14:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.29305v1"}, "authors": ["Junxuan Li", "Zijun Liu", "Ziyi Huang", "Peng Li", "Yuzhou Liu", "Ming Yan", "Yang Liu"], "canonical_source_url": "https://arxiv.org/abs/2608.29305", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-29305/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bbf657b2228bad25e668e3c9d7ec20c0ecab403d34b7e0b5608d350c46b8e50f", "submitted_at": "2026-08-29T14:45:09Z", "updated_at": "2026-08-29T14:45:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.29305v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.29305"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.29305"}], "name": "Learning Simple Test-Time Environments for LLM Web Agents", "publication": null, "published_at": "2026-08-29", "record_sha256": "ce8ce2fb6c963532e5b6c4e49693bd896b8151d94be24ed9989414a23f02d15d", "relations": [], "resource_id": "arxiv:2608.29305", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.29305v1", "topics": ["planning", "robustness", "web-navigation"], "updated_at": "2026-08-29"}, "text": "Learning Simple Test-Time Environments for LLM Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Test-Time Environment Decomposition, which uses trial interaction to break complex web observations into simpler modules. A label-free adaptation procedure learns from these sub-environments and studies whether their gains compose in full web tasks.\nPlatforms: web\nTopics: planning, robustness, web-navigation\nPublished: 2026-08-29\nAuthors: Junxuan Li, Zijun Liu, Ziyi Huang, Peng Li, Yuzhou Liu, Ming Yan, Yang Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.29305v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-29305/"}
{"id": "arxiv:2608.29596", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13b6833243a543eb156ce7f4d943f8a3c608b3b38b7eaa77605ca7484cdedc43", "submitted_at": "2026-08-30T06:36:42Z", "updated_at": "2026-09-28T17:25:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.29596v2"}, "authors": ["Sanket Badhe", "Deep Shah", "Priyanka Tiwari", "Nehal Kathrotia"], "canonical_source_url": "https://arxiv.org/abs/2608.29596", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-29596/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "13b6833243a543eb156ce7f4d943f8a3c608b3b38b7eaa77605ca7484cdedc43", "submitted_at": "2026-08-30T06:36:42Z", "updated_at": "2026-09-28T17:25:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.29596v2"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:54:12Z", "role": "primary", "sha256": "5f5bae19420f79231516d4bcf604ea18d4f930fc7790cde5a3e7fcb8d41a1126", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2508.19461%2C2508.19493%2C2509.21782%2C2509.25140%2C2510.03485%2C2510.07809%2C2510.14276%2C2510.27140%2C2510.27210%2C2511.13248%2C2512.08417%2C2601.07263%2C2602.13516%2C2603.00476%2C2603.11862%2C2603.18762%2C2603.23791%2C2604.00986%2C2604.16515%2C2604.16548%2C2604.24348%2C2605.05509%2C2605.08876%2C2605.20833%2C2605.24785%2C2607.00333%2C2607.18659%2C2607.25255%2C2608.00202%2C2608.08939%2C2608.24913%2C2608.29596&max_results=32"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:00:09.204016+00:00", "role": "primary", "sha256": "d112d0a83472b87de538d0290efdd2d0dd0d21e878acca5f383da5baa880fe88", "supports": ["scope", "environments", "notes"], "url": "https://arxiv.org/html/2608.29596v2"}], "identifiers": {"arxiv": "2608.29596"}, "kind": "survey", "last_verified_at": "2026-10-18T10:54:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.29596"}], "name": "A Systematic Survey of Agentic Skills: Architecture, Lifecycle, and Security", "publication": null, "published_at": "2026-08-30", "record_sha256": "e44e21eefe0a95ab5b75fff9ddade5b3c1dd872266ca1dce54663c848f1c52ae", "relations": [], "resource_id": "arxiv:2608.29596", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2608.29596v2", "topics": ["survey", "workflow", "memory", "safety-security"], "updated_at": "2026-09-28"}, "text": "A Systematic Survey of Agentic Skills: Architecture, Lifecycle, and Security\nResource type: survey\nReview status: curated\nScope: adjacent\n\nSurveys modular agentic skill architectures, skill lifecycles, and security risks in autonomous tool- and interface-using agents.\nPlatforms: cross-platform\nTopics: survey, workflow, memory, safety-security\nPublished: 2026-08-30\nAuthors: Sanket Badhe, Deep Shah, Priyanka Tiwari, Nehal Kathrotia\nRecorded arXiv version: v2\n\nAdjacent broad survey of procedural agent skills. Full-text Section 4.4.2 explicitly covers desktop, mobile, and web interface skills; surveyed systems and their reported results are not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.29596v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-29596/"}
{"id": "arxiv:2608.30038", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b17f858c9d6992f6a8d8f21cc79d468b0f1c81fc44084045791d9c3acacd6d0", "submitted_at": "2026-08-30T20:55:21Z", "updated_at": "2026-08-30T20:55:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.30038v1"}, "authors": ["Mingshuo Wang", "Hanqing Guo", "Huining Li", "Yuliang Fu", "Jing Xu", "Chenhan Xu"], "canonical_source_url": "https://arxiv.org/abs/2608.30038", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-30038/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7b17f858c9d6992f6a8d8f21cc79d468b0f1c81fc44084045791d9c3acacd6d0", "submitted_at": "2026-08-30T20:55:21Z", "updated_at": "2026-08-30T20:55:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2608.30038v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2608.30038"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.30038"}], "name": "ActReal: System-Level Mobile Agents Challenge Mobile Automation Detection", "publication": null, "published_at": "2026-08-30", "record_sha256": "783db91fefc824eb3e98f57d62da47846d12ef6af1a1ef190d72e7c4680f0cdc", "relations": [], "resource_id": "arxiv:2608.30038", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.30038v1", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-08-30"}, "text": "ActReal: System-Level Mobile Agents Challenge Mobile Automation Detection\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ActReal, which translates semantic mobile-agent actions into coordinated touch and sensor signals to study automation-detection weaknesses. It evaluates detectors that observe touch events alone or together with inertial measurements under a privileged executor threat model.\nPlatforms: unspecified\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2026-08-30\nAuthors: Mingshuo Wang, Hanqing Guo, Huining Li, Yuliang Fu, Jing Xu, Chenhan Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.30038v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-30038/"}
{"id": "arxiv:2608.30207", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87c3f58eebe2db437b1166ff8e83b336db4cde37684a7cd86f61c4b34fca7237", "submitted_at": "2026-08-31T03:39:43Z", "updated_at": "2026-10-03T10:41:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.30207v2"}, "authors": ["Chen Xiong", "Zhiyuan He", "Pin-Yu Chen", "Stjepan Picek", "Tsung-Yi Ho"], "canonical_source_url": "https://arxiv.org/abs/2608.30207", "citation_url": "https://awesomecua.github.io/resources/arxiv-2608-30207/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "87c3f58eebe2db437b1166ff8e83b336db4cde37684a7cd86f61c4b34fca7237", "submitted_at": "2026-08-31T03:39:43Z", "updated_at": "2026-10-03T10:41:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2608.30207v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2608.30207"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2608.30207"}], "name": "SIR: Self-improving Red-teaming for Compute Use Agents", "publication": null, "published_at": "2026-08-31", "record_sha256": "5a876cf180ea9746d03de8acf7342093d56dd3640430286a9b964c9b17d773fa", "relations": [], "resource_id": "arxiv:2608.30207", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2608.30207v2", "topics": ["safety-security", "evaluation", "data-generation"], "updated_at": "2026-10-03"}, "text": "SIR: Self-improving Red-teaming for Compute Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SIR, which iteratively refines a reusable library of prompt-injection strategies from failed computer-use attack trajectories. Its evaluation uses deterministic system-state checks and counts attacks only when the adversarial objective and legitimate task both complete.\nPlatforms: desktop\nTopics: safety-security, evaluation, data-generation\nPublished: 2026-08-31\nAuthors: Chen Xiong, Zhiyuan He, Pin-Yu Chen, Stjepan Picek, Tsung-Yi Ho\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2608.30207v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2608-30207/"}
{"id": "arxiv:2609.00028", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "34383fa6c6e5cca059c53fd0452d93eec9ac6a96271ddc14f81ddb96fdeaceec", "submitted_at": "2026-08-27T17:15:33Z", "updated_at": "2026-08-27T17:15:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00028v1"}, "authors": ["Venus Team", "Zhuohan Cai", "Haoxing Chen", "Jiaxuan Chen", "Weizhi Chen", "Changlong Gao", "Zhangxuan Gu", "Yuan Guo", "Yusong Hu", "Jianrong Jiang", "Jianguo Li", "Runze Li", "Jinzhen Lin", "Zhenyu Ma", "Changhua Meng", "Han Peng", "Xinyu Qiu", "Shuheng Shen", "Zhongyi Shui", "Weiqiang Wang", "Ming Wen", "Zhuoer Xu", "Hang Yan", "Kaiwen Yang", "Ruilin Yao", "Nanjun Yu", "Zhengwen Zeng", "Lianrui Zhang", "Yunzhu Zhang", "Zhe Zhao", "Beitong Zhou"], "canonical_source_url": "https://arxiv.org/abs/2609.00028", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-00028/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "34383fa6c6e5cca059c53fd0452d93eec9ac6a96271ddc14f81ddb96fdeaceec", "submitted_at": "2026-08-27T17:15:33Z", "updated_at": "2026-08-27T17:15:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00028v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.00028"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00028"}], "name": "UI-Venus-2 Technical Report", "publication": null, "published_at": "2026-08-27", "record_sha256": "1333281c61baa5719f9655a01cd6e59a78109faf34db87efdc5b333cc01a3179", "relations": [], "resource_id": "arxiv:2609.00028", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.00028v1", "topics": ["automation", "reinforcement-learning", "safety-security"], "updated_at": "2026-08-27"}, "text": "UI-Venus-2 Technical Report\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a GUI foundation agent trained through expanded environments, function-grounded tasks, and trajectory verification with safety-aware execution mechanisms.\nPlatforms: web, desktop, cross-platform\nTopics: automation, reinforcement-learning, safety-security\nPublished: 2026-08-27\nAuthors: Venus Team, Zhuohan Cai, Haoxing Chen, Jiaxuan Chen, Weizhi Chen, Changlong Gao, Zhangxuan Gu, Yuan Guo, Yusong Hu, Jianrong Jiang, Jianguo Li, Runze Li, Jinzhen Lin, Zhenyu Ma, Changhua Meng, Han Peng, Xinyu Qiu, Shuheng Shen, Zhongyi Shui, Weiqiang Wang, Ming Wen, Zhuoer Xu, Hang Yan, Kaiwen Yang, Ruilin Yao, Nanjun Yu, Zhengwen Zeng, Lianrui Zhang, Yunzhu Zhang, Zhe Zhao, Beitong Zhou\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.00028v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-00028/"}
{"id": "arxiv:2609.00048", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dba8eb3f30a7e88ad973080d07fbf272f63400f4f8fa858d5f2657735ce093c0", "submitted_at": "2026-08-30T05:19:43Z", "updated_at": "2026-08-30T05:19:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00048v1"}, "authors": ["Lin Fu", "Zheyuan Yang", "Tianhui Zhang", "Jinbiao Wei", "Guo Gan", "Boxu Liu", "Yilun Zhao", "Yu Rong"], "canonical_source_url": "https://arxiv.org/abs/2609.00048", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-00048/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "dba8eb3f30a7e88ad973080d07fbf272f63400f4f8fa858d5f2657735ce093c0", "submitted_at": "2026-08-30T05:19:43Z", "updated_at": "2026-08-30T05:19:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00048v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.00048"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00048"}], "name": "GUI-CC: Benchmarking Contextual Consistency of GUI World Models as Agent Environments", "publication": null, "published_at": "2026-08-30", "record_sha256": "557774ef60993c0a5042a5301b6fdfa27d444bf37b1ba88d241e33bb19a9edb0", "relations": [], "resource_id": "arxiv:2609.00048", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.00048v1", "topics": ["world-models", "evaluation", "long-horizon"], "updated_at": "2026-08-30"}, "text": "GUI-CC: Benchmarking Contextual Consistency of GUI World Models as Agent Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUI-CC, a benchmark for assessing GUI world models as multi-step agent environments. Separate reference-action and agent-loop tracks test whether generated screens preserve context, plausible transitions, and task progress throughout interaction.\nPlatforms: unspecified\nTopics: world-models, evaluation, long-horizon\nPublished: 2026-08-30\nAuthors: Lin Fu, Zheyuan Yang, Tianhui Zhang, Jinbiao Wei, Guo Gan, Boxu Liu, Yilun Zhao, Yu Rong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.00048v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-00048/"}
{"id": "arxiv:2609.00524", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "deb130d93bbbd6dfdc04c72d232a9dcfa1e92b0586831e5864070e7fe0de14dd", "submitted_at": "2026-09-01T00:44:19Z", "updated_at": "2026-09-01T00:44:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00524v1"}, "authors": ["Satwik Ram Kodandaram", "Monalika Padma Reddy", "Xiaojun Bi", "Jiawei Zhou", "I. V. Ramakrishnan", "Vikas Ashok"], "canonical_source_url": "https://arxiv.org/abs/2609.00524", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-00524/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "deb130d93bbbd6dfdc04c72d232a9dcfa1e92b0586831e5864070e7fe0de14dd", "submitted_at": "2026-09-01T00:44:19Z", "updated_at": "2026-09-01T00:44:19Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.00524v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.00524"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.00524"}], "name": "Are We There Yet? Assessing Computer-Use Agents for Blind Users' Accessible Interaction with Desktop Applications", "publication": null, "published_at": "2026-09-01", "record_sha256": "ba6d7995b6c0e9bd538fd37cb7b4f91f5e8e7136e3283dbf18da6e790f9d935b", "relations": [], "resource_id": "arxiv:2609.00524", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.00524v1", "topics": ["evaluation", "human-in-the-loop", "grounding"], "updated_at": "2026-09-01"}, "text": "Are We There Yet? Assessing Computer-Use Agents for Blind Users' Accessible Interaction with Desktop Applications\nResource type: paper\nReview status: curated\nScope: core\n\nA diary study deploys the screen-reader-accessible OLLA computer-use prototype with blind users across desktop applications. Recorded commands and interaction traces support model comparisons and analysis of grounding, planning, constraint tracking and termination failures.\nPlatforms: desktop\nTopics: evaluation, human-in-the-loop, grounding\nPublished: 2026-09-01\nAuthors: Satwik Ram Kodandaram, Monalika Padma Reddy, Xiaojun Bi, Jiawei Zhou, I. V. Ramakrishnan, Vikas Ashok\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.00524v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-00524/"}
{"id": "arxiv:2609.02057", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "967f1cdda69349da3280827c0b90301e4942c44398902ab6cfb868f1d3e94a84", "submitted_at": "2026-09-02T03:34:47Z", "updated_at": "2026-09-02T03:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02057v1"}, "authors": ["Sitong Pan", "Yipeng Shen", "Yilin Lu", "Caiwen Ding", "Lu Cheng", "Qianwen Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.02057", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-02057/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "967f1cdda69349da3280827c0b90301e4942c44398902ab6cfb868f1d3e94a84", "submitted_at": "2026-09-02T03:34:47Z", "updated_at": "2026-09-02T03:34:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02057v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.02057"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02057"}], "name": "Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision", "publication": null, "published_at": "2026-09-02", "record_sha256": "cb923aa3c020257626d03275a2b71c1e0a30c4b1ca872ee46ad49e603b65015d", "relations": [], "resource_id": "arxiv:2609.02057", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.02057v1", "topics": ["observability", "evaluation", "robustness", "interaction-history"], "updated_at": "2026-09-02"}, "text": "Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision\nResource type: paper\nReview status: curated\nScope: core\n\nPredicts web-agent failure risk from observable trajectory prefixes using cross-step behavior and black-box consistency queries. Key-step labels preserve valid early prefixes of failed runs, and evaluation tests early intervention and transfer across website categories.\nPlatforms: web\nTopics: observability, evaluation, robustness, interaction-history\nPublished: 2026-09-02\nAuthors: Sitong Pan, Yipeng Shen, Yilin Lu, Caiwen Ding, Lu Cheng, Qianwen Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.02057v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-02057/"}
{"id": "arxiv:2609.02149", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52849171f507e126093411ba81dc580d460b84f831358b08bd57cb54ca41b665", "submitted_at": "2026-09-02T06:03:18Z", "updated_at": "2026-09-10T15:52:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.02149v3"}, "authors": ["Yixiong Xiao", "Lang An", "Hucheng Yang", "Pinxue Ma", "Yongquan Chen", "Jingjia Cao", "Yusai Zhao", "Ting Wang", "Ting Liu", "Siqi Bao", "Jingbo Zhou", "Hua Wu"], "canonical_source_url": "https://arxiv.org/abs/2609.02149", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-02149/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "52849171f507e126093411ba81dc580d460b84f831358b08bd57cb54ca41b665", "submitted_at": "2026-09-02T06:03:18Z", "updated_at": "2026-09-10T15:52:42Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.02149v3"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.02149"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02149"}], "name": "OmegaUse-SOP: SOP Engineering for Professional Computer Use from Human Demonstrations", "publication": null, "published_at": "2026-09-02", "record_sha256": "b9a30d1550362ba6b80b46cacc03b4a23f50cbe94d35b99f1f90c703a707c154", "relations": [], "resource_id": "arxiv:2609.02149", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.02149v3", "topics": ["human-in-the-loop", "recording", "workflow", "automation"], "updated_at": "2026-09-10"}, "text": "OmegaUse-SOP: SOP Engineering for Professional Computer Use from Human Demonstrations\nResource type: paper\nReview status: curated\nScope: core\n\nOmegaUse-SOP converts expert GUI demonstrations into configurable procedures and refines them with domain rules and human input. It records, abstracts, grounds and verifies live execution, with a case study of photovoltaic simulation workflows in PVsyst 7.2.\nPlatforms: unspecified\nTopics: human-in-the-loop, recording, workflow, automation\nPublished: 2026-09-02\nAuthors: Yixiong Xiao, Lang An, Hucheng Yang, Pinxue Ma, Yongquan Chen, Jingjia Cao, Yusai Zhao, Ting Wang, Ting Liu, Siqi Bao, Jingbo Zhou, Hua Wu\nRecorded arXiv version: v3\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.02149v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-02149/"}
{"id": "arxiv:2609.02154", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b462dddd45e6b04a3bf6fa0bc73576a23a5f9617f61aeb257a845c0f456f731d", "submitted_at": "2026-09-02T06:06:43Z", "updated_at": "2026-09-02T06:06:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02154v1"}, "authors": ["Yimeng Liu", "Hua Huang"], "canonical_source_url": "https://arxiv.org/abs/2609.02154", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-02154/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b462dddd45e6b04a3bf6fa0bc73576a23a5f9617f61aeb257a845c0f456f731d", "submitted_at": "2026-09-02T06:06:43Z", "updated_at": "2026-09-02T06:06:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02154v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.02154"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02154"}], "name": "WiP: Characterizing and Defending Against Mobile-Agent-Driven MFA Automation", "publication": null, "published_at": "2026-09-02", "record_sha256": "ea1c1026b408dd52e55a13e347856b3a66530090d78c281a889a0d93cba198c2", "relations": [], "resource_id": "arxiv:2609.02154", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.02154v1", "topics": ["safety-security", "evaluation"], "updated_at": "2026-09-02"}, "text": "WiP: Characterizing and Defending Against Mobile-Agent-Driven MFA Automation\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how mobile automation can combine authentication factors within one agent environment and evaluates motion sensing as a user-presence signal.\nPlatforms: android\nTopics: safety-security, evaluation\nPublished: 2026-09-02\nAuthors: Yimeng Liu, Hua Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.02154v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-02154/"}
{"id": "arxiv:2609.02309", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec439fd2cfdf7319d77ef5ab58e50143bceb3f8205ccffa0d83555f10dece42a", "submitted_at": "2026-09-02T08:54:58Z", "updated_at": "2026-09-02T08:54:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02309v1"}, "authors": ["Bizhe Bai", "Jiakang Yuan", "Hongming Wu", "Xinyue Wang", "Jie Ren", "Siyao Chen", "Yuchen Ya", "Fan Bai", "Pai Peng", "Huafeng Qin", "Tao Chen"], "canonical_source_url": "https://arxiv.org/abs/2609.02309", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-02309/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ec439fd2cfdf7319d77ef5ab58e50143bceb3f8205ccffa0d83555f10dece42a", "submitted_at": "2026-09-02T08:54:58Z", "updated_at": "2026-09-02T08:54:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02309v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.02309"}, "kind": "survey", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02309"}], "name": "Efficient GUI Agents: A Systems Survey of Observation, Memory, Action, and Runtime Optimization", "publication": null, "published_at": "2026-09-02", "record_sha256": "adc44dfe425f8d31332b77465a748bc793d28076450fe15d2c5b0e8d42f86ee3", "relations": [], "resource_id": "arxiv:2609.02309", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.02309v1", "topics": ["survey"], "updated_at": "2026-09-02"}, "text": "Efficient GUI Agents: A Systems Survey of Observation, Memory, Action, and Runtime Optimization\nResource type: survey\nReview status: curated\nScope: core\n\nSurveys GUI-agent efficiency beyond task success: observation, context and memory, action, and planner or runtime cost.\nPlatforms: cross-platform\nTopics: survey\nPublished: 2026-09-02\nAuthors: Bizhe Bai, Jiakang Yuan, Hongming Wu, Xinyue Wang, Jie Ren, Siyao Chen, Yuchen Ya, Fan Bai, Pai Peng, Huafeng Qin, Tao Chen\nRecorded arXiv version: v1\n\nPrimary source: https://arxiv.org/abs/2609.02309v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-02309/"}
{"id": "arxiv:2609.02885", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8703474cb4f577b307731cf1a986f76f800470fde625b6ca4b3569d30e74f903", "submitted_at": "2026-09-02T17:59:40Z", "updated_at": "2026-09-02T17:59:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02885v1"}, "authors": ["Kelvin Li", "Dhruv Pendharkar", "Anish Pahilajani", "Chuyi Shang", "Leon Oks", "Leonid Karlinsky", "Rogerio Feris", "Trevor Darrell", "Roei Herzig"], "canonical_source_url": "https://arxiv.org/abs/2609.02885", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-02885/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8703474cb4f577b307731cf1a986f76f800470fde625b6ca4b3569d30e74f903", "submitted_at": "2026-09-02T17:59:40Z", "updated_at": "2026-09-02T17:59:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.02885v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.02885"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.02885"}], "name": "Discriminative World Models for Web Agents", "publication": null, "published_at": "2026-09-02", "record_sha256": "10663e585ae18e342710e1b8c49b1e2eb55051e5773f6d8e14c123ae42814d21", "relations": [], "resource_id": "arxiv:2609.02885", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.02885v1", "topics": ["world-models", "reward-models", "planning", "data-generation", "evaluation"], "updated_at": "2026-09-02"}, "text": "Discriminative World Models for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains discriminative web world models to distinguish the true next state from outcomes of alternative actions. Branching web trajectories support training, followed by tests of predicted-state matching, process-reward action ranking and end-to-end WebArena-Lite selection.\nPlatforms: web\nTopics: world-models, reward-models, planning, data-generation, evaluation\nPublished: 2026-09-02\nAuthors: Kelvin Li, Dhruv Pendharkar, Anish Pahilajani, Chuyi Shang, Leon Oks, Leonid Karlinsky, Rogerio Feris, Trevor Darrell, Roei Herzig\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.02885v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-02885/"}
{"id": "arxiv:2609.03438", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a59f7d54845de6efebf6365a188e61e41fdf31e9ebe894b1d01ce7b86d6be4ec", "submitted_at": "2026-09-03T06:48:41Z", "updated_at": "2026-09-03T06:48:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.03438v1"}, "authors": ["Zhaoyuan Huang", "Tianjie Ju", "Pengzhou Cheng", "Zheng Wu", "Yansi Li", "Chuanbiao Song", "Jun Lan", "Huijia Zhu", "Weiqiang Wang", "Zhuosheng Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.03438", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-03438/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a59f7d54845de6efebf6365a188e61e41fdf31e9ebe894b1d01ce7b86d6be4ec", "submitted_at": "2026-09-03T06:48:41Z", "updated_at": "2026-09-03T06:48:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.03438v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.03438"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.03438"}], "name": "Do GUI Agents Know When Not to Act? Enabling Conflict-Aware Termination for Multimodal GUI Agents", "publication": null, "published_at": "2026-09-03", "record_sha256": "2cf61c3ff6eeb945f0a721004278566af88633e52713c35b96406744ca3439c5", "relations": [], "resource_id": "arxiv:2609.03438", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.03438v1", "topics": ["safety-security", "evaluation", "robustness", "planning"], "updated_at": "2026-09-03"}, "text": "Do GUI Agents Know When Not to Act? Enabling Conflict-Aware Termination for Multimodal GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCONFLICTGUI evaluates termination under instruction conflicts and mismatches between instructions and interface context. CONFLICTGUARD couples feasibility checks with conditional action modulation to reduce unnecessary execution while tracking normal-task behavior.\nPlatforms: unspecified\nTopics: safety-security, evaluation, robustness, planning\nPublished: 2026-09-03\nAuthors: Zhaoyuan Huang, Tianjie Ju, Pengzhou Cheng, Zheng Wu, Yansi Li, Chuanbiao Song, Jun Lan, Huijia Zhu, Weiqiang Wang, Zhuosheng Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.03438v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-03438/"}
{"id": "arxiv:2609.04850", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61587798aefd8e9592ddaef756a13e02f0e4a987e4068ef582d0aaf66e8eb680", "submitted_at": "2026-09-04T08:09:47Z", "updated_at": "2026-09-04T08:09:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04850v1"}, "authors": ["Weide Zhan", "Qumu Shaqu", "Yuanqing Liu", "Peng Zhang", "Jiahao Liu", "Kam Him Lam", "Ning Gu", "Zhan Hu", "Tun Lu"], "canonical_source_url": "https://arxiv.org/abs/2609.04850", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-04850/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "61587798aefd8e9592ddaef756a13e02f0e4a987e4068ef582d0aaf66e8eb680", "submitted_at": "2026-09-04T08:09:47Z", "updated_at": "2026-09-04T08:09:47Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04850v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.04850"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.04850"}], "name": "ElderBench: Benchmarking Autonomous Mobile Agents for Older Adults", "publication": null, "published_at": "2026-09-04", "record_sha256": "f878c2e65fbb1b83f4b551257c69c595d3e06597b61f076f44245453c8e97d79", "relations": [], "resource_id": "arxiv:2609.04850", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.04850v1", "topics": ["evaluation", "human-in-the-loop", "mobile-navigation", "robustness"], "updated_at": "2026-09-04"}, "text": "ElderBench: Benchmarking Autonomous Mobile Agents for Older Adults\nResource type: paper\nReview status: curated\nScope: core\n\nElderBench evaluates mobile GUI agents on smartphone requests elicited from older adults, including indirect and underspecified language. Online and offline tests, instruction normalization and linguistic analyses examine how these request patterns contribute to failures.\nPlatforms: unspecified\nTopics: evaluation, human-in-the-loop, mobile-navigation, robustness\nPublished: 2026-09-04\nAuthors: Weide Zhan, Qumu Shaqu, Yuanqing Liu, Peng Zhang, Jiahao Liu, Kam Him Lam, Ning Gu, Zhan Hu, Tun Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.04850v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-04850/"}
{"id": "arxiv:2609.04869", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78db974038e0dde9a1347132382573fa9e2aafb1b0e1a574da3d769574329be9", "submitted_at": "2026-09-04T08:29:45Z", "updated_at": "2026-09-04T08:29:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04869v1"}, "authors": ["Longtao Hu", "Xiao Liang", "Linchao Zhu"], "canonical_source_url": "https://arxiv.org/abs/2609.04869", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-04869/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "78db974038e0dde9a1347132382573fa9e2aafb1b0e1a574da3d769574329be9", "submitted_at": "2026-09-04T08:29:45Z", "updated_at": "2026-09-04T08:29:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.04869v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.04869"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.04869"}], "name": "From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents", "publication": null, "published_at": "2026-09-04", "record_sha256": "9bb437964d3519ac7dfd2354851c01924cb88f3a3bbbe34ca738b90d20557eae", "relations": [], "resource_id": "arxiv:2609.04869", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.04869v1", "topics": ["memory", "workflow", "evaluation", "training-free"], "updated_at": "2026-09-04"}, "text": "From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAn online skill-evolution framework turns GUI interaction traces and evaluator feedback into a versioned library of reusable procedures without updating model parameters. Controlled OSWorld comparisons examine both the benefits of shared procedural memory and its domain-dependent instability under repeated revision.\nPlatforms: desktop\nTopics: memory, workflow, evaluation, training-free\nPublished: 2026-09-04\nAuthors: Longtao Hu, Xiao Liang, Linchao Zhu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.04869v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-04869/"}
{"id": "arxiv:2609.05374", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69eab19b6e45e1b2ddf9f29f6f0937d51e06a5bb7dea34dadb761361c82f08d4", "submitted_at": "2026-09-04T17:26:18Z", "updated_at": "2026-09-04T17:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05374v1"}, "authors": ["Haoting Shi", "Wenhao Wang", "Weicheng Fang", "Yaozhong Liang", "Tian Jin", "Pengxiang Zhao", "Guangyi Liu", "Siheng Chen", "Yanfeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.05374", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-05374/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "69eab19b6e45e1b2ddf9f29f6f0937d51e06a5bb7dea34dadb761361c82f08d4", "submitted_at": "2026-09-04T17:26:18Z", "updated_at": "2026-09-04T17:26:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05374v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.05374"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05374"}], "name": "CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents", "publication": null, "published_at": "2026-09-04", "record_sha256": "9cdf2e823f39b52121212cdee7af5339bf5ef601424b8b5e7081495a45be7ea0", "relations": [], "resource_id": "arxiv:2609.05374", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.05374v1", "topics": ["tool-integration", "data-generation", "sandbox"], "updated_at": "2026-09-04"}, "text": "CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTurns desktop applications into reproducible GUI-and-CLI environments and generates verified hybrid trajectories for agent post-training.\nPlatforms: desktop\nTopics: tool-integration, data-generation, sandbox\nPublished: 2026-09-04\nAuthors: Haoting Shi, Wenhao Wang, Weicheng Fang, Yaozhong Liang, Tian Jin, Pengxiang Zhao, Guangyi Liu, Siheng Chen, Yanfeng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.05374v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-05374/"}
{"id": "arxiv:2609.05511", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edbe52f304634dfa06e0f232090a4fe998c7ea111418f57942976d348c4e0997", "submitted_at": "2026-08-31T08:16:56Z", "updated_at": "2026-08-31T08:16:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05511v1"}, "authors": ["Bowei He", "Xiaokun Zhang", "Meng Ding", "Xue Liu"], "canonical_source_url": "https://arxiv.org/abs/2609.05511", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-05511/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "edbe52f304634dfa06e0f232090a4fe998c7ea111418f57942976d348c4e0997", "submitted_at": "2026-08-31T08:16:56Z", "updated_at": "2026-08-31T08:16:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05511v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.05511"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05511"}], "name": "SCAFFOLD: Self-Improving Web Agents via Recursive Parametric Skill Abstraction", "publication": null, "published_at": "2026-08-31", "record_sha256": "4528e260f41fb860e47e7f4d522b41d522c363d7443d6e935081ac4c55428f35", "relations": [], "resource_id": "arxiv:2609.05511", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.05511v1", "topics": ["memory", "automation", "web-navigation"], "updated_at": "2026-08-31"}, "text": "SCAFFOLD: Self-Improving Web Agents via Recursive Parametric Skill Abstraction\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SCAFFOLD, which extracts executable skills from web trajectories, composes them into a hierarchy, and compresses redundant skills through behavioral checks. It periodically distills skill-guided trajectories into the policy and studies repeated improvement across web benchmarks.\nPlatforms: web\nTopics: memory, automation, web-navigation\nPublished: 2026-08-31\nAuthors: Bowei He, Xiaokun Zhang, Meng Ding, Xue Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.05511v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-05511/"}
{"id": "arxiv:2609.05513", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaeb2a57b2bb92ed926c9b65fb40b6bbf272e0ad9bb7415716d24eab0aeb4402", "submitted_at": "2026-08-31T08:40:34Z", "updated_at": "2026-08-31T08:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05513v1"}, "authors": ["Jianwei Zhang", "Sihan Cao", "Pengcheng Zheng", "Ya Wen", "Pei Ke", "Kuien Liu", "Shen Gao", "Wei Dong", "Yang Yang", "Chaoning Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.05513", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-05513/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aaeb2a57b2bb92ed926c9b65fb40b6bbf272e0ad9bb7415716d24eab0aeb4402", "submitted_at": "2026-08-31T08:40:34Z", "updated_at": "2026-08-31T08:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.05513v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.05513"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.05513"}], "name": "When and What to Teach: Budget-Aware Online Adaptation for Web Agents", "publication": null, "published_at": "2026-08-31", "record_sha256": "6afc644a8fc3f82cee775e43443099616cc94cec87ab393dcbd31d489b2b4bf0", "relations": [], "resource_id": "arxiv:2609.05513", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.05513v1", "topics": ["efficiency", "data-generation", "web-navigation"], "updated_at": "2026-08-31"}, "text": "When and What to Teach: Budget-Aware Online Adaptation for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProposes an online teaching framework that decides which web-agent episodes warrant a stronger teacher and which trajectory turns merit training. A solvability gate and score-guided trimming aim to preserve first-attempt performance while reducing teaching and training costs.\nPlatforms: web\nTopics: efficiency, data-generation, web-navigation\nPublished: 2026-08-31\nAuthors: Jianwei Zhang, Sihan Cao, Pengcheng Zheng, Ya Wen, Pei Ke, Kuien Liu, Shen Gao, Wei Dong, Yang Yang, Chaoning Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.05513v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-05513/"}
{"id": "arxiv:2609.06530", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0d6c92c1bbf777119fdce5e9c4c5ded744c3459f8ae0b10ccc0fc75de11b237", "submitted_at": "2026-09-06T10:50:55Z", "updated_at": "2026-09-06T10:50:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06530v1"}, "authors": ["Zirui Shang", "Xin Shu", "Yang Liu", "Zhi Gao", "Xinxiao Wu", "Lifeng Fan"], "canonical_source_url": "https://arxiv.org/abs/2609.06530", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-06530/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a0d6c92c1bbf777119fdce5e9c4c5ded744c3459f8ae0b10ccc0fc75de11b237", "submitted_at": "2026-09-06T10:50:55Z", "updated_at": "2026-09-06T10:50:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06530v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.06530"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.06530"}], "name": "Selective Knowledge Control for Continual GUI Agent Learning over Application Streams", "publication": null, "published_at": "2026-09-06", "record_sha256": "90449b620f9611ca9b47b517f8eaa4acee86c4b38b4bb5f411977256bb8ac765", "relations": [], "resource_id": "arxiv:2609.06530", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.06530v1", "topics": ["memory", "robustness"], "updated_at": "2026-09-06"}, "text": "Selective Knowledge Control for Continual GUI Agent Learning over Application Streams\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces activation-conditioned gradient control for GUI agents learning successive applications. It protects inactive application-specific neurons, constrains updates to shared neurons, and accumulates a compact historical state to balance adaptation with retention.\nPlatforms: unspecified\nTopics: memory, robustness\nPublished: 2026-09-06\nAuthors: Zirui Shang, Xin Shu, Yang Liu, Zhi Gao, Xinxiao Wu, Lifeng Fan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.06530v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-06530/"}
{"id": "arxiv:2609.06792", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14790f9565861cdba0b24661f6bfe87d7572cbad9d76d9a0b2d7216e9bac7ea7", "submitted_at": "2026-09-06T19:22:53Z", "updated_at": "2026-09-06T19:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06792v1"}, "authors": ["Juyong Lee", "Woogyeol Jin", "Kimin Lee"], "canonical_source_url": "https://arxiv.org/abs/2609.06792", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-06792/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14790f9565861cdba0b24661f6bfe87d7572cbad9d76d9a0b2d7216e9bac7ea7", "submitted_at": "2026-09-06T19:22:53Z", "updated_at": "2026-09-06T19:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.06792v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.06792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.06792"}], "name": "Improving Proficiency and Efficiency of Android GUI Agents via Self-Generating Tool Actions", "publication": null, "published_at": "2026-09-06", "record_sha256": "7941fa615d39aeeca4e53387681e57fce536fa5f8a123b2cbd14fee65a4b7eca", "relations": [], "resource_id": "arxiv:2609.06792", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.06792v1", "topics": ["tool-integration", "automation", "mobile-navigation", "efficiency"], "updated_at": "2026-09-06"}, "text": "Improving Proficiency and Efficiency of Android GUI Agents via Self-Generating Tool Actions\nResource type: paper\nReview status: curated\nScope: core\n\nDroidTool generates application-state tools through proposal, implementation, testing and repair. Relational tests connect the tools’ preconditions and effects, and the resulting tools are evaluated as additions to Android GUI agents.\nPlatforms: android\nTopics: tool-integration, automation, mobile-navigation, efficiency\nPublished: 2026-09-06\nAuthors: Juyong Lee, Woogyeol Jin, Kimin Lee\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.06792v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-06792/"}
{"id": "arxiv:2609.07712", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4090a2c7e7730ea415564b6b418d85aad77a74a222cef35435b0aa93beafd7f", "submitted_at": "2026-09-07T16:25:13Z", "updated_at": "2026-09-07T16:25:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.07712v1"}, "authors": ["Jintian Feng", "Long Chen", "Xiao Yu", "Jiayi Dai", "Chenglong Liu", "Haoru Wang", "Zizhen Xue", "Yuxuan Shi", "Ziyang Wang", "Yichen Gong"], "canonical_source_url": "https://arxiv.org/abs/2609.07712", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-07712/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b4090a2c7e7730ea415564b6b418d85aad77a74a222cef35435b0aa93beafd7f", "submitted_at": "2026-09-07T16:25:13Z", "updated_at": "2026-09-07T16:25:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.07712v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.07712"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.07712"}], "name": "APPSim-Bench: Bridging Real-world Apps and Reproducible Evaluation for Mobile GUI Agents", "publication": null, "published_at": "2026-09-07", "record_sha256": "c98d2728695e6c88c31691f40899f6b3565abecff3409ac5f970e642618f04b1", "relations": [], "resource_id": "arxiv:2609.07712", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.07712v1", "topics": ["evaluation", "sandbox", "mobile-navigation"], "updated_at": "2026-09-07"}, "text": "APPSim-Bench: Bridging Real-world Apps and Reproducible Evaluation for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces controlled simulated mobile apps with deterministic backend data and outcome checks for reproducible evaluation of realistic interactions.\nPlatforms: unspecified\nTopics: evaluation, sandbox, mobile-navigation\nPublished: 2026-09-07\nAuthors: Jintian Feng, Long Chen, Xiao Yu, Jiayi Dai, Chenglong Liu, Haoru Wang, Zizhen Xue, Yuxuan Shi, Ziyang Wang, Yichen Gong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.07712v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-07712/"}
{"id": "arxiv:2609.09212", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3a7a81ffe43885d9c8362df4c8aff08000304fe60477f484a732caf35124b2", "submitted_at": "2026-09-06T03:49:45Z", "updated_at": "2026-09-06T03:49:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.09212v1"}, "authors": ["Zhihao Liu", "Hongyu Sun", "Zhiyuan Fu", "Xiaonan Duan", "Jice Wang", "Shangru Zhao", "Weizhi Meng", "Wuxin Yang", "Yangfan Zhou", "Yuqing Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.09212", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-09212/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5a3a7a81ffe43885d9c8362df4c8aff08000304fe60477f484a732caf35124b2", "submitted_at": "2026-09-06T03:49:45Z", "updated_at": "2026-09-06T03:49:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.09212v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.09212"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.09212"}], "name": "AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents", "publication": null, "published_at": "2026-09-06", "record_sha256": "629b914464775f5594f74d025f816b3b0d0d624a96d6794ddccef06aa566ff29", "relations": [], "resource_id": "arxiv:2609.09212", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.09212v1", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": "2026-09-06"}, "text": "AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAgentHijack evaluates whether visual patch effects propagate from screenshots through model outputs and action parsing into real environment changes. Its controlled web-page experiments distinguish intermediate attack signals from verified end-to-end consequences.\nPlatforms: web\nTopics: safety-security, evaluation, robustness\nPublished: 2026-09-06\nAuthors: Zhihao Liu, Hongyu Sun, Zhiyuan Fu, Xiaonan Duan, Jice Wang, Shangru Zhao, Weizhi Meng, Wuxin Yang, Yangfan Zhou, Yuqing Zhang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.09212v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-09212/"}
{"id": "arxiv:2609.10031", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3307d90ef09c16afaee160eacc9d2994d142be9b9a26ce9c4a57f61345e2e8ec", "submitted_at": "2026-09-09T11:04:53Z", "updated_at": "2026-09-09T11:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10031v1"}, "authors": ["Xiaolei Li", "Jialun Cao", "Zhijian Hou", "Yuzhi Zhao", "Yepang Liu", "Shing-Chi Cheung"], "canonical_source_url": "https://arxiv.org/abs/2609.10031", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-10031/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3307d90ef09c16afaee160eacc9d2994d142be9b9a26ce9c4a57f61345e2e8ec", "submitted_at": "2026-09-09T11:04:53Z", "updated_at": "2026-09-09T11:04:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10031v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:13Z", "role": "primary", "sha256": "ca895b2f0dcd74816fb5e775a7161a0932ff7120a72477791f5e4c0bb1ead88e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2608.17319%2C2608.18307%2C2608.20319%2C2608.20707%2C2608.20797%2C2608.21794%2C2608.21830%2C2608.21832%2C2608.21898%2C2608.22077%2C2608.22533%2C2608.22577%2C2608.22847%2C2608.24015%2C2608.24017%2C2608.24099%2C2608.24174%2C2608.24680%2C2608.24848%2C2608.25777%2C2608.26200%2C2608.26204%2C2608.27477%2C2608.27508%2C2608.27808%2C2608.27866%2C2608.29305%2C2608.30038%2C2608.30207%2C2609.00028%2C2609.00048%2C2609.00524%2C2609.02057%2C2609.02149%2C2609.02154%2C2609.02309%2C2609.02885%2C2609.03438%2C2609.04850%2C2609.04869%2C2609.05374%2C2609.05511%2C2609.05513%2C2609.05535%2C2609.06530%2C2609.06792%2C2609.07603%2C2609.07712%2C2609.09212%2C2609.10031&max_results=50"}], "identifiers": {"arxiv": "2609.10031"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:13Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10031"}], "name": "GraphDroid: Asynchronous LLM-Based Mobile App GUI Testing via History-Aware Exploration and Hybrid Intent Fulfillment", "publication": null, "published_at": "2026-09-09", "record_sha256": "79227c7b2ec2b84530c48fa2aa13391e3afde0207281e30544b2b3bfa32c75a8", "relations": [], "resource_id": "arxiv:2609.10031", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.10031v1", "topics": ["evaluation", "mobile-navigation", "memory", "interaction-history", "efficiency"], "updated_at": "2026-09-09"}, "text": "GraphDroid: Asynchronous LLM-Based Mobile App GUI Testing via History-Aware Exploration and Hybrid Intent Fulfillment\nResource type: paper\nReview status: curated\nScope: core\n\nGraphDroid combines clustered memory of visited GUI states with asynchronous test-intent generation. A hybrid executor reserves LLM calls for complex intents and uses heuristics for simpler ones, evaluated through Android app coverage, bug discovery and crash detection.\nPlatforms: android\nTopics: evaluation, mobile-navigation, memory, interaction-history, efficiency\nPublished: 2026-09-09\nAuthors: Xiaolei Li, Jialun Cao, Zhijian Hou, Yuzhi Zhao, Yepang Liu, Shing-Chi Cheung\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.10031v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-10031/"}
{"id": "arxiv:2609.10297", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20d3bc3137e163f6b467b8c464101478b2375fbab8aebe9179e2e1113b34f503", "submitted_at": "2026-09-09T15:12:48Z", "updated_at": "2026-09-09T15:12:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10297v1"}, "authors": ["Yuhao Wang", "Mu Qiao", "Xindong Zhang", "Yunzhi Zhuge", "Lei Zhang", "Huchuan Lu"], "canonical_source_url": "https://arxiv.org/abs/2609.10297", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-10297/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "20d3bc3137e163f6b467b8c464101478b2375fbab8aebe9179e2e1113b34f503", "submitted_at": "2026-09-09T15:12:48Z", "updated_at": "2026-09-09T15:12:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10297v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.10297"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10297"}], "name": "TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents", "publication": null, "published_at": "2026-09-09", "record_sha256": "985db8c8e032157f3a9539257bac5894b8bf38fba49417081fa94af57c143183", "relations": [], "resource_id": "arxiv:2609.10297", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.10297v1", "topics": ["efficiency", "interaction-history", "screen-understanding", "training-free"], "updated_at": "2026-09-09"}, "text": "TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTRACE orders screenshot evidence for monotonic visual-token reduction using layout priors, instruction relevance and feature novelty. Spatial token reserves and incremental KV contraction aim to preserve useful evidence for future GUI targets without repeatedly encoding retired frames.\nPlatforms: unspecified\nTopics: efficiency, interaction-history, screen-understanding, training-free\nPublished: 2026-09-09\nAuthors: Yuhao Wang, Mu Qiao, Xindong Zhang, Yunzhi Zhuge, Lei Zhang, Huchuan Lu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.10297v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-10297/"}
{"id": "arxiv:2609.10451", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ec17516bce87a8e349983acac14415a6ceeb12c167fce57aa37d6157cd3cd83", "submitted_at": "2026-09-09T16:59:59Z", "updated_at": "2026-09-09T16:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10451v1"}, "authors": ["Zixiang Chen", "Yuheng Lu", "Zihao Cheng", "Zeming Liu", "Jizeng Bai", "Ziye Huang", "Zhiyin Lin", "Zihan Li", "Yuhang Guo", "Yunhong Wang", "Haifeng Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.10451", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-10451/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5ec17516bce87a8e349983acac14415a6ceeb12c167fce57aa37d6157cd3cd83", "submitted_at": "2026-09-09T16:59:59Z", "updated_at": "2026-09-09T16:59:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.10451v1"}}, "environments": ["desktop", "android", "windows", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.10451"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.10451"}], "name": "JarvisGUI: Towards Cross-Device GUI Agents with Dynamic Task Composition", "publication": null, "published_at": "2026-09-09", "record_sha256": "4367545c329c2d93d7440cf2a55c1045a0fa49083a65f587ca0e81add10c246d", "relations": [], "resource_id": "arxiv:2609.10451", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.10451v1", "topics": ["evaluation", "cross-app", "planning", "long-horizon", "state-tracking"], "updated_at": "2026-09-09"}, "text": "JarvisGUI: Towards Cross-Device GUI Agents with Dynamic Task Composition\nResource type: paper\nReview status: curated\nScope: core\n\nJarvisGUI composes typed input-output task transformations into dynamic workflows spanning Android, Windows and Ubuntu devices. Its virtual-environment benchmark tests transfer of intermediate results, shared state and cross-device dependencies.\nPlatforms: desktop, android, windows, linux\nTopics: evaluation, cross-app, planning, long-horizon, state-tracking\nPublished: 2026-09-09\nAuthors: Zixiang Chen, Yuheng Lu, Zihao Cheng, Zeming Liu, Jizeng Bai, Ziye Huang, Zhiyin Lin, Zihan Li, Yuhang Guo, Yunhong Wang, Haifeng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.10451v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-10451/"}
{"id": "arxiv:2609.12394", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc15573bca49d453001ec22061f190e429704a5318751b041550e3610c469f4e", "submitted_at": "2026-09-11T03:36:35Z", "updated_at": "2026-09-15T06:36:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.12394v3"}, "authors": ["Tong Ye", "Kunyang Han", "Guozhi Wang", "Longqiang Luo", "Zhifeng Ding", "Yongxiang Zhang", "Xiaolei Shen", "Yuxuan Zhang", "Zhuping Zhang", "Tao Xu", "Yue Pan", "Yucheng Zhao", "Yupei Hu", "Yuanjiang Ouyang", "Danfeng Shen", "Runqi Lin", "Hongda Cai", "Zhaoxiong Wang", "Mengjia Yan", "Yingjie Zhong", "Chen Zhou", "Zeyu Zhang", "Xuwen Zhu", "Penggang Shi", "Mingcheng Luo", "Ziyang Wu", "Min Jin", "Mingfu Shen", "Zairong Xu", "Fan Zhang", "Hao Wang", "Liang Liu", "Zhulin Xie", "Lijun Yao", "Xiao Liang", "Liangmin Wen", "Liqiang Feng", "Feilong Wu", "Min Hu", "Min Chen", "Guanjing Xiong", "Xiaohu Ruan", "Xiaoxin Chen"], "canonical_source_url": "https://arxiv.org/abs/2609.12394", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-12394/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cc15573bca49d453001ec22061f190e429704a5318751b041550e3610c469f4e", "submitted_at": "2026-09-11T03:36:35Z", "updated_at": "2026-09-15T06:36:04Z", "version": 3, "versioned_url": "https://arxiv.org/abs/2609.12394v3"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.12394"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.12394"}], "name": "BlueLM-GUI Technical Report: A Real-Device-Centric Flywheel for Self-Improving Mobile GUI Agents", "publication": null, "published_at": "2026-09-11", "record_sha256": "aab835068bb1070ef070294c46946bef00c9cf944aa58fdce457f50058cc5c61", "relations": [], "resource_id": "arxiv:2609.12394", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.12394v3", "topics": ["mobile-navigation", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-15"}, "text": "BlueLM-GUI Technical Report: A Real-Device-Centric Flywheel for Self-Improving Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a mobile-agent training pipeline centered on real-device rollouts, recovery of useful supervision from failures, and evolving evaluation tasks.\nPlatforms: android\nTopics: mobile-navigation, reinforcement-learning, data-generation\nPublished: 2026-09-11\nAuthors: Tong Ye, Kunyang Han, Guozhi Wang, Longqiang Luo, Zhifeng Ding, Yongxiang Zhang, Xiaolei Shen, Yuxuan Zhang, Zhuping Zhang, Tao Xu, Yue Pan, Yucheng Zhao, Yupei Hu, Yuanjiang Ouyang, Danfeng Shen, Runqi Lin, Hongda Cai, Zhaoxiong Wang, Mengjia Yan, Yingjie Zhong, Chen Zhou, Zeyu Zhang, Xuwen Zhu, Penggang Shi, Mingcheng Luo, Ziyang Wu, Min Jin, Mingfu Shen, Zairong Xu, Fan Zhang, Hao Wang, Liang Liu, Zhulin Xie, Lijun Yao, Xiao Liang, Liangmin Wen, Liqiang Feng, Feilong Wu, Min Hu, Min Chen, Guanjing Xiong, Xiaohu Ruan, Xiaoxin Chen\nRecorded arXiv version: v3\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.12394v3\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-12394/"}
{"id": "arxiv:2609.13287", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d7e9643c0543f088e5d2b17699c843a7643c611315f20ce92e83aea849b931a", "submitted_at": "2026-09-09T05:53:39Z", "updated_at": "2026-09-09T05:53:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13287v1"}, "authors": ["Zhangxuan Gu", "Haoxing Chen", "Qi Qin", "Yi Xin", "Kai Gan", "Lin Liu", "Long Cui", "Xiaomei Wang", "Beitong Zhou", "Yunzhu Zhang", "Zhengwen Zeng", "Changlong Gao", "Weizhi Chen", "Rongchao Zhang", "Haoyuan Wu", "Shuheng Shen", "Changhua Meng", "Weiqiang Wang", "Jianguo Li", "Zhenzhong Lan"], "canonical_source_url": "https://arxiv.org/abs/2609.13287", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-13287/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2d7e9643c0543f088e5d2b17699c843a7643c611315f20ce92e83aea849b931a", "submitted_at": "2026-09-09T05:53:39Z", "updated_at": "2026-09-09T05:53:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13287v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.13287"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13287"}], "name": "LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents", "publication": null, "published_at": "2026-09-09", "record_sha256": "92fd95cfe51dbcc4d6e0780063e02592bbd1c0a3416d7cc685045c207115707c", "relations": [], "resource_id": "arxiv:2609.13287", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.13287v1", "topics": ["grounding", "imitation-learning", "efficiency", "screen-understanding"], "updated_at": "2026-09-09"}, "text": "LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nLLaDA-UI adapts block-wise diffusion generation to a vision-language GUI agent through multimodal pretraining and GUI supervised fine-tuning. It evaluates grounding and navigation across multiple platform settings.\nPlatforms: web, desktop\nTopics: grounding, imitation-learning, efficiency, screen-understanding\nPublished: 2026-09-09\nAuthors: Zhangxuan Gu, Haoxing Chen, Qi Qin, Yi Xin, Kai Gan, Lin Liu, Long Cui, Xiaomei Wang, Beitong Zhou, Yunzhu Zhang, Zhengwen Zeng, Changlong Gao, Weizhi Chen, Rongchao Zhang, Haoyuan Wu, Shuheng Shen, Changhua Meng, Weiqiang Wang, Jianguo Li, Zhenzhong Lan\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.13287v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-13287/"}
{"id": "arxiv:2609.13491", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b615736812b466b0df7133b7070184a1b7c8dacbe4c73fe32b878b654bb78d", "submitted_at": "2026-09-11T19:57:22Z", "updated_at": "2026-09-11T19:57:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13491v1"}, "authors": ["Alexandru Ianta", "Eleni Stroulia"], "canonical_source_url": "https://arxiv.org/abs/2609.13491", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-13491/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "29b615736812b466b0df7133b7070184a1b7c8dacbe4c73fe32b878b654bb78d", "submitted_at": "2026-09-11T19:57:22Z", "updated_at": "2026-09-11T19:57:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13491v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.13491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13491"}], "name": "Token Efficient Task Execution via Application Behavior Modeling for Web Agents", "publication": null, "published_at": "2026-09-11", "record_sha256": "33d7ee69344064d0ebcf8b90854b989e35afb53de3d71619ab3c5b3321d628ef", "relations": [], "resource_id": "arxiv:2609.13491", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.13491v1", "topics": ["memory", "efficiency", "web-navigation"], "updated_at": "2026-09-11"}, "text": "Token Efficient Task Execution via Application Behavior Modeling for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nOdoBot builds a model of web-application behavior from successful task demonstrations and uses it to guide task execution. Its Canvas LMS evaluation compares token use and completion with existing web agents.\nPlatforms: web\nTopics: memory, efficiency, web-navigation\nPublished: 2026-09-11\nAuthors: Alexandru Ianta, Eleni Stroulia\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.13491v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-13491/"}
{"id": "arxiv:2609.13548", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce929c347ed5160b1f8455cc9d7e73ac2cc1b678e9d386d00dd4b2ef146b56d", "submitted_at": "2026-09-11T21:24:40Z", "updated_at": "2026-09-11T21:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13548v1"}, "authors": ["Xinyun Cao", "Adriana Szekeres", "Fazle Elahi Faisal"], "canonical_source_url": "https://arxiv.org/abs/2609.13548", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-13548/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ce929c347ed5160b1f8455cc9d7e73ac2cc1b678e9d386d00dd4b2ef146b56d", "submitted_at": "2026-09-11T21:24:40Z", "updated_at": "2026-09-11T21:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13548v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.13548"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13548"}], "name": "AutoTailor: Automatic, User-Aligned Capability Selection and Adaptation for Web Agents", "publication": null, "published_at": "2026-09-11", "record_sha256": "eee8bb3744fab08a7198483e7ad78b8402f21961a16253b9512b63427bdb6d6f", "relations": [], "resource_id": "arxiv:2609.13548", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.13548v1", "topics": ["tool-integration", "web-navigation", "workflow", "efficiency"], "updated_at": "2026-09-11"}, "text": "AutoTailor: Automatic, User-Aligned Capability Selection and Adaptation for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAutoTailor compiles web trajectories into reusable browser-automation APIs, then filters and dynamically revises the capability inventory. WebArena experiments distinguish API-only operation from an agent that can fall back to ordinary reasoning and browser actions.\nPlatforms: web\nTopics: tool-integration, web-navigation, workflow, efficiency\nPublished: 2026-09-11\nAuthors: Xinyun Cao, Adriana Szekeres, Fazle Elahi Faisal\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.13548v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-13548/"}
{"id": "arxiv:2609.13873", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1eda33b358d4c4caf8c1fb0b1bd92698c3b37cee995d820f40ecffe82010edb4", "submitted_at": "2026-09-12T10:55:54Z", "updated_at": "2026-09-12T10:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13873v1"}, "authors": ["Qihang Cen", "Tianshuo Cong", "Da Song", "Xinlei He", "Jiaxing Song", "Ke Xu", "Qi Li"], "canonical_source_url": "https://arxiv.org/abs/2609.13873", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-13873/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1eda33b358d4c4caf8c1fb0b1bd92698c3b37cee995d820f40ecffe82010edb4", "submitted_at": "2026-09-12T10:55:54Z", "updated_at": "2026-09-12T10:55:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.13873v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.13873"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.13873"}], "name": "PriMobiBench: Characterizing Visual Privacy Leakage in VLM-Driven Mobile GUI Agents", "publication": null, "published_at": "2026-09-12", "record_sha256": "2c1595cba4b4660b44fffcac07f9d2c37949348cd81f860ad01293ef70e4591d", "relations": [], "resource_id": "arxiv:2609.13873", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.13873v1", "topics": ["evaluation", "safety-security", "screen-understanding", "mobile-navigation"], "updated_at": "2026-09-12"}, "text": "PriMobiBench: Characterizing Visual Privacy Leakage in VLM-Driven Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPriMobiBench evaluates direct privacy leakage and user profiling from mobile GUI screenshot trajectories, supported by the MobiLeak trace dataset. It also tests masking sensitive, task-irrelevant interface elements before cloud processing and measures the resulting privacy and utility trade-off.\nPlatforms: unspecified\nTopics: evaluation, safety-security, screen-understanding, mobile-navigation\nPublished: 2026-09-12\nAuthors: Qihang Cen, Tianshuo Cong, Da Song, Xinlei He, Jiaxing Song, Ke Xu, Qi Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.13873v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-13873/"}
{"id": "arxiv:2609.15307", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e2081a382ae815b48eb3166d18ee7c6e53afb4cd6d75538f3ff3e29f2ed5954", "submitted_at": "2026-09-14T09:56:10Z", "updated_at": "2026-09-14T09:56:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15307v1"}, "authors": ["Yizhou Liu", "Fei Tang", "Yuchen Yan", "Zhengxi Lu", "Songqin Nong", "Tao Jiang", "Wenhao Xu", "Wenqi Zhang", "Weiming Lu", "Jun Xiao", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2609.15307", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-15307/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3e2081a382ae815b48eb3166d18ee7c6e53afb4cd6d75538f3ff3e29f2ed5954", "submitted_at": "2026-09-14T09:56:10Z", "updated_at": "2026-09-14T09:56:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15307v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.15307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.15307"}], "name": "Learning from Reliable Negatives: Confidence-Anchored Test-Time Adaptation for GUI Grounding", "publication": null, "published_at": "2026-09-14", "record_sha256": "bfd0a0705605ecd6a8d37adf5ec6e1c07fcbc7af85b75bdec62d952f213dd768", "relations": [], "resource_id": "arxiv:2609.15307", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.15307v1", "topics": ["grounding", "robustness", "evaluation"], "updated_at": "2026-09-14"}, "text": "Learning from Reliable Negatives: Confidence-Anchored Test-Time Adaptation for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nConfidence-Anchored Learning uses coordinate-token confidence to filter pseudo-supervision for GUI grounding. Its negative-only variant performs label-free test-time updates to reduce reliance on potentially incorrect positive examples.\nPlatforms: unspecified\nTopics: grounding, robustness, evaluation\nPublished: 2026-09-14\nAuthors: Yizhou Liu, Fei Tang, Yuchen Yan, Zhengxi Lu, Songqin Nong, Tao Jiang, Wenhao Xu, Wenqi Zhang, Weiming Lu, Jun Xiao, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.15307v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-15307/"}
{"id": "arxiv:2609.15457", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a98337941972fadb725906b9d1315a0ca5921859bd04780cacbabb0500e6a5", "submitted_at": "2026-09-14T12:22:26Z", "updated_at": "2026-09-14T12:22:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15457v1"}, "authors": ["Shengjie Jin", "Zelong Sun", "Hengbo Xu", "Yanbiao Ma", "Zhiwu Lu"], "canonical_source_url": "https://arxiv.org/abs/2609.15457", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-15457/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8a98337941972fadb725906b9d1315a0ca5921859bd04780cacbabb0500e6a5", "submitted_at": "2026-09-14T12:22:26Z", "updated_at": "2026-09-14T12:22:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.15457v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.15457"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.15457"}], "name": "AnchorGUI: Asymmetric Memory for Dual-Scale Learning in GUI Navigation", "publication": null, "published_at": "2026-09-14", "record_sha256": "641ac7f5b03149592dc8d6502d9d6cd9ba3f29859ca6b346c8b86834de0b87a5", "relations": [], "resource_id": "arxiv:2609.15457", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.15457v1", "topics": ["memory", "interaction-history", "state-tracking", "self-reflection", "efficiency"], "updated_at": "2026-09-14"}, "text": "AnchorGUI: Asymmetric Memory for Dual-Scale Learning in GUI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nAnchorGUI compares expected and observed UI transitions and selectively retains screenshots when they disagree. This asymmetric memory supports immediate correction and focuses cross-attempt experience distillation on likely failure steps.\nPlatforms: android\nTopics: memory, interaction-history, state-tracking, self-reflection, efficiency\nPublished: 2026-09-14\nAuthors: Shengjie Jin, Zelong Sun, Hengbo Xu, Yanbiao Ma, Zhiwu Lu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.15457v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-15457/"}
{"id": "arxiv:2609.16251", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f5fd0a3024311f9aba44a2971c9c752ee2c38444139823746b90a7c9a6d5ed4", "submitted_at": "2026-09-14T19:19:05Z", "updated_at": "2026-09-21T14:30:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.16251v2"}, "authors": ["Zihan Dong", "Yuanzhe Liu", "Zhiyuan Ma", "Qishi Zhan", "Dehan Kong", "Guohao Li", "Kaixin Li"], "canonical_source_url": "https://arxiv.org/abs/2609.16251", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-16251/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f5fd0a3024311f9aba44a2971c9c752ee2c38444139823746b90a7c9a6d5ed4", "submitted_at": "2026-09-14T19:19:05Z", "updated_at": "2026-09-21T14:30:34Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.16251v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.16251"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16251"}], "name": "CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design", "publication": null, "published_at": "2026-09-14", "record_sha256": "a8443c68360123ee6f4efe393f7f14aaf4b15e13c102c41163242c6f0f5ea32a", "relations": [], "resource_id": "arxiv:2609.16251", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.16251v2", "topics": ["evaluation", "long-horizon", "workflow", "science"], "updated_at": "2026-09-21"}, "text": "CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design\nResource type: paper\nReview status: curated\nScope: core\n\nCADWorld evaluates screenshot-and-GUI agents on long FreeCAD workflows, from modeling and assembly to manufacturing and simulation. Executable checks inspect saved engineering artifacts for geometry, constraints, construction structure and downstream state.\nPlatforms: desktop\nTopics: evaluation, long-horizon, workflow, science\nPublished: 2026-09-14\nAuthors: Zihan Dong, Yuanzhe Liu, Zhiyuan Ma, Qishi Zhan, Dehan Kong, Guohao Li, Kaixin Li\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.16251v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-16251/"}
{"id": "arxiv:2609.16635", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8dcd34beebb8a4660d8d7b9f4c76c3f86069926b9856d26b1c20d87dec83bb5", "submitted_at": "2026-09-15T04:56:17Z", "updated_at": "2026-09-15T04:56:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16635v1"}, "authors": ["Yao Zhao", "Aditya Shanmugham", "Swastik Roy", "Yanxun Xu"], "canonical_source_url": "https://arxiv.org/abs/2609.16635", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-16635/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e8dcd34beebb8a4660d8d7b9f4c76c3f86069926b9856d26b1c20d87dec83bb5", "submitted_at": "2026-09-15T04:56:17Z", "updated_at": "2026-09-15T04:56:17Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16635v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.16635"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16635"}], "name": "EchoPath: Execution-Level Replayable Memory for GUI Agents", "publication": null, "published_at": "2026-09-15", "record_sha256": "f0a011a1fde01ffbf10e56689cea046bc0d30fd10a1de9666d0a524de158cf72", "relations": [], "resource_id": "arxiv:2609.16635", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.16635v1", "topics": ["memory", "workflow", "automation", "grounding", "efficiency", "robustness"], "updated_at": "2026-09-15"}, "text": "EchoPath: Execution-Level Replayable Memory for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEchoPath turns validated GUI trajectories into parameterized replay procedures with explicit runtime preconditions. Image-based target matching adjusts stored coordinates, while ambiguous steps fall back to bounded repair or fresh planning.\nPlatforms: unspecified\nTopics: memory, workflow, automation, grounding, efficiency, robustness\nPublished: 2026-09-15\nAuthors: Yao Zhao, Aditya Shanmugham, Swastik Roy, Yanxun Xu\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.16635v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-16635/"}
{"id": "arxiv:2609.16732", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b86516b7695d23ffc9179c0829209dc28b96c3c819525eab62a293a68ab5daf0", "submitted_at": "2026-09-15T07:02:10Z", "updated_at": "2026-09-15T07:02:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16732v1"}, "authors": ["Heng Li", "Fulin Zhao", "Zhe Geng", "Zhiyuan Yao", "Wei Yuan", "Xiapu Luo"], "canonical_source_url": "https://arxiv.org/abs/2609.16732", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-16732/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b86516b7695d23ffc9179c0829209dc28b96c3c819525eab62a293a68ab5daf0", "submitted_at": "2026-09-15T07:02:10Z", "updated_at": "2026-09-15T07:02:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.16732v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.16732"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.16732"}], "name": "When Agents See Differently: Exposing UI Desynchronization Threats in Mobile Agents", "publication": null, "published_at": "2026-09-15", "record_sha256": "79be9c0aefb4ef33ac65d53e1fc4e50bc4733f131d924f24d0b42247e4c5479f", "relations": [], "resource_id": "arxiv:2609.16732", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.16732v1", "topics": ["safety-security", "screen-understanding", "evaluation", "robustness"], "updated_at": "2026-09-15"}, "text": "When Agents See Differently: Exposing UI Desynchronization Threats in Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nInvestigates mismatches between what humans see on a mobile display and what agents receive through screenshots or accessibility metadata. An automated evaluation framework embeds instruction-agnostic perturbations into functional APK clones and compares static and dynamic agent deviations with human noticeability.\nPlatforms: android\nTopics: safety-security, screen-understanding, evaluation, robustness\nPublished: 2026-09-15\nAuthors: Heng Li, Fulin Zhao, Zhe Geng, Zhiyuan Yao, Wei Yuan, Xiapu Luo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.16732v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-16732/"}
{"id": "arxiv:2609.17653", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "975662f216cc9a6b89856d4457f2125b733620e80bfac1b6424775253071d2e6", "submitted_at": "2026-09-15T17:24:06Z", "updated_at": "2026-09-15T17:24:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.17653v1"}, "authors": ["Bofan Chen", "Boxuan Zhang", "Fei Tang", "Zhengxi Lu", "Yong Du", "Tongbo Chen", "Weiming Lu", "Jun Xiao", "Yueting Zhuang", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2609.17653", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-17653/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "975662f216cc9a6b89856d4457f2125b733620e80bfac1b6424775253071d2e6", "submitted_at": "2026-09-15T17:24:06Z", "updated_at": "2026-09-15T17:24:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.17653v1"}}, "environments": ["android", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.17653"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.17653"}], "name": "Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents", "publication": null, "published_at": "2026-09-15", "record_sha256": "d6355c78a8d2548ea27ee5aa0823dc302711027c2d136f1a10759ea46e817dac", "relations": [], "resource_id": "arxiv:2609.17653", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.17653v1", "topics": ["memory", "training-free", "self-reflection", "robustness", "long-horizon", "workflow"], "updated_at": "2026-09-15"}, "text": "Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nEvoSkill-GUI revises reusable skill files from execution feedback through an executor–critic loop. It studies how training-free updates to localization, recovery and procedure information help subsequent mobile and desktop tasks.\nPlatforms: android, desktop\nTopics: memory, training-free, self-reflection, robustness, long-horizon, workflow\nPublished: 2026-09-15\nAuthors: Bofan Chen, Boxuan Zhang, Fei Tang, Zhengxi Lu, Yong Du, Tongbo Chen, Weiming Lu, Jun Xiao, Yueting Zhuang, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.17653v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-17653/"}
{"id": "arxiv:2609.17885", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab0dbea71d0b8bfe035d3b0a71d6ed99fbd261202976cdc322a0769074827d76", "submitted_at": "2026-09-15T22:23:17Z", "updated_at": "2026-09-23T01:31:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.17885v2"}, "authors": ["Kratika Bhagtani", "Kusha Sridhar", "Maziyar Baran Pouyan", "Yuying Zhao", "Eugene Siow"], "canonical_source_url": "https://arxiv.org/abs/2609.17885", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-17885/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ab0dbea71d0b8bfe035d3b0a71d6ed99fbd261202976cdc322a0769074827d76", "submitted_at": "2026-09-15T22:23:17Z", "updated_at": "2026-09-23T01:31:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.17885v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.17885"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.17885"}], "name": "ERPBench: A State-Grounded Evaluation Paradigm for Computer-Use Agents in Enterprise Software", "publication": null, "published_at": "2026-09-15", "record_sha256": "2a457567e2fa9ebb563e1b3c74c6197d733336e4e2b5f18af53e2fbf0c0ec13d", "relations": [], "resource_id": "arxiv:2609.17885", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.17885v2", "topics": ["evaluation", "workflow", "human-in-the-loop", "robustness"], "updated_at": "2026-09-23"}, "text": "ERPBench: A State-Grounded Evaluation Paradigm for Computer-Use Agents in Enterprise Software\nResource type: paper\nReview status: curated\nScope: core\n\nERPBench evaluates screenshot-only agents in enterprise software using database values to check the business records produced by their actions. It pairs these state-based checks with a harness for human approval and analyzes cases where visible completion masks incorrect persisted data.\nPlatforms: unspecified\nTopics: evaluation, workflow, human-in-the-loop, robustness\nPublished: 2026-09-15\nAuthors: Kratika Bhagtani, Kusha Sridhar, Maziyar Baran Pouyan, Yuying Zhao, Eugene Siow\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.17885v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-17885/"}
{"id": "arxiv:2609.18690", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccbc85f4d2598b4cb1e6e1a96b3b9184ddeba838cb610658070fd2d1ac9c8d06", "submitted_at": "2026-09-16T14:06:38Z", "updated_at": "2026-09-16T14:06:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.18690v1"}, "authors": ["Liyang Fan", "Xinping Bi", "Yitai Li", "Shuaimin Li", "Hui Li", "Min Yang"], "canonical_source_url": "https://arxiv.org/abs/2609.18690", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-18690/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ccbc85f4d2598b4cb1e6e1a96b3b9184ddeba838cb610658070fd2d1ac9c8d06", "submitted_at": "2026-09-16T14:06:38Z", "updated_at": "2026-09-16T14:06:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.18690v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.18690"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.18690"}], "name": "RankGround: Efficient High-Resolution GUI Grounding via Lightweight Reranker-Guided Crop Selection", "publication": null, "published_at": "2026-09-16", "record_sha256": "7540889767b79f6995de2bf8e7e94bd03249dca136e595f303147ef718eecde3", "relations": [], "resource_id": "arxiv:2609.18690", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.18690v1", "topics": ["grounding", "efficiency", "data-generation", "screen-understanding"], "updated_at": "2026-09-16"}, "text": "RankGround: Efficient High-Resolution GUI Grounding via Lightweight Reranker-Guided Crop Selection\nResource type: paper\nReview status: curated\nScope: core\n\nRankGround trains a lightweight reranker to select a promising screenshot crop before one VLM localization call. Ranking supervision uses containment and boundary-aware augmentation, with pointwise and listwise training stages.\nPlatforms: unspecified\nTopics: grounding, efficiency, data-generation, screen-understanding\nPublished: 2026-09-16\nAuthors: Liyang Fan, Xinping Bi, Yitai Li, Shuaimin Li, Hui Li, Min Yang\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.18690v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-18690/"}
{"id": "arxiv:2609.19523", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e7fb5704932a99147ff7c64a82a6a4041565f3edf6506a6252f75f832834d96", "submitted_at": "2026-09-17T00:29:46Z", "updated_at": "2026-10-08T04:53:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.19523v2"}, "authors": ["Yinzhu Quan", "Zefang Liu"], "canonical_source_url": "https://arxiv.org/abs/2609.19523", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-19523/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5e7fb5704932a99147ff7c64a82a6a4041565f3edf6506a6252f75f832834d96", "submitted_at": "2026-09-17T00:29:46Z", "updated_at": "2026-10-08T04:53:11Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.19523v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.19523"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.19523"}], "name": "EconSkills: Studying Skill Transfer and Retrieval for Web Agents on Live Economic Data", "publication": null, "published_at": "2026-09-17", "record_sha256": "b9e78ffd41c7fb761a3bce37e3bcf749a510393cdbe85df5c06e1f0dac09f569", "relations": [], "resource_id": "arxiv:2609.19523", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.19523v2", "topics": ["memory", "web-navigation", "evaluation", "workflow"], "updated_at": "2026-10-08"}, "text": "EconSkills: Studying Skill Transfer and Retrieval for Web Agents on Live Economic Data\nResource type: paper\nReview status: curated\nScope: core\n\nEconSkills converts verified economic-data browsing traces into parameterized procedures with navigation, verification and recovery guidance. Its evaluation separates known-skill transfer from library retrieval, exposing the limits of approximate matches to new tasks.\nPlatforms: web\nTopics: memory, web-navigation, evaluation, workflow\nPublished: 2026-09-17\nAuthors: Yinzhu Quan, Zefang Liu\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.19523v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-19523/"}
{"id": "arxiv:2609.19843", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70a7588abcf345418ed759fb2bdb96fc60c6e1444a3e0dc98766510c755f8478", "submitted_at": "2026-09-17T07:54:15Z", "updated_at": "2026-09-17T07:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.19843v1"}, "authors": ["Haya Halimeh", "Sascha Kaltenpoth", "Kevin Bösch", "Oliver Müller"], "canonical_source_url": "https://arxiv.org/abs/2609.19843", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-19843/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "70a7588abcf345418ed759fb2bdb96fc60c6e1444a3e0dc98766510c755f8478", "submitted_at": "2026-09-17T07:54:15Z", "updated_at": "2026-09-17T07:54:15Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.19843v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.19843"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.19843"}], "name": "A Dual-Process Perspective on Nudge Susceptibility in LLM-Based GUI Agents", "publication": null, "published_at": "2026-09-17", "record_sha256": "03f4c75eeef4867dfbcc6b28b6a3ea301aa7f2ecf064818c5c469a58a24d8991", "relations": [], "resource_id": "arxiv:2609.19843", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.19843v1", "topics": ["safety-security", "evaluation", "ui-design", "robustness"], "updated_at": "2026-09-17"}, "text": "A Dual-Process Perspective on Nudge Susceptibility in LLM-Based GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how automatic defaults and reflective social-influence cues affect GUI-agent shopping decisions under different reasoning configurations. A randomized simulated-shopping experiment finds that reasoning changes the type of susceptibility rather than uniformly removing interface influence.\nPlatforms: unspecified\nTopics: safety-security, evaluation, ui-design, robustness\nPublished: 2026-09-17\nAuthors: Haya Halimeh, Sascha Kaltenpoth, Kevin Bösch, Oliver Müller\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.19843v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-19843/"}
{"id": "arxiv:2609.22000", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b312c5a4d638ec9f9a7408fd9ce2d8cbf74fa462d48c8aa37f59e4ad3bd8edc", "submitted_at": "2026-09-18T17:00:56Z", "updated_at": "2026-09-21T07:09:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22000v2"}, "authors": ["Shuai Bai", "Jiayong Deng", "Sicheng Fan", "Yikun Fu", "Chang Gao", "Xuhao Hu", "Mianqiu Huang", "Yizhen Jiang", "Yuheng Jing", "Dehui Kong", "Keliang Li", "Ning Li", "Wanli Li", "Dayiheng Liu", "Dunjie Lu", "Changwei Luo", "Que Shen", "Zheyuan Wang", "Zijian Wang", "Jie Wu", "Gao Wu", "Zhihui Xie", "Rui Xie", "Haiyang Xu", "An Yang", "Jiakang Yuan", "Yanming Zhang", "Jiajun Zhang", "Xi Zhang", "Zhenru Zhang", "Zhuo Zhen", "Mingkang Zhu", "Bowen Zhou"], "canonical_source_url": "https://arxiv.org/abs/2609.22000", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-22000/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b312c5a4d638ec9f9a7408fd9ce2d8cbf74fa462d48c8aa37f59e4ad3bd8edc", "submitted_at": "2026-09-18T17:00:56Z", "updated_at": "2026-09-21T07:09:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22000v2"}}, "environments": ["desktop", "linux", "macos", "windows", "android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.22000"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22000"}], "name": "RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents", "publication": null, "published_at": "2026-09-18", "record_sha256": "4842709e61f10b737af7cc1a228eac39fead534bd3375fda6e04954eb5a50799", "relations": [], "resource_id": "arxiv:2609.22000", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.22000v2", "topics": ["evaluation", "data-generation", "tool-integration", "workflow"], "updated_at": "2026-09-21"}, "text": "RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nRecreationWorld gives hybrid agents running reference applications to explore and recreate using native GUI controls and coding tools. Its multi-platform benchmark checks interaction-dependent behavior with programmatic and visual assertions, while trajectory-training transfer is evaluated separately.\nPlatforms: desktop, linux, macos, windows, android, web\nTopics: evaluation, data-generation, tool-integration, workflow\nPublished: 2026-09-18\nAuthors: Shuai Bai, Jiayong Deng, Sicheng Fan, Yikun Fu, Chang Gao, Xuhao Hu, Mianqiu Huang, Yizhen Jiang, Yuheng Jing, Dehui Kong, Keliang Li, Ning Li, Wanli Li, Dayiheng Liu, Dunjie Lu, Changwei Luo, Que Shen, Zheyuan Wang, Zijian Wang, Jie Wu, Gao Wu, Zhihui Xie, Rui Xie, Haiyang Xu, An Yang, Jiakang Yuan, Yanming Zhang, Jiajun Zhang, Xi Zhang, Zhenru Zhang, Zhuo Zhen, Mingkang Zhu, Bowen Zhou\nRecorded arXiv version: v2\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.22000v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-22000/"}
{"id": "arxiv:2609.22083", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4dfb0703133032426ff8f40ed6a2c8131025b07658c0b643603f8d50ba163d16", "submitted_at": "2026-09-18T17:59:34Z", "updated_at": "2026-09-18T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.22083v1"}, "authors": ["Mingfei Gao", "Rui Tian", "Haiming Gang", "Bohan Zhai", "Le Zhang", "Yuanzheng Gong", "Di Feng", "Ege Özsoy", "Kaixin Ma", "Vishwesh Kirthivasan", "Oğuzhan Fatih Kar", "Roman Bachmann", "Anders Boesen Lindbo Larsen", "Afshin Dehghan"], "canonical_source_url": "https://arxiv.org/abs/2609.22083", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-22083/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4dfb0703133032426ff8f40ed6a2c8131025b07658c0b643603f8d50ba163d16", "submitted_at": "2026-09-18T17:59:34Z", "updated_at": "2026-09-18T17:59:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.22083v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.22083"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22083"}], "name": "MintAct: A Unified Visual Agent for Digital Environments", "publication": null, "published_at": "2026-09-18", "record_sha256": "12a8c5f0d9abe5a3be876e34a62a0389d4a8c72c2618c32266a6a0741f7e1790", "relations": [], "resource_id": "arxiv:2609.22083", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.22083v1", "topics": ["grounding", "reinforcement-learning", "data-generation", "tool-integration"], "updated_at": "2026-09-18"}, "text": "MintAct: A Unified Visual Agent for Digital Environments\nResource type: paper\nReview status: curated\nScope: core\n\nMintAct trains a visual-agent family for grounding, navigation and tool use across mobile, desktop and web environments. Its shared infrastructure supports trajectory collection and asynchronous online RL while controlling cross-domain sampling and off-policy drift.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning, data-generation, tool-integration\nPublished: 2026-09-18\nAuthors: Mingfei Gao, Rui Tian, Haiming Gang, Bohan Zhai, Le Zhang, Yuanzheng Gong, Di Feng, Ege Özsoy, Kaixin Ma, Vishwesh Kirthivasan, Oğuzhan Fatih Kar, Roman Bachmann, Anders Boesen Lindbo Larsen, Afshin Dehghan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.22083v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-22083/"}
{"id": "arxiv:2609.22178", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a88d287d5433009962a905dbdbf03359e1183e66e4a6015441ec28405f64a6ee", "submitted_at": "2026-08-27T05:43:50Z", "updated_at": "2026-09-22T06:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22178v2"}, "authors": ["Zeyu Kang", "Zhenyun Yin", "Yang Zhang", "Shan He", "Shanzhe Lei", "Yanjiu Zhong", "Xinquan Chen", "Xuhong Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.22178", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-22178/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a88d287d5433009962a905dbdbf03359e1183e66e4a6015441ec28405f64a6ee", "submitted_at": "2026-08-27T05:43:50Z", "updated_at": "2026-09-22T06:39:18Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.22178v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.22178"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.22178"}], "name": "Beyond Task Completion: Training Capable and Safe Computer-Use Agents", "publication": null, "published_at": "2026-08-27", "record_sha256": "6f121d1aef75ad70d695847c6ef0eca5cb40fb3ea2cae11b86d08b9765d5418d", "relations": [], "resource_id": "arxiv:2609.22178", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.22178v2", "topics": ["safety-security", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-22"}, "text": "Beyond Task Completion: Training Capable and Safe Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SCOPE, which trains computer-use agents on ordinary task demonstrations, safe continuations around hazards, and explicit refusals. Its SCOPE-Gen pipeline pairs verifiable tasks with environmental-risk variants to support joint capability and safety training.\nPlatforms: desktop\nTopics: safety-security, reinforcement-learning, data-generation\nPublished: 2026-08-27\nAuthors: Zeyu Kang, Zhenyun Yin, Yang Zhang, Shan He, Shanzhe Lei, Yanjiu Zhong, Xinquan Chen, Xuhong Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.22178v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-22178/"}
{"id": "arxiv:2609.24277", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64936af65d0ca351b19da3f63325f2bc2a5c9e0951e5a7806cd4dd82b18784a7", "submitted_at": "2026-09-21T08:42:08Z", "updated_at": "2026-09-21T08:42:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24277v1"}, "authors": ["Yunxiang Li", "Xixin Wu", "Helen Meng"], "canonical_source_url": "https://arxiv.org/abs/2609.24277", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-24277/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "64936af65d0ca351b19da3f63325f2bc2a5c9e0951e5a7806cd4dd82b18784a7", "submitted_at": "2026-09-21T08:42:08Z", "updated_at": "2026-09-21T08:42:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24277v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.24277"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.24277"}], "name": "How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation", "publication": null, "published_at": "2026-09-21", "record_sha256": "079b85c273d6c72939d168ef3c975f8e8a64064160d8c6048e62c83f3c9f5bb7", "relations": [], "resource_id": "arxiv:2609.24277", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.24277v1", "topics": ["grounding", "efficiency", "evaluation", "robustness"], "updated_at": "2026-09-21"}, "text": "How Many Pixels Is a Digit Worth? Place-Aware Coordinate Entropy for GUI Agent Confidence Estimation\nResource type: paper\nReview status: curated\nScope: core\n\nPACE weights coordinate-digit entropy by place value to estimate confidence in GUI clicks from one forward pass. It evaluates correctness ranking and selective accuracy against repeated-sampling alternatives on ScreenSpot-Pro and ScreenSpot-v2.\nPlatforms: unspecified\nTopics: grounding, efficiency, evaluation, robustness\nPublished: 2026-09-21\nAuthors: Yunxiang Li, Xixin Wu, Helen Meng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.24277v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-24277/"}
{"id": "arxiv:2609.24890", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d3cb920a677108910530788eb68c263127dba9510274f182910aed9c2f848c9", "submitted_at": "2026-09-21T16:55:24Z", "updated_at": "2026-09-21T16:55:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24890v1"}, "authors": ["Zhilin Wang", "Shaokun Zhang", "Yifan Zhang", "Hao Zhang", "Jin Xu", "Binfeng Xu", "Jian Hu", "Yunheng Zou", "Karan Sapra", "Andrew Tao", "Jan Kautz", "Yi Dong"], "canonical_source_url": "https://arxiv.org/abs/2609.24890", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-24890/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0d3cb920a677108910530788eb68c263127dba9510274f182910aed9c2f848c9", "submitted_at": "2026-09-21T16:55:24Z", "updated_at": "2026-09-21T16:55:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.24890v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.24890"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.24890"}], "name": "OSWorld-Pro: Process-based Evaluation for Computer Use Agents", "publication": null, "published_at": "2026-09-21", "record_sha256": "753d60dbc30fd615ef40146777e8f890c06589b129c38de1da3eee6f92c1e64b", "relations": [], "resource_id": "arxiv:2609.24890", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.24890v1", "topics": ["evaluation", "observability"], "updated_at": "2026-09-21"}, "text": "OSWorld-Pro: Process-based Evaluation for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAdds annotated intermediate subgoals to computer-use tasks and uses process-oriented evaluation to identify action and progress failures.\nPlatforms: desktop\nTopics: evaluation, observability\nPublished: 2026-09-21\nAuthors: Zhilin Wang, Shaokun Zhang, Yifan Zhang, Hao Zhang, Jin Xu, Binfeng Xu, Jian Hu, Yunheng Zou, Karan Sapra, Andrew Tao, Jan Kautz, Yi Dong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.24890v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-24890/"}
{"id": "arxiv:2609.25769", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d184b111aa2a9ad65dd57391b72bcef369c2bcd01ceb8e4f00f7a2320db856d8", "submitted_at": "2026-09-22T07:02:35Z", "updated_at": "2026-09-22T07:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.25769v1"}, "authors": ["Yan Zhang", "Pei Fu", "Daiqing Wu", "Huawen Shen", "Ruoceng Zhang", "Shaojie Zhang", "Jiahui Yang", "Yu Zhou", "Can Ma", "Zhenbo Luo", "Jian Luan"], "canonical_source_url": "https://arxiv.org/abs/2609.25769", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-25769/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d184b111aa2a9ad65dd57391b72bcef369c2bcd01ceb8e4f00f7a2320db856d8", "submitted_at": "2026-09-22T07:02:35Z", "updated_at": "2026-09-22T07:02:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.25769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.25769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.25769"}], "name": "Towards Omni-dimensional GUI Agent Navigation with Masked Trajectory Prediction", "publication": null, "published_at": "2026-09-22", "record_sha256": "6897fba3420ff396efc7d8aca7b0234f820d4c1485becb92e6a09a1b918dd588", "relations": [], "resource_id": "arxiv:2609.25769", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.25769v1", "topics": ["imitation-learning", "action-prediction", "state-tracking", "planning"], "updated_at": "2026-09-22"}, "text": "Towards Omni-dimensional GUI Agent Navigation with Masked Trajectory Prediction\nResource type: paper\nReview status: curated\nScope: core\n\nMaP unifies GUI-navigation objectives by masking and predicting components of multi-turn trajectories. Role-aware adapters address heterogeneous training data, and evaluations compare the approach with direct task-mixture training across navigation benchmarks.\nPlatforms: unspecified\nTopics: imitation-learning, action-prediction, state-tracking, planning\nPublished: 2026-09-22\nAuthors: Yan Zhang, Pei Fu, Daiqing Wu, Huawen Shen, Ruoceng Zhang, Shaojie Zhang, Jiahui Yang, Yu Zhou, Can Ma, Zhenbo Luo, Jian Luan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.25769v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-25769/"}
{"id": "arxiv:2609.27273", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d241c1ea9731bc815eef5df9c409d410608adb052616406d955d7d33cecd9b9", "submitted_at": "2026-09-23T03:00:00Z", "updated_at": "2026-09-27T20:49:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.27273v2"}, "authors": ["Yuxuan Li", "Will Epperson", "Wesley Deng", "Zezhou Huang"], "canonical_source_url": "https://arxiv.org/abs/2609.27273", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-27273/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8d241c1ea9731bc815eef5df9c409d410608adb052616406d955d7d33cecd9b9", "submitted_at": "2026-09-23T03:00:00Z", "updated_at": "2026-09-27T20:49:01Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.27273v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.27273"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27273"}], "name": "CAVEAT: Towards Robust Computer-Use Agents in Incentive-Misaligned Environments", "publication": null, "published_at": "2026-09-23", "record_sha256": "68b5531b09e0f3199cf67415244f78f0744895962784a0849f201c925a409685", "relations": [], "resource_id": "arxiv:2609.27273", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.27273v2", "topics": ["evaluation", "robustness", "safety-security"], "updated_at": "2026-09-27"}, "text": "CAVEAT: Towards Robust Computer-Use Agents in Incentive-Misaligned Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CAVEAT, a controlled marketplace benchmark for testing whether delegated agents preserve user objectives when interfaces favor other outcomes. It diagnoses decision failures and studies harness interventions and targeted training to improve resistance to marketplace steering.\nPlatforms: web\nTopics: evaluation, robustness, safety-security\nPublished: 2026-09-23\nAuthors: Yuxuan Li, Will Epperson, Wesley Deng, Zezhou Huang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.27273v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-27273/"}
{"id": "arxiv:2609.27307", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ae6c83c017a05c2916f31e388f6584cc2cab7b87e1eb0038c33a707bbd66e24", "submitted_at": "2026-09-23T03:38:11Z", "updated_at": "2026-09-23T03:38:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27307v1"}, "authors": ["Yan Zhang", "Daiqing Wu", "Huawen Shen", "Liang Li", "Gang Cao", "Zhi Gong", "Wei Dai", "Xiaode Zhang", "Can Ma", "Yu Zhou"], "canonical_source_url": "https://arxiv.org/abs/2609.27307", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-27307/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ae6c83c017a05c2916f31e388f6584cc2cab7b87e1eb0038c33a707bbd66e24", "submitted_at": "2026-09-23T03:38:11Z", "updated_at": "2026-09-23T03:38:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27307v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.27307"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27307"}], "name": "Learn How to Act from Your Own Interactions: On-Policy Self-Distillation for GUI Agents", "publication": null, "published_at": "2026-09-23", "record_sha256": "347bc203b3efbb94c4273c4efbba38b5944e9fb28ebfd769d5ba3cde87155245", "relations": [], "resource_id": "arxiv:2609.27307", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.27307v1", "topics": ["imitation-learning", "memory", "interaction-history", "mobile-navigation"], "updated_at": "2026-09-23"}, "text": "Learn How to Act from Your Own Interactions: On-Policy Self-Distillation for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-SD-v2 extends on-policy self-distillation from grounding to multi-turn GUI interaction. It first strengthens responses to privileged guidance, then selectively distills step-specific reasoning and memory support for later actions.\nPlatforms: android\nTopics: imitation-learning, memory, interaction-history, mobile-navigation\nPublished: 2026-09-23\nAuthors: Yan Zhang, Daiqing Wu, Huawen Shen, Liang Li, Gang Cao, Zhi Gong, Wei Dai, Xiaode Zhang, Can Ma, Yu Zhou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.27307v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-27307/"}
{"id": "arxiv:2609.27353", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b2bda1c50fd47fabbe77c40a2938979a6a16b407b40f11737fc6f1fc795d090", "submitted_at": "2026-09-23T04:49:07Z", "updated_at": "2026-09-23T04:49:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27353v1"}, "authors": ["Chengguang Gan", "Yunhao Liang", "QingHao Zhang", "Shiwen Ni"], "canonical_source_url": "https://arxiv.org/abs/2609.27353", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-27353/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b2bda1c50fd47fabbe77c40a2938979a6a16b407b40f11737fc6f1fc795d090", "submitted_at": "2026-09-23T04:49:07Z", "updated_at": "2026-09-23T04:49:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27353v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.27353"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27353"}], "name": "Guides That Cause Actions: An Offline Study of Guide-Action Mutual Reinforcement in Multimodal Web Agents", "publication": null, "published_at": "2026-09-23", "record_sha256": "30309af9e96899cbf9c108594f7fdb190377e7c4dbb9ebba69ceb06dbeefd20f", "relations": [], "resource_id": "arxiv:2609.27353", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.27353v1", "topics": ["evaluation", "action-prediction", "imitation-learning", "grounding"], "updated_at": "2026-09-23"}, "text": "Guides That Cause Actions: An Offline Study of Guide-Action Mutual Reinforcement in Multimodal Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nWebMRE pairs guidance sentences with grounded actions in an offline benchmark derived from successful WebArena traces. Controlled decoding-prefix interventions test how guide content affects action selection, while separating offline prediction metrics from live browser task success.\nPlatforms: web\nTopics: evaluation, action-prediction, imitation-learning, grounding\nPublished: 2026-09-23\nAuthors: Chengguang Gan, Yunhao Liang, QingHao Zhang, Shiwen Ni\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.27353v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-27353/"}
{"id": "arxiv:2609.27461", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a89489bb1ff3001a1d806c3eefd8a6aa19503d1dfbe0265b916766119f96ec2", "submitted_at": "2026-09-23T07:27:23Z", "updated_at": "2026-09-23T07:27:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27461v1"}, "authors": ["Zeyu Zhang", "Dingyi Rong", "Zijian Chen", "Zicheng Zhang", "Xiongkuo Min", "Guangtao Zhai"], "canonical_source_url": "https://arxiv.org/abs/2609.27461", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-27461/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3a89489bb1ff3001a1d806c3eefd8a6aa19503d1dfbe0265b916766119f96ec2", "submitted_at": "2026-09-23T07:27:23Z", "updated_at": "2026-09-23T07:27:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27461v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.27461"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27461"}], "name": "Invisible in Space, Visible in Time: Motion Vision CAPTCHA against GUI Agents", "publication": null, "published_at": "2026-09-23", "record_sha256": "686ef6e3579a89e7ba877e706f7166649e3760ad26d5356e109ed36b8b3bf234", "relations": [], "resource_id": "arxiv:2609.27461", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.27461v1", "topics": ["evaluation", "screen-understanding", "video", "safety-security"], "updated_at": "2026-09-23"}, "text": "Invisible in Space, Visible in Time: Motion Vision CAPTCHA against GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMVCAP-Bench tests live browser CAPTCHAs whose foreground meaning emerges through motion against dynamic backgrounds. A foreground-only control separates temporal camouflage difficulty from answer format and browser interaction.\nPlatforms: web\nTopics: evaluation, screen-understanding, video, safety-security\nPublished: 2026-09-23\nAuthors: Zeyu Zhang, Dingyi Rong, Zijian Chen, Zicheng Zhang, Xiongkuo Min, Guangtao Zhai\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.27461v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-27461/"}
{"id": "arxiv:2609.27770", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80f6fb973e9f8d62e515d02105d08be774fc32ed5b20a7ecee92b45d112415aa", "submitted_at": "2026-08-16T19:54:53Z", "updated_at": "2026-08-16T19:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27770v1"}, "authors": ["Zhitong Guo", "Jing Yu Koh", "Ruiyu Li"], "canonical_source_url": "https://arxiv.org/abs/2609.27770", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-27770/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "80f6fb973e9f8d62e515d02105d08be774fc32ed5b20a7ecee92b45d112415aa", "submitted_at": "2026-08-16T19:54:53Z", "updated_at": "2026-08-16T19:54:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.27770v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.27770"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.27770"}], "name": "Improving LLM-based Autonomous Web Agents with Filtering", "publication": null, "published_at": "2026-08-16", "record_sha256": "6bb46c807f4b82ec93c881fd1fd8f8661365b4873c290a4c21ef18f9651cf1a6", "relations": [], "resource_id": "arxiv:2609.27770", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.27770v1", "topics": ["web-navigation", "screen-understanding", "efficiency", "evaluation"], "updated_at": "2026-08-16"}, "text": "Improving LLM-based Autonomous Web Agents with Filtering\nResource type: paper\nReview status: curated\nScope: core\n\nTrains HTML-element relevance filters on Mind2Web trajectories and transfers them to WebArena to reduce irrelevant browser context. It also evaluates a zero-shot ColBERT retriever and documents baseline web-agent failure modes.\nPlatforms: web\nTopics: web-navigation, screen-understanding, efficiency, evaluation\nPublished: 2026-08-16\nAuthors: Zhitong Guo, Jing Yu Koh, Ruiyu Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.27770v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-27770/"}
{"id": "arxiv:2609.29892", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572ba5397fc1e4bb138febc3e7004e6e4f244e0ba8b628fb91debef4a854561f", "submitted_at": "2026-09-24T14:38:08Z", "updated_at": "2026-09-24T14:38:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.29892v1"}, "authors": ["Tingyu Qu", "Weigao Sun", "Yuecheng Liu", "Yucheng Zhao", "Yi Zhu", "Yifeng Ding", "Qiyi Wang", "Sihan Cao", "Pengkun Jiao", "Hanlei Xie", "Xiongwei Wu", "Qichao Wang", "Haodong Zhang", "Jiajun Liu", "Yuhao Wang", "Yuqing Xie", "Junpeng Zhao", "Long Chen", "Ming Ma", "Sihan Yang", "Ziwang Zhao", "Yanhao Jia", "Liangquan Gong", "Feida Zhu", "Yiran Zhong", "Steven Hoi"], "canonical_source_url": "https://arxiv.org/abs/2609.29892", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-29892/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "572ba5397fc1e4bb138febc3e7004e6e4f244e0ba8b628fb91debef4a854561f", "submitted_at": "2026-09-24T14:38:08Z", "updated_at": "2026-09-24T14:38:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.29892v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2609.29892v1"], "reason": "Record the paper-specific arXiv document declaration; it does not establish code, data or model-weight permissions."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:43:20.402537Z", "role": "primary", "sha256": "d4385231cea2965a48895eeebd2dd6000265dd7b43731a0ad88b0373a21d0308", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2609.29892v1"}], "identifiers": {"arxiv": "2609.29892"}, "kind": "paper", "last_verified_at": "2026-10-20T15:43:20.402537Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox (#license-tr) declares License: CC BY 4.0 for this paper. This declaration is document-only.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:43:20.402537Z", "url": "https://arxiv.org/html/2609.29892v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.29892"}], "name": "Qwen-Planner-Agent: A Closed-Loop AI-for-AI Framework for Real-World Mobile Planner Agents", "publication": null, "published_at": "2026-09-24", "record_sha256": "5d415ab20e6e239481269bc43aeb520214650edc863b811938afda5b15f8ce11", "relations": [], "resource_id": "arxiv:2609.29892", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.29892v1", "topics": ["planning", "data-generation", "reinforcement-learning", "memory", "mobile-navigation", "efficiency"], "updated_at": "2026-09-24"}, "text": "Qwen-Planner-Agent: A Closed-Loop AI-for-AI Framework for Real-World Mobile Planner Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains a mobile planner with supervised trajectories and competence-adaptive RL, then couples typed tool execution with memory, skills and offline model-harness refinement.\nPlatforms: unspecified\nTopics: planning, data-generation, reinforcement-learning, memory, mobile-navigation, efficiency\nPublished: 2026-09-24\nAuthors: Tingyu Qu, Weigao Sun, Yuecheng Liu, Yucheng Zhao, Yi Zhu, Yifeng Ding, Qiyi Wang, Sihan Cao, Pengkun Jiao, Hanlei Xie, Xiongwei Wu, Qichao Wang, Haodong Zhang, Jiajun Liu, Yuhao Wang, Yuqing Xie, Junpeng Zhao, Long Chen, Ming Ma, Sihan Yang, Ziwang Zhao, Yanhao Jia, Liangquan Gong, Feida Zhu, Yiran Zhong, Steven Hoi\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2609.29892v1\"], \"reason\": \"Record the paper-specific arXiv document declaration; it does not establish code, data or model-weight permissions.\"}}}\n\nPrimarily typed tool actions and structured feedback, not pixel-coordinate GUI control; backends combine programmatic sandboxes, LLM simulation and selected real-device sessions. No specific OS is established. Model/harness updates are reviewed offline, not autonomous serving-time changes. Co-evolution scores use an evolve/development set; output-cost estimates exclude input, tools, devices and harness processing. Memory comparisons differ in history access and processing cost; user authorization relies on trusted upstream services. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2609.29892v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-29892/"}
{"id": "arxiv:2609.30186", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b387c1cfe77f6e06477415944b3923189a1f0ca30be0680f01815cb65b0067bc", "submitted_at": "2026-09-24T17:30:32Z", "updated_at": "2026-09-24T17:30:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30186v1"}, "authors": ["Linghua Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.30186", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-30186/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b387c1cfe77f6e06477415944b3923189a1f0ca30be0680f01815cb65b0067bc", "submitted_at": "2026-09-24T17:30:32Z", "updated_at": "2026-09-24T17:30:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30186v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.30186"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30186"}], "name": "Jev-Mobile: Jev as an Executor for Mobile GUI Agents", "publication": null, "published_at": "2026-09-24", "record_sha256": "433adbd8f7bd6bd69bbaf4dc9af7e568bf80e8c74a453bfc545870bcd3253e67", "relations": [], "resource_id": "arxiv:2609.30186", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.30186v1", "topics": ["mobile-navigation", "efficiency", "planning"], "updated_at": "2026-09-24"}, "text": "Jev-Mobile: Jev as an Executor for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates occasional visual-language planning from frequent lightweight action selection over an accessibility-tree-based action space.\nPlatforms: android\nTopics: mobile-navigation, efficiency, planning\nPublished: 2026-09-24\nAuthors: Linghua Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.30186v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-30186/"}
{"id": "arxiv:2609.30604", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6ed206d8f8fec53c03956fffcafec6b55d94a47fc705bbc8b77c832b76419e6", "submitted_at": "2026-09-24T22:36:10Z", "updated_at": "2026-09-24T22:36:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30604v1"}, "authors": ["Alexander Gill", "Md Farhan Ishmam", "Xuyen Nguyen", "Neha Bhat", "Parker Henry DeYoung", "Fateme Hashemi Chaleshtori", "Nathan Stringham", "Kenneth Marino", "Ana Marasović"], "canonical_source_url": "https://arxiv.org/abs/2609.30604", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-30604/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d6ed206d8f8fec53c03956fffcafec6b55d94a47fc705bbc8b77c832b76419e6", "submitted_at": "2026-09-24T22:36:10Z", "updated_at": "2026-09-24T22:36:10Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30604v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.30604"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30604"}], "name": "The Hard Part Comes After Search: Benchmarking Web Agents on Synthesizing, Organizing, and Displaying Knowledge", "publication": null, "published_at": "2026-09-24", "record_sha256": "49c820c4499f391be8a154589a6e72ee31706ca55a37b22116ac3a15ee303d25", "relations": [], "resource_id": "arxiv:2609.30604", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.30604v1", "topics": ["evaluation", "web-navigation", "workflow", "long-horizon"], "updated_at": "2026-09-24"}, "text": "The Hard Part Comes After Search: Benchmarking Web Agents on Synthesizing, Organizing, and Displaying Knowledge\nResource type: paper\nReview status: curated\nScope: core\n\nKNOWS benchmarks browser workflows that retrieve and synthesize information into documents, presentations and spreadsheets. Task-specific evaluators combine deterministic checks with model judgments, distinguishing partial progress from usable end-to-end artifact completion.\nPlatforms: web\nTopics: evaluation, web-navigation, workflow, long-horizon\nPublished: 2026-09-24\nAuthors: Alexander Gill, Md Farhan Ishmam, Xuyen Nguyen, Neha Bhat, Parker Henry DeYoung, Fateme Hashemi Chaleshtori, Nathan Stringham, Kenneth Marino, Ana Marasović\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.30604v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-30604/"}
{"id": "arxiv:2609.30887", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "838635f067b706c7b36a948c8100b30bf4cacd218470aaed6970e5186fa44fb9", "submitted_at": "2026-09-25T06:50:04Z", "updated_at": "2026-09-25T06:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30887v1"}, "authors": ["Yuchen Sun", "Chenglin Cai", "Gongjie Zhang", "Tianyu Xia", "Quyu Kong", "Panrong Tong", "Zhengwen Zeng", "Long Chen", "Steven Hoi", "Chongyang Zhang", "Yue Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.30887", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-30887/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "838635f067b706c7b36a948c8100b30bf4cacd218470aaed6970e5186fa44fb9", "submitted_at": "2026-09-25T06:50:04Z", "updated_at": "2026-09-25T06:50:04Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.30887v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:30.162100+00:00", "role": "primary", "sha256": "096098b62f82867ee39dfd7cf44095edc5c6a9307351265592e485d82eadf111", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202609200000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:53:44.035673+00:00", "role": "primary", "sha256": "2b6b8147cf7b0e1fcded85f397c272bcf535ed179789e76c3ff4f15aa3fe53a7", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.30887v1"}], "identifiers": {"arxiv": "2609.30887"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:30.162100+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.30887"}], "name": "From Tapping to Hopping: Augmenting Mobile GUI Agents with App-Native Deeplinks", "publication": null, "published_at": "2026-09-25", "record_sha256": "c9b17bfbec44b2b210bd11f57cb05941ca8bc513582668b095a01a4fe2cdb8eb", "relations": [], "resource_id": "arxiv:2609.30887", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.30887v1", "topics": ["mobile-navigation", "tool-integration", "data-generation", "imitation-learning", "reinforcement-learning", "efficiency"], "updated_at": "2026-09-25"}, "text": "From Tapping to Hopping: Augmenting Mobile GUI Agents with App-Native Deeplinks\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-Hopper combines verified Android deeplinks with GUI actions, using landing-screen descriptions, hybrid trajectory training and reinforcement learning to select navigation shortcuts and recover through GUI control.\nPlatforms: android\nTopics: mobile-navigation, tool-integration, data-generation, imitation-learning, reinforcement-learning, efficiency\nPublished: 2026-09-25\nAuthors: Yuchen Sun, Chenglin Cai, Gongjie Zhang, Tianyu Xia, Quyu Kong, Panrong Tong, Zhengwen Zeng, Long Chen, Steven Hoi, Chongyang Zhang, Yue Wang\nRecorded arXiv version: v1\n\nReviewed the complete retained Atom abstract and selected exact-v1 full text. The Atom has incomplete clauses; its text/hash remain unchanged and this annotation uses the explicit methods. Evaluation is Android-only, with a selected real-app subset. Landing-Activity verification is not task completion. Model/catalog release is prospective in the paper. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2609.30887v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-30887/"}
{"id": "arxiv:2609.31364", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ce76f34abe391289643321027a33ac3cac8b5bd64eaa901f9a5fbf7c5561362", "submitted_at": "2026-09-25T14:59:14Z", "updated_at": "2026-09-25T14:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31364v1"}, "authors": ["Kiana Hooshanfar", "Amirhossein Kazerouni", "Alireza Hosseini", "Michael Brudno", "Babak Taati"], "canonical_source_url": "https://arxiv.org/abs/2609.31364", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-31364/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9ce76f34abe391289643321027a33ac3cac8b5bd64eaa901f9a5fbf7c5561362", "submitted_at": "2026-09-25T14:59:14Z", "updated_at": "2026-09-25T14:59:14Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31364v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2609.31364v1"], "reason": "Document declaration only. The paper does not establish code, model-weight or underlying dataset permissions through this license."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:01:29.156954Z", "role": "primary", "sha256": "cd4ba28778b725c63f4a7329f683b2a287db2e46ed228a0d47006fcd95bbce80", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2609.31364v1"}], "identifiers": {"arxiv": "2609.31364"}, "kind": "paper", "last_verified_at": "2026-10-20T19:01:29.156954Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox (#license-tr) declares License: CC BY 4.0 for this paper document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:01:29.156954Z", "url": "https://arxiv.org/html/2609.31364v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.31364"}], "name": "OpenVAM: Open-World Visual Attention Modeling with VLMs", "publication": null, "published_at": "2026-09-25", "record_sha256": "68d37a910c30e42ccbb7591cc9fa898dd260cfeb5e04982240c655e16fa78339", "relations": [], "resource_id": "arxiv:2609.31364", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2609.31364v1", "topics": ["screen-understanding", "ui-design", "evaluation", "data-generation"], "updated_at": "2026-09-25"}, "text": "OpenVAM: Open-World Visual Attention Modeling with VLMs\nResource type: paper\nReview status: curated\nScope: adjacent\n\nCombines dense saliency prediction with explanations of attended elements, including webpage layouts; evaluates human-attention agreement and textual rationales rather than agent actions.\nPlatforms: web\nTopics: screen-understanding, ui-design, evaluation, data-generation\nPublished: 2026-09-25\nAuthors: Kiana Hooshanfar, Amirhossein Kazerouni, Alireza Hosseini, Michael Brudno, Babak Taati\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2609.31364v1\"], \"reason\": \"Document declaration only. The paper does not establish code, model-weight or underlying dataset permissions through this license.\"}}}\n\nStatic human-attention modeling on U-EYE webpage screenshots, with FIWI held out from training; no live-browser, agent-action or task-success evaluation. Gemini-generated reference rationales are expert-checked, and JScore compares generated text with references; human validation covers a separate 100-image subset. Reference text and heatmap alignment do not establish causal explanations of gaze or temporal scanpaths. Several baseline scores come from prior papers. Code/data/weight releases and licenses were not verified; no independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2609.31364v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-31364/"}
{"id": "arxiv:2609.31957", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa2bcfa05fcbd6937af6ba3d23e892e5a105248939538dcdc9d7b015daf38593", "submitted_at": "2026-09-25T20:04:24Z", "updated_at": "2026-09-25T20:04:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31957v1"}, "authors": ["Zhenhao Zhang", "Zhaoyu Fan", "Haohan Ying", "Jingwen Hu", "Hancen Fan", "Junhao Zhou", "Zitian Chen", "Linchao Zhu"], "canonical_source_url": "https://arxiv.org/abs/2609.31957", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-31957/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fa2bcfa05fcbd6937af6ba3d23e892e5a105248939538dcdc9d7b015daf38593", "submitted_at": "2026-09-25T20:04:24Z", "updated_at": "2026-09-25T20:04:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.31957v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.31957"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.31957"}], "name": "CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs", "publication": null, "published_at": "2026-09-25", "record_sha256": "7b1602406a69e319c806f0d5b7c2c8656112f48bb005b97e283351fe837170a2", "relations": [], "resource_id": "arxiv:2609.31957", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.31957v1", "topics": ["data-generation", "grounding", "reinforcement-learning"], "updated_at": "2026-09-25"}, "text": "CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CaptchaArena, a dataset of execution-verified interactive CAPTCHA solutions with screenshot–action trajectories, reasoning annotations, and pixel masks for irregular targets. The work trains a single policy across CAPTCHA types using supervised learning followed by verifier-reward reinforcement learning.\nPlatforms: unspecified\nTopics: data-generation, grounding, reinforcement-learning\nPublished: 2026-09-25\nAuthors: Zhenhao Zhang, Zhaoyu Fan, Haohan Ying, Jingwen Hu, Hancen Fan, Junhao Zhou, Zitian Chen, Linchao Zhu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.31957v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-31957/"}
{"id": "arxiv:2609.32036", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12976e58eb48d9c4709fb03508ac7ed7eeafffd3df675d3dccfd09bdc354f3db", "submitted_at": "2026-09-25T22:07:29Z", "updated_at": "2026-09-25T22:07:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32036v1"}, "authors": ["Chenyue Li", "Xiaoxiao Sun", "Yubo Deng", "Qinlin Zhao", "Serena Yeung-Levy", "Yuhui Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.32036", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32036/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "12976e58eb48d9c4709fb03508ac7ed7eeafffd3df675d3dccfd09bdc354f3db", "submitted_at": "2026-09-25T22:07:29Z", "updated_at": "2026-09-25T22:07:29Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32036v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32036"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32036"}], "name": "ScreenHaystack: Finding Blind Zones in GUI Grounding", "publication": null, "published_at": "2026-09-25", "record_sha256": "79889a3b620718d9a7d95bcb7a054d803e18ec1ee5ab0ea318eac80bc00a2129", "relations": [], "resource_id": "arxiv:2609.32036", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32036v1", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-09-25"}, "text": "ScreenHaystack: Finding Blind Zones in GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nMoves controlled target icons across high-resolution interfaces to diagnose spatial blind zones in grounding models and study targeted augmentation.\nPlatforms: cross-platform\nTopics: grounding, evaluation, robustness\nPublished: 2026-09-25\nAuthors: Chenyue Li, Xiaoxiao Sun, Yubo Deng, Qinlin Zhao, Serena Yeung-Levy, Yuhui Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32036v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32036/"}
{"id": "arxiv:2609.32130", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "719385c97b9d174bbebc34ac4a196971af61521489f7bf032a7d807ae3c8d4c1", "submitted_at": "2026-09-26T01:13:44Z", "updated_at": "2026-09-26T01:13:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32130v1"}, "authors": ["Beining Wu", "Jun Huang", "Yanxiao Zhao"], "canonical_source_url": "https://arxiv.org/abs/2609.32130", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32130/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "719385c97b9d174bbebc34ac4a196971af61521489f7bf032a7d807ae3c8d4c1", "submitted_at": "2026-09-26T01:13:44Z", "updated_at": "2026-09-26T01:13:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32130v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32130"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32130"}], "name": "REBASE: Device-Cloud Experience Coherence for GUI Agents Across App Updates", "publication": null, "published_at": "2026-09-26", "record_sha256": "eef9f73b3931e035f37e6750b3d718fcbc70403d4084b006bf6ef07dfbdf1670", "relations": [], "resource_id": "arxiv:2609.32130", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32130v1", "topics": ["memory", "robustness", "efficiency", "mobile-navigation"], "updated_at": "2026-09-26"}, "text": "REBASE: Device-Cloud Experience Coherence for GUI Agents Across App Updates\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces REBASE, a device–cloud protocol for repairing cached GUI experience after app updates. It combines cloud replay, device-side step verification, and progressively richer failure reports to produce repairs tied to application versions.\nPlatforms: unspecified\nTopics: memory, robustness, efficiency, mobile-navigation\nPublished: 2026-09-26\nAuthors: Beining Wu, Jun Huang, Yanxiao Zhao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32130v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32130/"}
{"id": "arxiv:2609.32166", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f343a941e12c0fb8c4e12dc156e7c8b37f753d74836d8e42542d504bee883e51", "submitted_at": "2026-09-26T02:41:56Z", "updated_at": "2026-09-26T02:41:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32166v1"}, "authors": ["Taehwan Park", "Changmin Lee", "Hayeon Lee", "Taesik Gong"], "canonical_source_url": "https://arxiv.org/abs/2609.32166", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32166/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f343a941e12c0fb8c4e12dc156e7c8b37f753d74836d8e42542d504bee883e51", "submitted_at": "2026-09-26T02:41:56Z", "updated_at": "2026-09-26T02:41:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32166v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32166"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32166"}], "name": "PastForward: Faster On-Device GUI Agents via Computational Experience Reuse", "publication": null, "published_at": "2026-09-26", "record_sha256": "52583119fac08b80d9ea989057452e8a37a390a5f409b8a3cb3ef5123ee01be6", "relations": [], "resource_id": "arxiv:2609.32166", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32166v1", "topics": ["memory", "efficiency", "mobile-navigation"], "updated_at": "2026-09-26"}, "text": "PastForward: Faster On-Device GUI Agents via Computational Experience Reuse\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PastForward, which reuses earlier output sequences as verified decoding proposals and starts future action computation from predicted GUI transitions. Screen checks determine whether speculative computation and cached model states can be retained.\nPlatforms: android\nTopics: memory, efficiency, mobile-navigation\nPublished: 2026-09-26\nAuthors: Taehwan Park, Changmin Lee, Hayeon Lee, Taesik Gong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32166v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32166/"}
{"id": "arxiv:2609.32351", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc6fd87d2299e2881eab9a2db96e4022663e2d94da5549dcde166c0c8b8be11", "submitted_at": "2026-09-26T08:14:50Z", "updated_at": "2026-09-26T08:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32351v1"}, "authors": ["Yangzhe Peng", "Xiaoyang Wang", "Yiyang Zhao", "Lijun Wu", "Kun He"], "canonical_source_url": "https://arxiv.org/abs/2609.32351", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32351/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7cc6fd87d2299e2881eab9a2db96e4022663e2d94da5549dcde166c0c8b8be11", "submitted_at": "2026-09-26T08:14:50Z", "updated_at": "2026-09-26T08:14:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32351v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.32351"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32351"}], "name": "From Trajectories to Grounded Preferences: Process Preference Synthesis via Interaction Element Graphs for Web PRMs", "publication": null, "published_at": "2026-09-26", "record_sha256": "ae1b6596d2bb9d29a4b2531a9849574efec0c3ba2bd2aa8b3252a509dd523691", "relations": [], "resource_id": "arxiv:2609.32351", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32351v1", "topics": ["data-generation", "reward-models", "web-navigation"], "updated_at": "2026-09-26"}, "text": "From Trajectories to Grounded Preferences: Process Preference Synthesis via Interaction Element Graphs for Web PRMs\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SURFPRM, which turns web demonstrations into an interaction-element graph for synthesizing grounded negative actions. Its contrastive preference data trains process reward models to distinguish plausible on-page alternatives and guide downstream trajectory search.\nPlatforms: web\nTopics: data-generation, reward-models, web-navigation\nPublished: 2026-09-26\nAuthors: Yangzhe Peng, Xiaoyang Wang, Yiyang Zhao, Lijun Wu, Kun He\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32351v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32351/"}
{"id": "arxiv:2609.32385", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f091cab38c932e08d21afa936dfe241c93c0b2139aba1d1ce7a6d1513ad5da1a", "submitted_at": "2026-09-26T08:58:50Z", "updated_at": "2026-09-26T08:58:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32385v1"}, "authors": ["Chuhan Zhang", "Qi Xie", "Ziyue Wang", "Jianing Yin", "Yunfan Zhou", "Dazhen Deng", "Yingcai Wu"], "canonical_source_url": "https://arxiv.org/abs/2609.32385", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32385/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f091cab38c932e08d21afa936dfe241c93c0b2139aba1d1ce7a6d1513ad5da1a", "submitted_at": "2026-09-26T08:58:50Z", "updated_at": "2026-09-26T08:58:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32385v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32385"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32385"}], "name": "DashAct: A Progressive Diagnostic Benchmark for GUI Agents in Interactive Dashboard Analysis", "publication": null, "published_at": "2026-09-26", "record_sha256": "b0c0454eb32c7fc6cfa98b0b3ae1fa235f115014f30dd6b0a2262bd3cbc64ee9", "relations": [], "resource_id": "arxiv:2609.32385", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32385v1", "topics": ["evaluation", "grounding", "planning"], "updated_at": "2026-09-26"}, "text": "DashAct: A Progressive Diagnostic Benchmark for GUI Agents in Interactive Dashboard Analysis\nResource type: paper\nReview status: curated\nScope: core\n\nDiagnoses dashboard-agent failures by progressively restoring interaction context, next-action support, and visual target information within the same task.\nPlatforms: unspecified\nTopics: evaluation, grounding, planning\nPublished: 2026-09-26\nAuthors: Chuhan Zhang, Qi Xie, Ziyue Wang, Jianing Yin, Yunfan Zhou, Dazhen Deng, Yingcai Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32385v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32385/"}
{"id": "arxiv:2609.32511", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff1c4eff95094ea8ecf37b7bc621eebbd5e4a677b0e11240e802a72b9b229c7f", "submitted_at": "2026-09-26T11:59:39Z", "updated_at": "2026-09-26T11:59:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32511v1"}, "authors": ["Jinming Hu", "Haodong Zhao", "Qi Jia", "Die Chen", "Tianhang Zhao", "Sufeng Duan", "Gongshen Liu"], "canonical_source_url": "https://arxiv.org/abs/2609.32511", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32511/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff1c4eff95094ea8ecf37b7bc621eebbd5e4a677b0e11240e802a72b9b229c7f", "submitted_at": "2026-09-26T11:59:39Z", "updated_at": "2026-09-26T11:59:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32511v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56Z", "role": "primary", "sha256": "ce7210dccecb2896202639e6f7be41876fee20de1135b0d23278cead2d695339", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.32511v1"}], "identifiers": {"arxiv": "2609.32511"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32511"}], "name": "Learning from Others, Acting for You: Cross-User Memory Sharing for LLM Agents", "publication": null, "published_at": "2026-09-26", "record_sha256": "2e3a14a408cbfc7e4a6fa172a01b9fd4e8edd5b24ba2be4496c6a9c5e2fe2ff9", "relations": [], "resource_id": "arxiv:2609.32511", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32511v1", "topics": ["memory", "web-navigation", "action-prediction", "evaluation"], "updated_at": "2026-09-26"}, "text": "Learning from Others, Acting for You: Cross-User Memory Sharing for LLM Agents\nResource type: paper\nReview status: curated\nScope: core\n\nShareMem consolidates reusable experience across users while keeping preference retrieval user-bound; its web evaluation predicts target elements and actions on recorded Mind2Web states.\nPlatforms: web\nTopics: memory, web-navigation, action-prediction, evaluation\nPublished: 2026-09-26\nAuthors: Jinming Hu, Haodong Zhao, Qi Jia, Die Chen, Tianhang Zhao, Sufeng Duan, Gongshen Liu\nRecorded arXiv version: v1\n\nMain Mind2Web evaluation uses 252 recorded tasks assigned to 20 simulated users, frozen demonstration-derived memory and task-macro step metrics; it is not a live-browser completion test. The main split withholds same-site local experience, and sharing gains shrink when local coverage rises. Mind2Web has no preference channel; preference claims come from separate service and coding settings. User-bound retrieval is not a demonstrated privacy or security guarantee. Selected sections were source-read; no independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2609.32511v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32511/"}
{"id": "arxiv:2609.32600", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d53f2512b99828849f047f63412264437a59b730c117c801de252e6dc80a33fc", "submitted_at": "2026-09-26T13:18:00Z", "updated_at": "2026-09-26T13:18:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32600v1"}, "authors": ["Prince Zizhuang Wang", "Chenhao Liang", "Zelong Xu", "Aojie Yuan", "Xiaolin Zhou", "Haiyue Zhang", "Yue Zhao", "Xiyang Hu", "Shuli Jiang"], "canonical_source_url": "https://arxiv.org/abs/2609.32600", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32600/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d53f2512b99828849f047f63412264437a59b730c117c801de252e6dc80a33fc", "submitted_at": "2026-09-26T13:18:00Z", "updated_at": "2026-09-26T13:18:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32600v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32600"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32600"}], "name": "CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering", "publication": null, "published_at": "2026-09-26", "record_sha256": "420ed82aff008e5b07eed6ad8fb3e74116e98eed669a48a617a19ad124923a67", "relations": [], "resource_id": "arxiv:2609.32600", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32600v1", "topics": ["evaluation", "workflow", "tool-integration"], "updated_at": "2026-09-26"}, "text": "CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CUA-SWE, a benchmark and execution environment that combines code changes with graphical interaction and visual inspection of running software. Task-specific tests verify repairs and changes whose required information may be available only through the application interface.\nPlatforms: unspecified\nTopics: evaluation, workflow, tool-integration\nPublished: 2026-09-26\nAuthors: Prince Zizhuang Wang, Chenhao Liang, Zelong Xu, Aojie Yuan, Xiaolin Zhou, Haiyue Zhang, Yue Zhao, Xiyang Hu, Shuli Jiang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32600v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32600/"}
{"id": "arxiv:2609.32679", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d641819d168eedc07270607da33f3197c15f2f90cf99f538c3839eed506f3308", "submitted_at": "2026-09-26T14:35:45Z", "updated_at": "2026-09-26T14:35:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32679v1"}, "authors": ["Dongsheng Liu", "Chao Jin", "Wenkui Yang", "Hejin Wang", "Junwei Yang", "Zeren Zhang", "Ziwei Chen", "Huaibo Huang", "Jie Cao", "Ran He"], "canonical_source_url": "https://arxiv.org/abs/2609.32679", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32679/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d641819d168eedc07270607da33f3197c15f2f90cf99f538c3839eed506f3308", "submitted_at": "2026-09-26T14:35:45Z", "updated_at": "2026-09-26T14:35:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32679v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.32679"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32679"}], "name": "The GUI Is Not the State: Diagnosing State Aliasing in GUI World Models", "publication": null, "published_at": "2026-09-26", "record_sha256": "c4cf315e41a87e6a18497ded8b3567b53e166bec8579f824a4a560454f1a019e", "relations": [], "resource_id": "arxiv:2609.32679", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32679v1", "topics": ["world-models", "state-tracking", "evaluation"], "updated_at": "2026-09-26"}, "text": "The GUI Is Not the State: Diagnosing State Aliasing in GUI World Models\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces StateAliasBench to isolate GUI world-model ambiguities in which matching visible screens and actions can lead to different futures. It augments frozen world models with structured state recovered from interaction history and evaluates the effect on prediction and agent execution.\nPlatforms: android\nTopics: world-models, state-tracking, evaluation\nPublished: 2026-09-26\nAuthors: Dongsheng Liu, Chao Jin, Wenkui Yang, Hejin Wang, Junwei Yang, Zeren Zhang, Ziwei Chen, Huaibo Huang, Jie Cao, Ran He\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32679v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32679/"}
{"id": "arxiv:2609.32750", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4083bbe61c846d93663cd41d6335a2d6a54e361df9c3bdcd636b312c436d1a4", "submitted_at": "2026-09-26T16:16:08Z", "updated_at": "2026-09-26T16:16:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32750v1"}, "authors": ["Xin Yan", "Zhengbo Jiao", "Jiaqi Liu", "Zhenglin Wan", "SiYuan Ma", "Xuliang Yu", "Tianyi Jiang", "Chubin Zhang", "Pengfei Zhou", "Wangbo Zhao", "Xingrui Yu", "Bo An", "Yang You", "Ivor Tsang"], "canonical_source_url": "https://arxiv.org/abs/2609.32750", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32750/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d4083bbe61c846d93663cd41d6335a2d6a54e361df9c3bdcd636b312c436d1a4", "submitted_at": "2026-09-26T16:16:08Z", "updated_at": "2026-09-26T16:16:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32750v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.32750"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32750"}], "name": "CUA-Sandbox: Efficient Environments for Computer-Use Agent Reinforcement Learning", "publication": null, "published_at": "2026-09-26", "record_sha256": "702367857dc57c56f4021e8f580f173c1d481d7fd4edf3a434e7e08ea83c263b", "relations": [], "resource_id": "arxiv:2609.32750", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32750v1", "topics": ["sandbox", "reinforcement-learning", "efficiency"], "updated_at": "2026-09-26"}, "text": "CUA-Sandbox: Efficient Environments for Computer-Use Agent Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CUA-Sandbox, which separates each rollout’s mutable state from shared initialized software runtimes. State-scoped execution and transactional resets or branches aim to reduce training-environment costs while retaining original interfaces and evaluators.\nPlatforms: web, desktop\nTopics: sandbox, reinforcement-learning, efficiency\nPublished: 2026-09-26\nAuthors: Xin Yan, Zhengbo Jiao, Jiaqi Liu, Zhenglin Wan, SiYuan Ma, Xuliang Yu, Tianyi Jiang, Chubin Zhang, Pengfei Zhou, Wangbo Zhao, Xingrui Yu, Bo An, Yang You, Ivor Tsang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32750v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32750/"}
{"id": "arxiv:2609.32915", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "802342ce2c486c30c84c33c897becb0348831034d4d11f8d3d3861ec0380b65c", "submitted_at": "2026-09-26T20:09:16Z", "updated_at": "2026-09-26T20:09:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32915v1"}, "authors": ["Asif Shahriar", "Md Nafiu Rahman", "Sadif Ahmed", "Farig Sadeque", "Md Rizwan Parvez"], "canonical_source_url": "https://arxiv.org/abs/2609.32915", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-32915/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "802342ce2c486c30c84c33c897becb0348831034d4d11f8d3d3861ec0380b65c", "submitted_at": "2026-09-26T20:09:16Z", "updated_at": "2026-09-26T20:09:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.32915v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.32915"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.32915"}], "name": "AgentTell: Behavioural Side-Channel Leakage in Browser-Use Agents", "publication": null, "published_at": "2026-09-26", "record_sha256": "a75aeffb173152ef8df7f385dd841cb19824d83b3e232d4f7992582480fe442c", "relations": [], "resource_id": "arxiv:2609.32915", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.32915v1", "topics": ["safety-security", "evaluation", "memory"], "updated_at": "2026-09-26"}, "text": "AgentTell: Behavioural Side-Channel Leakage in Browser-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AgentTell, a benchmark for privacy leakage through browser actions after an agent learns a private fact on another website. It examines whether agents choose revealing options despite nondisclosure instructions and whether their final reports accurately describe that behavior.\nPlatforms: web\nTopics: safety-security, evaluation, memory\nPublished: 2026-09-26\nAuthors: Asif Shahriar, Md Nafiu Rahman, Sadif Ahmed, Farig Sadeque, Md Rizwan Parvez\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.32915v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-32915/"}
{"id": "arxiv:2609.33102", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c380079220050fbe354509238a5a79c7df2917a245490e4ec0cff512f7565ab", "submitted_at": "2026-09-27T02:20:08Z", "updated_at": "2026-09-27T02:20:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33102v1"}, "authors": ["Ben Hagag", "William L. Anderson", "Srija Chakraborty", "Christian Schroeder de Witt"], "canonical_source_url": "https://arxiv.org/abs/2609.33102", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-33102/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3c380079220050fbe354509238a5a79c7df2917a245490e4ec0cff512f7565ab", "submitted_at": "2026-09-27T02:20:08Z", "updated_at": "2026-09-27T02:20:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33102v1"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:28.833839+00:00", "role": "primary", "sha256": "3ea667f4187ba601011e5ff00c4d63676b6d9903d12b80a0d7d5dd9920f00c65", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2609.33102v1"}], "identifiers": {"arxiv": "2609.33102"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:28.833839+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33102"}], "name": "ORBIT: A Framework for Multi-Agent Safety and Security Evaluations", "publication": null, "published_at": "2026-09-27", "record_sha256": "dcc880847fb86530bb1a47ba0a3cc0b7d35fef5ac2601c4688721d8b3624cb0d", "relations": [], "resource_id": "arxiv:2609.33102", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.33102v1", "topics": ["evaluation", "safety-security", "web-navigation"], "updated_at": "2026-09-27"}, "text": "ORBIT: A Framework for Multi-Agent Safety and Security Evaluations\nResource type: paper\nReview status: curated\nScope: core\n\nInspect-based framework that varies multi-agent roles, communication, attacks and defenses, with BrowserGym and virtual-desktop GUI adapters alongside coding and other tool scenarios.\nPlatforms: web, desktop\nTopics: evaluation, safety-security, web-navigation\nPublished: 2026-09-27\nAuthors: Ben Hagag, William L. Anderson, Srija Chakraborty, Christian Schroeder de Witt\nRecorded arXiv version: v1\n\nBrowserART uses an accessibility-tree action interface, while the desktop adapter uses mouse, keyboard and screenshots. The adapted tasks are not claimed to preserve the constructs of their source benchmarks. Experiments cover a representative configuration slice with fixed attack templates; the OS-Harm results are imported from prior work using a different model and stack. Browser defense tables also differ in baseline versus defended judges and in judged benign denominators. Human judge validation used one grader. These results do not establish a universally effective defense. The linked repository and its code/data licenses were not independently inspected, and no experiment was run.\n\nPrimary source: https://arxiv.org/abs/2609.33102v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-33102/"}
{"id": "arxiv:2609.33304", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "563661718e6e64d62f0601bbc5c214828219b9495a729ae685ba1ae9599332f3", "submitted_at": "2026-09-27T07:15:57Z", "updated_at": "2026-09-27T07:15:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33304v1"}, "authors": ["Fuyao Zhang", "Xuan Wang", "Zherui Li", "Jiaming Zhang", "Longtao Huang", "Wei Yang Bryan Lim"], "canonical_source_url": "https://arxiv.org/abs/2609.33304", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-33304/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "563661718e6e64d62f0601bbc5c214828219b9495a729ae685ba1ae9599332f3", "submitted_at": "2026-09-27T07:15:57Z", "updated_at": "2026-09-27T07:15:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33304v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.33304"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33304"}], "name": "Relevance Does Not Imply Applicability: Experience Activation for Personal GUI Agents", "publication": null, "published_at": "2026-09-27", "record_sha256": "e0c5ab429fc9cdb219d85115566ba849a9cfe63da0b80c3fff537358c3773156", "relations": [], "resource_id": "arxiv:2609.33304", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.33304v1", "topics": ["memory", "planning", "training-free"], "updated_at": "2026-09-27"}, "text": "Relevance Does Not Imply Applicability: Experience Activation for Personal GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ExpActivator, which matches the current GUI state to relevant historical states before supplying past actions as guidance. A separate activation rule decides whether a recurring user intent is warranted by the current time and situation or should be withheld.\nPlatforms: unspecified\nTopics: memory, planning, training-free\nPublished: 2026-09-27\nAuthors: Fuyao Zhang, Xuan Wang, Zherui Li, Jiaming Zhang, Longtao Huang, Wei Yang Bryan Lim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.33304v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-33304/"}
{"id": "arxiv:2609.33646", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88daacf5fda62f59a08b504a633ad984ee1d03791bba24e89c223c0686adf34e", "submitted_at": "2026-09-27T15:11:30Z", "updated_at": "2026-09-27T15:11:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33646v1"}, "authors": ["Keliang Li", "Heng Wang", "Chen Hu", "Daxin Jiang", "Hong Chang", "Shiguang Shan"], "canonical_source_url": "https://arxiv.org/abs/2609.33646", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-33646/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "88daacf5fda62f59a08b504a633ad984ee1d03791bba24e89c223c0686adf34e", "submitted_at": "2026-09-27T15:11:30Z", "updated_at": "2026-09-27T15:11:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.33646v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:17Z", "role": "primary", "sha256": "5d7c9ff24833b254d33a41a7c4247e830e4d9c371881a1e16151bad5c71cc8b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.10297%2C2609.10451%2C2609.12394%2C2609.12404%2C2609.13287%2C2609.13491%2C2609.13548%2C2609.13873%2C2609.14239%2C2609.15134%2C2609.15307%2C2609.15364%2C2609.15457%2C2609.16251%2C2609.16635%2C2609.16732%2C2609.16900%2C2609.17653%2C2609.17885%2C2609.18411%2C2609.18690%2C2609.19125%2C2609.19523%2C2609.19843%2C2609.22083%2C2609.22724%2C2609.24277%2C2609.24890%2C2609.25769%2C2609.27273%2C2609.27307%2C2609.27353%2C2609.27461%2C2609.29892%2C2609.30186%2C2609.30604%2C2609.30887%2C2609.31957%2C2609.32036%2C2609.32130%2C2609.32166%2C2609.32351%2C2609.32385%2C2609.32600%2C2609.32679%2C2609.32750%2C2609.32915%2C2609.33304%2C2609.33646%2C2609.34113&max_results=50"}], "identifiers": {"arxiv": "2609.33646"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.33646"}], "name": "Probe to Act: Elevating Browser-Use Agent via Active Visual Probing", "publication": null, "published_at": "2026-09-27", "record_sha256": "8baa0ce9b0142acf4e625374bec802b646887ae6833d83446f70cb563b805b1d", "relations": [], "resource_id": "arxiv:2609.33646", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.33646v1", "topics": ["grounding", "memory", "web-navigation"], "updated_at": "2026-09-27"}, "text": "Probe to Act: Elevating Browser-Use Agent via Active Visual Probing\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Probe to Act, which lets browser agents request visual evidence linking DOM handles and screen regions before committing an operation. It retains selected interaction evidence as memory and studies both prompted use and distillation into open models.\nPlatforms: web\nTopics: grounding, memory, web-navigation\nPublished: 2026-09-27\nAuthors: Keliang Li, Heng Wang, Chen Hu, Daxin Jiang, Hong Chang, Shiguang Shan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.33646v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-33646/"}
{"id": "arxiv:2609.34113", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f4116455785579ece9a2ff1bae304b62f1cbe13a91db8d146bd281a27d19eb0", "submitted_at": "2026-09-28T01:48:48Z", "updated_at": "2026-09-29T09:05:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.34113v2"}, "authors": ["Shaoqing Zhang", "Kehai Chen", "Xuefeng Bai", "Zhuosheng Zhang", "Pengfei Zhang", "Yang Xiang", "Min Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.34113", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-34113/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f4116455785579ece9a2ff1bae304b62f1cbe13a91db8d146bd281a27d19eb0", "submitted_at": "2026-09-28T01:48:48Z", "updated_at": "2026-09-29T09:05:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.34113v2"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.34113"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34113"}], "name": "GUITAR: Structured Failure Diagnosis of GUI Agents via State Transitions", "publication": null, "published_at": "2026-09-28", "record_sha256": "284b1f45bdd8cbe95e60ea9200f7c737b7764a9fb258c2f08424393e4498f058", "relations": [], "resource_id": "arxiv:2609.34113", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.34113v2", "topics": ["evaluation", "observability", "robustness"], "updated_at": "2026-09-29"}, "text": "GUITAR: Structured Failure Diagnosis of GUI Agents via State Transitions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GUITAR, which groups visually different screens into functional states and analyzes agent failures over a state-transition graph. It uses the resulting bottlenecks to diagnose recurring mobile and web errors and evaluate targeted guidance.\nPlatforms: android, web\nTopics: evaluation, observability, robustness\nPublished: 2026-09-28\nAuthors: Shaoqing Zhang, Kehai Chen, Xuefeng Bai, Zhuosheng Zhang, Pengfei Zhang, Yang Xiang, Min Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.34113v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-34113/"}
{"id": "arxiv:2609.34139", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ee7ddc170afcd2fa09d1c8d12a07bbd6aac692bc12ffc27bd39f1de0c912dfa", "submitted_at": "2026-09-28T02:21:13Z", "updated_at": "2026-09-28T02:21:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34139v1"}, "authors": ["Tien Tran", "Namho Koh", "Daiki E. Matsunaga", "Ayush Jain", "Kee Eung Kim"], "canonical_source_url": "https://arxiv.org/abs/2609.34139", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-34139/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0ee7ddc170afcd2fa09d1c8d12a07bbd6aac692bc12ffc27bd39f1de0c912dfa", "submitted_at": "2026-09-28T02:21:13Z", "updated_at": "2026-09-28T02:21:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34139v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.34139"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34139"}], "name": "Same Tasks, Different Apps: Why Mobile GUI Agents Fail to Generalize?", "publication": null, "published_at": "2026-09-28", "record_sha256": "74a85e3105d9edbd5e96b27f44f6db8916c7e8894ef758fd375e06b684f082a3", "relations": [], "resource_id": "arxiv:2609.34139", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.34139v1", "topics": ["evaluation", "robustness", "mobile-navigation"], "updated_at": "2026-09-28"}, "text": "Same Tasks, Different Apps: Why Mobile GUI Agents Fail to Generalize?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AnyAppBench, a live Android benchmark that holds user goals fixed while changing the application used to achieve them. It combines category-controlled task pairs with failure annotations to distinguish task understanding from application-specific execution skill.\nPlatforms: android\nTopics: evaluation, robustness, mobile-navigation\nPublished: 2026-09-28\nAuthors: Tien Tran, Namho Koh, Daiki E. Matsunaga, Ayush Jain, Kee Eung Kim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.34139v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-34139/"}
{"id": "arxiv:2609.34321", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fbaeb0b4f891d64cc12a3d5663b6804df6c7db82e6a7ccfd54aa10e837010de", "submitted_at": "2026-09-28T05:00:45Z", "updated_at": "2026-09-28T05:00:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34321v1"}, "authors": ["Ziqiang Wang", "Li Gu", "Zhixiang Chi", "Linlian Jiang", "Zihuan Jiang", "Linqiang Guo", "Siobhan Reid", "Zhi Liu", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.34321", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-34321/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7fbaeb0b4f891d64cc12a3d5663b6804df6c7db82e6a7ccfd54aa10e837010de", "submitted_at": "2026-09-28T05:00:45Z", "updated_at": "2026-09-28T05:00:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34321v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.34321"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34321"}], "name": "One Rollout Is All You Get: Fully Test-Time Adaptation for GUI Agents", "publication": null, "published_at": "2026-09-28", "record_sha256": "5e2f753648e27e60b9ffd97d8794c327e236db963c454e7e73a4bb24d8a6a8d3", "relations": [], "resource_id": "arxiv:2609.34321", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.34321v1", "topics": ["memory", "evaluation", "robustness"], "updated_at": "2026-09-28"}, "text": "One Rollout Is All You Get: Fully Test-Time Adaptation for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nDefines a test-time adaptation setting with one counted attempt per arriving GUI task and no ground-truth feedback. Its SOLO method filters or relabels completed episodes, then updates a small adapter from a short window of the agent’s own experience.\nPlatforms: web\nTopics: memory, evaluation, robustness\nPublished: 2026-09-28\nAuthors: Ziqiang Wang, Li Gu, Zhixiang Chi, Linlian Jiang, Zihuan Jiang, Linqiang Guo, Siobhan Reid, Zhi Liu, Yang Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.34321v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-34321/"}
{"id": "arxiv:2609.34769", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f81fd07262dccf6783729e4336fbd77835f2fcdea030eef8dc2272e7ef82b8fc", "submitted_at": "2026-09-28T09:49:25Z", "updated_at": "2026-09-28T09:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34769v1"}, "authors": ["Bingo Zhang", "Haochuan Lu", "Zongjie Li", "Genjian Li", "Ari Yu Zhang", "Chaozheng Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.34769", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-34769/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f81fd07262dccf6783729e4336fbd77835f2fcdea030eef8dc2272e7ef82b8fc", "submitted_at": "2026-09-28T09:49:25Z", "updated_at": "2026-09-28T09:49:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34769v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.34769"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34769"}], "name": "LongPuzzleBench: Evaluating GUI Agents on Long-Horizon Visual Puzzles", "publication": null, "published_at": "2026-09-28", "record_sha256": "94882267abeb31971223619f3b834146fc9ff501a90772a1483a0a1315695e3e", "relations": [], "resource_id": "arxiv:2609.34769", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2609.34769v1", "topics": ["evaluation", "planning", "long-horizon", "state-tracking"], "updated_at": "2026-09-28"}, "text": "LongPuzzleBench: Evaluating GUI Agents on Long-Horizon Visual Puzzles\nResource type: paper\nReview status: curated\nScope: adjacent\n\nLongPuzzleBench evaluates coherent long-horizon planning on persistent visual puzzle boards through native GUI actions. It compares GUI-only and code-assisted agents and uses controlled hints and memory diagnostics to examine decisions that preserve or eliminate future solutions.\nPlatforms: unspecified\nTopics: evaluation, planning, long-horizon, state-tracking\nPublished: 2026-09-28\nAuthors: Bingo Zhang, Haochuan Lu, Zongjie Li, Genjian Li, Ari Yu Zhang, Chaozheng Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the complete retained original arXiv abstract. Experiments, reported performance and artifact availability have not been independently reproduced. Agents execute real GUI moves in six puzzle games; the tasks diagnose computer-control planning but remain game-based rather than general productivity workflows.\n\nPrimary source: https://arxiv.org/abs/2609.34769v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-34769/"}
{"id": "arxiv:2609.34974", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9bc3b57957e6c4f1af399379c4275366563b94a8b8fe01ec83e54715cfc8d11", "submitted_at": "2026-09-28T11:53:24Z", "updated_at": "2026-09-28T11:53:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34974v1"}, "authors": ["Ruozhao Yang", "Mingfei Cheng", "Xiaofei Xie"], "canonical_source_url": "https://arxiv.org/abs/2609.34974", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-34974/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a9bc3b57957e6c4f1af399379c4275366563b94a8b8fe01ec83e54715cfc8d11", "submitted_at": "2026-09-28T11:53:24Z", "updated_at": "2026-09-28T11:53:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.34974v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.34974"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.34974"}], "name": "Before Acting, Change the State: Prospective State Intervention for Web Agents under Deceptive Interfaces", "publication": null, "published_at": "2026-09-28", "record_sha256": "400ff6e0da616efbaeec99063d84d6e8ac01c2423757ca8aac84b216ea31aebd", "relations": [], "resource_id": "arxiv:2609.34974", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.34974v1", "topics": ["safety-security", "planning", "robustness"], "updated_at": "2026-09-28"}, "text": "Before Acting, Change the State: Prospective State Intervention for Web Agents under Deceptive Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Veer, a runtime defense that changes task-relevant web state when an otherwise valid proposed action would produce an unauthorized consequence. It plans a prospective intervention and grounds and verifies its execution before the base agent continues.\nPlatforms: web\nTopics: safety-security, planning, robustness\nPublished: 2026-09-28\nAuthors: Ruozhao Yang, Mingfei Cheng, Xiaofei Xie\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.34974v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-34974/"}
{"id": "arxiv:2609.35026", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5090c6b7879e38102b1180ee1e4dfd1552af777172e2f94bc0f9a684fb59dc7", "submitted_at": "2026-09-28T12:24:23Z", "updated_at": "2026-09-28T12:24:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35026v1"}, "authors": ["Anton Emelyanov", "Maria Tikhonova", "Zaven Martirosian", "Sergei Averkiev", "Alena Fenogenova"], "canonical_source_url": "https://arxiv.org/abs/2609.35026", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-35026/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d5090c6b7879e38102b1180ee1e4dfd1552af777172e2f94bc0f9a684fb59dc7", "submitted_at": "2026-09-28T12:24:23Z", "updated_at": "2026-09-28T12:24:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35026v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.35026"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35026"}], "name": "WebPageBench: Event-Level Verification and Controlled UI-Variant Generation for Web Agents", "publication": null, "published_at": "2026-09-28", "record_sha256": "1b0b645b890acd2b0bbfdef8979908f254cc7ca9f6c33b13b83b26b67ff9e38e", "relations": [], "resource_id": "arxiv:2609.35026", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.35026v1", "topics": ["evaluation", "robustness"], "updated_at": "2026-09-28"}, "text": "WebPageBench: Event-Level Verification and Controlled UI-Variant Generation for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nVerifies web tasks through typed interface events and generates controlled UI variations while preserving instructions and success conditions.\nPlatforms: web\nTopics: evaluation, robustness\nPublished: 2026-09-28\nAuthors: Anton Emelyanov, Maria Tikhonova, Zaven Martirosian, Sergei Averkiev, Alena Fenogenova\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.35026v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-35026/"}
{"id": "arxiv:2609.35671", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ffb5951994cbccef0700ce850a1fae096a59b87334070750caf0b7b42d61c4b", "submitted_at": "2026-09-28T17:26:56Z", "updated_at": "2026-10-05T01:08:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.35671v2"}, "authors": ["Yangqin Jiang", "Lingrui Xu", "Chao Huang"], "canonical_source_url": "https://arxiv.org/abs/2609.35671", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-35671/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4ffb5951994cbccef0700ce850a1fae096a59b87334070750caf0b7b42d61c4b", "submitted_at": "2026-09-28T17:26:56Z", "updated_at": "2026-10-05T01:08:41Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.35671v2"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.35671"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35671"}], "name": "PhoneCLI: From App Interfaces to Callable Commands for Mobile Agents", "publication": null, "published_at": "2026-09-28", "record_sha256": "69202bda2f11c59e5b7fdb49de5063902192eb74f250d2a9d7c8747f7fb8362d", "relations": [], "resource_id": "arxiv:2609.35671", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.35671v2", "topics": ["automation", "efficiency", "mobile-navigation"], "updated_at": "2026-10-05"}, "text": "PhoneCLI: From App Interfaces to Callable Commands for Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PhoneCLI, which explores mobile app interfaces and compiles navigation paths into reusable commands without internal app APIs or instrumentation. At execution time it verifies compiled paths and falls back to a visual agent for open-ended interaction or failed navigation.\nPlatforms: android\nTopics: automation, efficiency, mobile-navigation\nPublished: 2026-09-28\nAuthors: Yangqin Jiang, Lingrui Xu, Chao Huang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.35671v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-35671/"}
{"id": "arxiv:2609.35814", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197cfb8424584955bc85602d6559037ad31ee815f89b45dc8f9f18c8636552ad", "submitted_at": "2026-09-20T16:16:23Z", "updated_at": "2026-09-20T16:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35814v1"}, "authors": ["Xunjian Yin", "Tianchen Guan", "Jinao Wang", "Weili Cao", "Daisy Xinlei Lin", "Royce Cheng-Yue", "Keagan Long", "Kyle Wong", "Bhuwan Dhingra", "Xiangjun Wang", "Shuyan Zhou"], "canonical_source_url": "https://arxiv.org/abs/2609.35814", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-35814/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "197cfb8424584955bc85602d6559037ad31ee815f89b45dc8f9f18c8636552ad", "submitted_at": "2026-09-20T16:16:23Z", "updated_at": "2026-09-20T16:16:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35814v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.35814"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35814"}], "name": "Constructing Challenging Browser-Use Tasks by Controlled Environment Interventions", "publication": null, "published_at": "2026-09-20", "record_sha256": "3582dfff8f34a6c1e0b6e3eaad72b9f3054cb3670e4cb40a86e4d2921f6e4311", "relations": [], "resource_id": "arxiv:2609.35814", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.35814v1", "topics": ["evaluation", "robustness", "data-generation", "state-tracking"], "updated_at": "2026-09-20"}, "text": "Constructing Challenging Browser-Use Tasks by Controlled Environment Interventions\nResource type: paper\nReview status: curated\nScope: core\n\nBreakingWeb pairs solvable browser tasks with controlled environment interventions that preserve the goal, target and backend success criterion. The paired evaluations isolate robustness failures, including agents declaring completion without the required state change.\nPlatforms: web\nTopics: evaluation, robustness, data-generation, state-tracking\nPublished: 2026-09-20\nAuthors: Xunjian Yin, Tianchen Guan, Jinao Wang, Weili Cao, Daisy Xinlei Lin, Royce Cheng-Yue, Keagan Long, Kyle Wong, Bhuwan Dhingra, Xiangjun Wang, Shuyan Zhou\nRecorded arXiv version: v1\n\nThe annotation was reviewed against the complete retained official abstract. Experimental findings and artifact availability have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.35814v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-35814/"}
{"id": "arxiv:2609.35904", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7efd2eafa1984d7c8f4bf6ce79a94cd2ca4189da330bdd056b4be9c07f7f9dc1", "submitted_at": "2026-09-28T04:02:07Z", "updated_at": "2026-09-28T04:02:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35904v1"}, "authors": ["Ziqi Zhang", "Shaohui Li", "Bing Li"], "canonical_source_url": "https://arxiv.org/abs/2609.35904", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-35904/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7efd2eafa1984d7c8f4bf6ce79a94cd2ca4189da330bdd056b4be9c07f7f9dc1", "submitted_at": "2026-09-28T04:02:07Z", "updated_at": "2026-09-28T04:02:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.35904v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.35904"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.35904"}], "name": "Structured Interaction, Visual Localization, and Robust Execution for Complex Web Tasks: A Technical Report on the WebRetriever Challenge", "publication": null, "published_at": "2026-09-28", "record_sha256": "d4e6af3711f1b5dbfb821281a5868debfd66d123c0bb4d03684e38f017a9975d", "relations": [], "resource_id": "arxiv:2609.35904", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.35904v1", "topics": ["web-navigation", "grounding", "robustness", "memory"], "updated_at": "2026-09-28"}, "text": "Structured Interaction, Visual Localization, and Robust Execution for Complex Web Tasks: A Technical Report on the WebRetriever Challenge\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes a WebRetriever agent that combines semantic webpage operations with visual localization when structured representations are insufficient. Hierarchical context management and fault-aware execution support processing tasks across concurrent browsers.\nPlatforms: web\nTopics: web-navigation, grounding, robustness, memory\nPublished: 2026-09-28\nAuthors: Ziqi Zhang, Shaohui Li, Bing Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.35904v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-35904/"}
{"id": "arxiv:2609.36201", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92c63d4cf58375aa7068570f8a98c145e24f8d393c5807c0aff3ed67f18b9f0", "submitted_at": "2026-09-28T20:01:35Z", "updated_at": "2026-09-28T20:01:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36201v1"}, "authors": ["Jianxing Chen", "Xiao Yu", "Shipra Agrawal", "Zhou Yu"], "canonical_source_url": "https://arxiv.org/abs/2609.36201", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-36201/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e92c63d4cf58375aa7068570f8a98c145e24f8d393c5807c0aff3ed67f18b9f0", "submitted_at": "2026-09-28T20:01:35Z", "updated_at": "2026-09-28T20:01:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36201v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.36201"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36201"}], "name": "SCOUT: Synergizing Reasoning and Tool-Use for Computer-Use Safety", "publication": null, "published_at": "2026-09-28", "record_sha256": "b60374384862256bc91a3110b49ab7d3ccf92e4cfde17daee0abe0eb8ae1fac9", "relations": [], "resource_id": "arxiv:2609.36201", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.36201v1", "topics": ["safety-security", "evaluation", "tool-integration"], "updated_at": "2026-09-28"}, "text": "SCOUT: Synergizing Reasoning and Tool-Use for Computer-Use Safety\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SCOUT, a safety verifier that first derives task-specific completion and safety rubrics from a computer-use trajectory, then actively probes the resulting environment for evidence. It evaluates whether these two stages detect harmful consequences that screenshots alone can miss.\nPlatforms: unspecified\nTopics: safety-security, evaluation, tool-integration\nPublished: 2026-09-28\nAuthors: Jianxing Chen, Xiao Yu, Shipra Agrawal, Zhou Yu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.36201v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-36201/"}
{"id": "arxiv:2609.36488", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fe39365133228daa770445f19002f6abe34ca1151cf4952a225ba0001de5250", "submitted_at": "2026-09-29T01:46:48Z", "updated_at": "2026-09-29T01:46:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36488v1"}, "authors": ["Dongchan Shin", "Xing Han Lù", "Jiaqi Deng", "Jay Gala", "Tomás Vergara Browne", "Jaewon Moon", "Fengyuan Liu", "Alexandre Drouin", "Siva Reddy", "Alexandre Lacoste"], "canonical_source_url": "https://arxiv.org/abs/2609.36488", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-36488/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2fe39365133228daa770445f19002f6abe34ca1151cf4952a225ba0001de5250", "submitted_at": "2026-09-29T01:46:48Z", "updated_at": "2026-09-29T01:46:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36488v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.36488"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36488"}], "name": "AdaptArena: Evaluating Test-Time Personalization of Web Agents", "publication": null, "published_at": "2026-09-29", "record_sha256": "87dceb17792c5b4a3f832f9b1e376033ba62ce879f793babc88268b7d73f1a74", "relations": [], "resource_id": "arxiv:2609.36488", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.36488v1", "topics": ["evaluation", "memory", "web-navigation"], "updated_at": "2026-09-29"}, "text": "AdaptArena: Evaluating Test-Time Personalization of Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AdaptArena to test whether web agents can infer unstated user preferences from relevant past trajectories and carry those preferences into new tasks. Its AdaptiveAgent baseline separates preference-retrieval challenges from failures in downstream web execution.\nPlatforms: web\nTopics: evaluation, memory, web-navigation\nPublished: 2026-09-29\nAuthors: Dongchan Shin, Xing Han Lù, Jiaqi Deng, Jay Gala, Tomás Vergara Browne, Jaewon Moon, Fengyuan Liu, Alexandre Drouin, Siva Reddy, Alexandre Lacoste\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.36488v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-36488/"}
{"id": "arxiv:2609.36923", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3af30198735ca9b9d192e15acbb7f8a37fe222fa2c0c1d2ccff448ef5e79f55", "submitted_at": "2026-09-29T07:41:34Z", "updated_at": "2026-09-29T07:41:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36923v1"}, "authors": ["Bin Kang", "Jiarui Ouyang", "Li Jiang", "Bin Chen", "Zhuotao Tian"], "canonical_source_url": "https://arxiv.org/abs/2609.36923", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-36923/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e3af30198735ca9b9d192e15acbb7f8a37fe222fa2c0c1d2ccff448ef5e79f55", "submitted_at": "2026-09-29T07:41:34Z", "updated_at": "2026-09-29T07:41:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36923v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.36923"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36923"}], "name": "PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval", "publication": null, "published_at": "2026-09-29", "record_sha256": "8dd67cb9199e731875e92ed2fec19721bfb926206485cb7449dba2062ceed5af", "relations": [], "resource_id": "arxiv:2609.36923", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.36923v1", "topics": ["memory", "world-models", "robustness", "long-horizon"], "updated_at": "2026-09-29"}, "text": "PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval\nResource type: paper\nReview status: curated\nScope: core\n\nPresents PrecogUI, which retrieves prior anomaly and success patterns and predicts symbolic UI transitions before choosing actions. It also introduces InterfereBench, generated for evaluating long GUI workflows with disruptive events and recovery requirements.\nPlatforms: unspecified\nTopics: memory, world-models, robustness, long-horizon\nPublished: 2026-09-29\nAuthors: Bin Kang, Jiarui Ouyang, Li Jiang, Bin Chen, Zhuotao Tian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.36923v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-36923/"}
{"id": "arxiv:2609.36927", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9890482c07ef44f6e274c802bae5decd9d883b381545b3034ebd4001cdb34", "submitted_at": "2026-09-29T07:44:22Z", "updated_at": "2026-09-29T07:44:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36927v1"}, "authors": ["Hyewon Suh", "Thanh Minh Nguyen", "Chih-Lun Lee", "Darrow Hartman", "Lizhao Liu", "Xin Eric Wang", "Ang Li", "Jiachen Yang"], "canonical_source_url": "https://arxiv.org/abs/2609.36927", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-36927/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fdd9890482c07ef44f6e274c802bae5decd9d883b381545b3034ebd4001cdb34", "submitted_at": "2026-09-29T07:44:22Z", "updated_at": "2026-09-29T07:44:22Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36927v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.36927"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36927"}], "name": "Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient Execution", "publication": null, "published_at": "2026-09-29", "record_sha256": "04ca32cc8bb34da11944562ffcee4bf42673ffc8867f76ad5d1781d65ea7a1b4", "relations": [], "resource_id": "arxiv:2609.36927", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.36927v1", "topics": ["automation", "workflow", "efficiency"], "updated_at": "2026-09-29"}, "text": "Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient Execution\nResource type: paper\nReview status: curated\nScope: core\n\nLearns reusable computer-use policies that encode stable workflow logic in programs while delegating grounding and state checks to neural models. Policy iteration executes variants, diagnoses failures, revises the program, and adds verification before state-changing actions.\nPlatforms: desktop\nTopics: automation, workflow, efficiency\nPublished: 2026-09-29\nAuthors: Hyewon Suh, Thanh Minh Nguyen, Chih-Lun Lee, Darrow Hartman, Lizhao Liu, Xin Eric Wang, Ang Li, Jiachen Yang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.36927v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-36927/"}
{"id": "arxiv:2609.36939", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cee8117e302aa190d38c8298bf479936a5a27aa9b85716a76a2a90499069c", "submitted_at": "2026-09-29T07:51:57Z", "updated_at": "2026-09-29T07:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36939v1"}, "authors": ["Shengtian Yang", "Ziyu Xiong", "Kaibing Yang", "Guangfeng Cai", "Yewen Li", "Peng Jiang", "Gai Kun", "Qingpeng Cai", "Lei Feng"], "canonical_source_url": "https://arxiv.org/abs/2609.36939", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-36939/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "d45cee8117e302aa190d38c8298bf479936a5a27aa9b85716a76a2a90499069c", "submitted_at": "2026-09-29T07:51:57Z", "updated_at": "2026-09-29T07:51:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.36939v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.36939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.36939"}], "name": "SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents", "publication": null, "published_at": "2026-09-29", "record_sha256": "af9fc675268e26e79a6f9f7e34a0eba7b9c0a05b600a4a8ac16b3d8a04ab485e", "relations": [], "resource_id": "arxiv:2609.36939", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.36939v1", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": "2026-09-29"}, "text": "SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Spatial Credit Assignment, which uses click coordinates to refine group-relative reinforcement-learning updates for GUI grounding. It distinguishes failed clicks by spatial evidence while leaving the deployed action policy unchanged.\nPlatforms: unspecified\nTopics: grounding, reinforcement-learning, evaluation\nPublished: 2026-09-29\nAuthors: Shengtian Yang, Ziyu Xiong, Kaibing Yang, Guangfeng Cai, Yewen Li, Peng Jiang, Gai Kun, Qingpeng Cai, Lei Feng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.36939v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-36939/"}
{"id": "arxiv:2609.37053", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e032f67257a52dbc59310d921d802e4dc85af3af110d22bb6d005d202bbaa6b2", "submitted_at": "2026-09-29T08:58:51Z", "updated_at": "2026-09-29T08:58:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37053v1"}, "authors": ["Mei Wu", "Rui Xie", "Runyu Zhang", "Yuqiang Li", "Tianfan Fu", "Bo Chen", "Kai Yu", "Xin Chen", "Lu Chen"], "canonical_source_url": "https://arxiv.org/abs/2609.37053", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-37053/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e032f67257a52dbc59310d921d802e4dc85af3af110d22bb6d005d202bbaa6b2", "submitted_at": "2026-09-29T08:58:51Z", "updated_at": "2026-09-29T08:58:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37053v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.37053"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37053"}], "name": "MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows", "publication": null, "published_at": "2026-09-29", "record_sha256": "af14f829f9a646238cf2a38dfc8840ce4e0d5ea9860a84421eda47823dbe226e", "relations": [], "resource_id": "arxiv:2609.37053", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.37053v1", "topics": ["science", "evaluation", "cross-app"], "updated_at": "2026-09-29"}, "text": "MatToolBench: Benchmarking Multimodal Agents in Real-World Materials Science Workflows\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MatToolBench, a Windows-based materials-science benchmark combining GUI operation, scientific-software scripting, and database queries. Expert-defined subcriteria support partial-credit scoring, while the evaluation examines domain knowledge gaps and cross-tool handoff failures.\nPlatforms: desktop, windows\nTopics: science, evaluation, cross-app\nPublished: 2026-09-29\nAuthors: Mei Wu, Rui Xie, Runyu Zhang, Yuqiang Li, Tianfan Fu, Bo Chen, Kai Yu, Xin Chen, Lu Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.37053v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-37053/"}
{"id": "arxiv:2609.37176", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "945184e8fc790979a1f7e466ef6b4e727ddd11c3f285be0371fa4dda42615f40", "submitted_at": "2026-09-29T10:01:52Z", "updated_at": "2026-09-29T10:01:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37176v1"}, "authors": ["Giulio Segalini", "Zhi Wen Soi", "Jérémie Decouchant", "Lydia Chen"], "canonical_source_url": "https://arxiv.org/abs/2609.37176", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-37176/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "945184e8fc790979a1f7e466ef6b4e727ddd11c3f285be0371fa4dda42615f40", "submitted_at": "2026-09-29T10:01:52Z", "updated_at": "2026-09-29T10:01:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.37176v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.37176"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37176"}], "name": "Absorbed in Inertia: Activation Analysis for Computer-Use Agents", "publication": null, "published_at": "2026-09-29", "record_sha256": "dff1467587cd0b6efb6262cce77cdc3965715bf6f21159d7752ee09c927fcace", "relations": [], "resource_id": "arxiv:2609.37176", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.37176v1", "topics": ["robustness", "observability", "memory"], "updated_at": "2026-09-29"}, "text": "Absorbed in Inertia: Activation Analysis for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies repetitive, unproductive actions in desktop agents through their internal activation states. It proposes Reset, Reroute, Restore, a context intervention that temporarily resets an agent’s trajectory and later restores it to help escape repeated ineffective behavior.\nPlatforms: desktop\nTopics: robustness, observability, memory\nPublished: 2026-09-29\nAuthors: Giulio Segalini, Zhi Wen Soi, Jérémie Decouchant, Lydia Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.37176v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-37176/"}
{"id": "arxiv:2609.37686", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aab7f586cb58a1f2d20267fd810ba34d6d4d166e3da715d2b3ff0b77e464473e", "submitted_at": "2026-09-29T14:32:35Z", "updated_at": "2026-10-09T08:15:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.37686v2"}, "authors": ["Hongcheng Gao", "Hailong Qu", "Yu Lei", "Henghui Sun", "Haoyang Li", "Yipeng Wei", "Naihao Xue", "Xiaohan Yu", "Zhuo Tao", "Yihe Zang", "Yajiao Wang", "Jingyi Tang", "Yi Li", "Jingjing Zhou", "Jie Luo", "Bohan Zeng", "Chengyu Shen", "Hao Jiang", "Chong Chen", "Bowen Qu", "Olive Huang", "Zeqiang Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.37686", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-37686/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aab7f586cb58a1f2d20267fd810ba34d6d4d166e3da715d2b3ff0b77e464473e", "submitted_at": "2026-09-29T14:32:35Z", "updated_at": "2026-10-09T08:15:21Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.37686v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.37686"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.37686"}], "name": "EngiWorld: What Can Frontier Agents Deliver in Professional Engineering Environments?", "publication": null, "published_at": "2026-09-29", "record_sha256": "e6853c7c3bcd4fa47adcd16a523b58e4ff3de69142c3edc810ca365fafc39db9", "relations": [], "resource_id": "arxiv:2609.37686", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.37686v2", "topics": ["evaluation", "science", "cross-app"], "updated_at": "2026-10-09"}, "text": "EngiWorld: What Can Frontier Agents Deliver in Professional Engineering Environments?\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces professional engineering tasks across software domains, with GUI and CLI access and artifact checks for geometric, physical, and design validity.\nPlatforms: desktop\nTopics: evaluation, science, cross-app\nPublished: 2026-09-29\nAuthors: Hongcheng Gao, Hailong Qu, Yu Lei, Henghui Sun, Haoyang Li, Yipeng Wei, Naihao Xue, Xiaohan Yu, Zhuo Tao, Yihe Zang, Yajiao Wang, Jingyi Tang, Yi Li, Jingjing Zhou, Jie Luo, Bohan Zeng, Chengyu Shen, Hao Jiang, Chong Chen, Bowen Qu, Olive Huang, Zeqiang Wang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.37686v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-37686/"}
{"id": "arxiv:2609.38008", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b394677ebe37782950cfc56c0be65ca7a93f34e5550ec9425d91921bf1799847", "submitted_at": "2026-09-29T16:57:25Z", "updated_at": "2026-09-29T16:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38008v1"}, "authors": ["Tongbo Chen", "Junbo Niu", "Zhengxi Lu", "Niu Lian", "Fei Tang", "Yuchen Yan", "Yike Hong", "Yong Du", "Yizhou Liu", "Bofan Chen", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2609.38008", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-38008/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b394677ebe37782950cfc56c0be65ca7a93f34e5550ec9425d91921bf1799847", "submitted_at": "2026-09-29T16:57:25Z", "updated_at": "2026-09-29T16:57:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38008v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.38008"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38008"}], "name": "HybridCUA: Learning to Orchestrate GUI and CLI for Computer-Use Agents", "publication": null, "published_at": "2026-09-29", "record_sha256": "3da7d9abb4eaebc094c15c1aaac1e8beed3cc166fd1a01b53920c0d713a707e5", "relations": [], "resource_id": "arxiv:2609.38008", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.38008v1", "topics": ["tool-integration", "reinforcement-learning", "data-generation"], "updated_at": "2026-09-29"}, "text": "HybridCUA: Learning to Orchestrate GUI and CLI for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains computer agents on verified GUI, command-line, and interleaved trajectories to learn when and how to combine interaction modes.\nPlatforms: desktop, windows\nTopics: tool-integration, reinforcement-learning, data-generation\nPublished: 2026-09-29\nAuthors: Tongbo Chen, Junbo Niu, Zhengxi Lu, Niu Lian, Fei Tang, Yuchen Yan, Yike Hong, Yong Du, Yizhou Liu, Bofan Chen, Yongliang Shen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.38008v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-38008/"}
{"id": "arxiv:2609.38184", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14844abefab6b4eb464a1c7675f9de7f460849872ebb776230724ed5828feed", "submitted_at": "2026-08-07T07:51:13Z", "updated_at": "2026-08-07T07:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38184v1"}, "authors": ["Yongjiang Wu", "Junyuan Zhang", "Ada Chen", "Kuiyi Gao", "Wenxuan Wang"], "canonical_source_url": "https://arxiv.org/abs/2609.38184", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-38184/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f14844abefab6b4eb464a1c7675f9de7f460849872ebb776230724ed5828feed", "submitted_at": "2026-08-07T07:51:13Z", "updated_at": "2026-08-07T07:51:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38184v1"}}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.38184"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38184"}], "name": "RealGUINoise: An Interactive Cross-Platform Benchmark for GUI Agent Robustness under Real-World Interface Noise", "publication": null, "published_at": "2026-08-07", "record_sha256": "f583fdf7dff64d8cace008dade8a3961dbfc2c08f008440803dbb6dccedbeeed", "relations": [], "resource_id": "arxiv:2609.38184", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.38184v1", "topics": ["robustness", "evaluation", "safety-security"], "updated_at": "2026-08-07"}, "text": "RealGUINoise: An Interactive Cross-Platform Benchmark for GUI Agent Robustness under Real-World Interface Noise\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates interactive GUI agents under varied realistic interface noise, comparing task completion, safety, and behavior with clean-environment runs.\nPlatforms: web, desktop, cross-platform\nTopics: robustness, evaluation, safety-security\nPublished: 2026-08-07\nAuthors: Yongjiang Wu, Junyuan Zhang, Ada Chen, Kuiyi Gao, Wenxuan Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.38184v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-38184/"}
{"id": "arxiv:2609.38778", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb910e64ebb65bd49e8729e4373c1c81ce4b2d0a32647a5abaf15f52737e578d", "submitted_at": "2026-09-30T02:11:02Z", "updated_at": "2026-09-30T02:11:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38778v1"}, "authors": ["Hongbo Zhang", "Liuyang Song", "Quanquan Li", "Daqian Yang", "Yan Wen", "Zhengtao Yao"], "canonical_source_url": "https://arxiv.org/abs/2609.38778", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-38778/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bb910e64ebb65bd49e8729e4373c1c81ce4b2d0a32647a5abaf15f52737e578d", "submitted_at": "2026-09-30T02:11:02Z", "updated_at": "2026-09-30T02:11:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38778v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.38778"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38778"}], "name": "Action Conditioned Bisimulation For GUI Agent Memory", "publication": null, "published_at": "2026-09-30", "record_sha256": "7f0f9b2b00fdfdafdaf9fcd357fa031826f0e8f9559690d63498e444fb5c2567", "relations": [], "resource_id": "arxiv:2609.38778", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.38778v1", "topics": ["memory", "web-navigation"], "updated_at": "2026-09-30"}, "text": "Action Conditioned Bisimulation For GUI Agent Memory\nResource type: paper\nReview status: curated\nScope: core\n\nDefines a GUI-memory merge rule that treats two states as equivalent only when shared actions lead to matching outcomes and successor states. The training-free method replaces observation-similarity merging and is tested in closed-loop MiniWoB++ interaction.\nPlatforms: web\nTopics: memory, web-navigation\nPublished: 2026-09-30\nAuthors: Hongbo Zhang, Liuyang Song, Quanquan Li, Daqian Yang, Yan Wen, Zhengtao Yao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.38778v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-38778/"}
{"id": "arxiv:2609.38867", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bded059a37101b0e6a91ba662e0d70e5e010ddb8c3557598dff2ba82afa45e34", "submitted_at": "2026-09-30T03:19:25Z", "updated_at": "2026-09-30T03:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38867v1"}, "authors": ["Terumi Chiba", "Guangzhi Sun", "Zheqi Yuan", "Chao Zhang"], "canonical_source_url": "https://arxiv.org/abs/2609.38867", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-38867/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bded059a37101b0e6a91ba662e0d70e5e010ddb8c3557598dff2ba82afa45e34", "submitted_at": "2026-09-30T03:19:25Z", "updated_at": "2026-09-30T03:19:25Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.38867v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2609.38867"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.38867"}], "name": "Talk2Agent: Benchmarking Voice Interfaces for Text Agents", "publication": null, "published_at": "2026-09-30", "record_sha256": "4bc6604e98970dc279c3ec79efb75be3f2d6fe2887dc3a16ea3f9c5bbe720495", "relations": [], "resource_id": "arxiv:2609.38867", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.38867v1", "topics": ["evaluation", "robustness"], "updated_at": "2026-09-30"}, "text": "Talk2Agent: Benchmarking Voice Interfaces for Text Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Talk2Agent, a benchmark of spoken instructions for computer-use tasks that measures whether voice interfaces preserve task-critical information. It pairs human speech with an execution-free intention-based metric and checks how that metric relates to downstream task completion.\nPlatforms: desktop\nTopics: evaluation, robustness\nPublished: 2026-09-30\nAuthors: Terumi Chiba, Guangzhi Sun, Zheqi Yuan, Chao Zhang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.38867v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-38867/"}
{"id": "arxiv:2609.39547", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3594620e97b41f3a7481871588b63ffa4aa8822e8beb42334b2cd888bcfcfaaf", "submitted_at": "2026-09-30T11:45:51Z", "updated_at": "2026-09-30T11:45:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39547v1"}, "authors": ["Bo Han", "Qianyi Wang", "Shuai Liu", "Xiong Zifan", "Changqiao Wu", "Yuanfa Li", "Pengzhi Gao", "Wei Liu", "Jian Luan", "Heng Qu", "Yunpeng Song", "Zhongmin Cai"], "canonical_source_url": "https://arxiv.org/abs/2609.39547", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-39547/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3594620e97b41f3a7481871588b63ffa4aa8822e8beb42334b2cd888bcfcfaaf", "submitted_at": "2026-09-30T11:45:51Z", "updated_at": "2026-09-30T11:45:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39547v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.39547"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.39547"}], "name": "Learning Reliable GUI Agents under Imperfect Priors", "publication": null, "published_at": "2026-09-30", "record_sha256": "db0e46181ea157ff6f34397f8973fa5688a5b01a3aac68801720bd841822594e", "relations": [], "resource_id": "arxiv:2609.39547", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.39547v1", "topics": ["data-generation", "memory", "robustness"], "updated_at": "2026-09-30"}, "text": "Learning Reliable GUI Agents under Imperfect Priors\nResource type: paper\nReview status: curated\nScope: core\n\nCombines structured GUI exploration with training designed to tolerate imperfect retrieved experience. It builds state-transition graphs and synthetic task trajectories, then injects realistic forms of interface drift so agents learn when to rely on or reject prior guidance.\nPlatforms: unspecified\nTopics: data-generation, memory, robustness\nPublished: 2026-09-30\nAuthors: Bo Han, Qianyi Wang, Shuai Liu, Xiong Zifan, Changqiao Wu, Yuanfa Li, Pengzhi Gao, Wei Liu, Jian Luan, Heng Qu, Yunpeng Song, Zhongmin Cai\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.39547v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-39547/"}
{"id": "arxiv:2609.39903", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "690cd75934bae65311240cd4e6db7b8fd607056bd60799bf70cddd03efda9261", "submitted_at": "2026-09-30T15:03:05Z", "updated_at": "2026-09-30T15:03:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39903v1"}, "authors": ["Dingyuan Dai", "Heli Qi", "Lei Liu", "Yinxi Li", "Baiding Chen", "Zijun Dou", "Qingcheng Zeng", "Qi Kang", "Oliver Sun", "Eric Wang", "Bo Zhou", "Haixin Wang", "Yufan Du", "Shi Bo", "Ruihan Lin", "Mengqi Yuan", "Dunjie Lu", "Steven Dillmann", "Yiming Shi", "Tina Su", "Amy Xin", "Minghao Liu", "Xi Wang", "Xu Huang", "Ge Zhang", "Pengyu Nie", "Zhen Yang", "Jie Tang", "Juanzi Li", "Weihao Xuan", "Tianyu Liu"], "canonical_source_url": "https://arxiv.org/abs/2609.39903", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-39903/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "690cd75934bae65311240cd4e6db7b8fd607056bd60799bf70cddd03efda9261", "submitted_at": "2026-09-30T15:03:05Z", "updated_at": "2026-09-30T15:03:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.39903v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.39903"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.39903"}], "name": "OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software", "publication": null, "published_at": "2026-09-30", "record_sha256": "2d4df19b75830a5c8f14103ddec46b1d784390dc3037ac54acaf00299ccac08b", "relations": [], "resource_id": "arxiv:2609.39903", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.39903v1", "topics": ["evaluation", "science"], "updated_at": "2026-09-30"}, "text": "OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software\nResource type: paper\nReview status: curated\nScope: core\n\nBenchmarks computer agents on scientific-software workflows with task-specific checks of application states and generated research artifacts.\nPlatforms: desktop\nTopics: evaluation, science\nPublished: 2026-09-30\nAuthors: Dingyuan Dai, Heli Qi, Lei Liu, Yinxi Li, Baiding Chen, Zijun Dou, Qingcheng Zeng, Qi Kang, Oliver Sun, Eric Wang, Bo Zhou, Haixin Wang, Yufan Du, Shi Bo, Ruihan Lin, Mengqi Yuan, Dunjie Lu, Steven Dillmann, Yiming Shi, Tina Su, Amy Xin, Minghao Liu, Xi Wang, Xu Huang, Ge Zhang, Pengyu Nie, Zhen Yang, Jie Tang, Juanzi Li, Weihao Xuan, Tianyu Liu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.39903v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-39903/"}
{"id": "arxiv:2609.40253", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f3745b6b497c8819dc990bdb5d74e7af463d23654cbc7cdd76ade0faf2b3e0c", "submitted_at": "2026-09-30T17:40:09Z", "updated_at": "2026-10-01T17:50:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.40253v2"}, "authors": ["Yong Du", "Tongbo Chen", "Zhengxi Lu", "Yizhou Liu", "Bofan Chen", "Tao Jiang", "Wenhao Xu", "Yongliang Shen"], "canonical_source_url": "https://arxiv.org/abs/2609.40253", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-40253/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "8f3745b6b497c8819dc990bdb5d74e7af463d23654cbc7cdd76ade0faf2b3e0c", "submitted_at": "2026-09-30T17:40:09Z", "updated_at": "2026-10-01T17:50:14Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2609.40253v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.40253"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.40253"}], "name": "ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents", "publication": null, "published_at": "2026-09-30", "record_sha256": "ef0fc61ecc49685fff0d2c89385a370ab4a1ce5b19797f041d60a63a4ee90ee2", "relations": [], "resource_id": "arxiv:2609.40253", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.40253v2", "topics": ["reinforcement-learning", "evaluation"], "updated_at": "2026-10-01"}, "text": "ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ComputerSD, which turns feedback from executed GUI transitions into online self-distillation guidance and step-level scores. A GUI analyzer supplies these signals, which are combined with trajectory-level reinforcement learning in asynchronous training.\nPlatforms: desktop\nTopics: reinforcement-learning, evaluation\nPublished: 2026-09-30\nAuthors: Yong Du, Tongbo Chen, Zhengxi Lu, Yizhou Liu, Bofan Chen, Tao Jiang, Wenhao Xu, Yongliang Shen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.40253v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-40253/"}
{"id": "arxiv:2609.40284", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "502912cdc35afd61beeb0202d55fc8feefcd92cd3719cd0e85bc71561d6f5a3c", "submitted_at": "2026-09-30T17:48:06Z", "updated_at": "2026-09-30T17:48:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.40284v1"}, "authors": ["Pranjal Aggarwal", "Lawrence Keunho Jang", "Sean Welleck", "Daniel Fried", "Ruslan Salakhutdinov", "Jing Yu Koh"], "canonical_source_url": "https://arxiv.org/abs/2609.40284", "citation_url": "https://awesomecua.github.io/resources/arxiv-2609-40284/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "502912cdc35afd61beeb0202d55fc8feefcd92cd3719cd0e85bc71561d6f5a3c", "submitted_at": "2026-09-30T17:48:06Z", "updated_at": "2026-09-30T17:48:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2609.40284v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2609.40284"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2609.40284"}], "name": "cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents", "publication": null, "published_at": "2026-09-30", "record_sha256": "b3757e4dd7143bb4d0e7f4187fe2317560b7b5d737e76a0275e0146ecdd570e4", "relations": [], "resource_id": "arxiv:2609.40284", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2609.40284v1", "topics": ["evaluation", "efficiency"], "updated_at": "2026-09-30"}, "text": "cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStandardizes virtual machines, interaction pipelines, and agent interfaces to compare computer-use speed, cost, and task performance.\nPlatforms: desktop\nTopics: evaluation, efficiency\nPublished: 2026-09-30\nAuthors: Pranjal Aggarwal, Lawrence Keunho Jang, Sean Welleck, Daniel Fried, Ruslan Salakhutdinov, Jing Yu Koh\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2609.40284v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2609-40284/"}
{"id": "arxiv:2610.00948", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c43224c865fba2a9f4c826a8e912ab0c07297716e4b7640bf312cce9c9c510e", "submitted_at": "2026-10-01T02:31:20Z", "updated_at": "2026-10-01T02:31:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00948v1"}, "authors": ["Geyi Yang", "Zikun Qu", "Xiang Li", "Zhiyong Wang", "Min Zhang", "Shipei Zeng", "Zhongxiang Dai"], "canonical_source_url": "https://arxiv.org/abs/2610.00948", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-00948/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5c43224c865fba2a9f4c826a8e912ab0c07297716e4b7640bf312cce9c9c510e", "submitted_at": "2026-10-01T02:31:20Z", "updated_at": "2026-10-01T02:31:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.00948v1"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.00948"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.00948"}], "name": "GUI-HARVEST: Self-Improving GUI Agents through Evidence-Driven Harness Evolution", "publication": null, "published_at": "2026-10-01", "record_sha256": "cbbadf2213bde254b5da91fb2b3ed5704773f9b6647dbcd28271891f27cae273", "relations": [], "resource_id": "arxiv:2610.00948", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.00948v1", "topics": ["automation", "observability", "robustness"], "updated_at": "2026-10-01"}, "text": "GUI-HARVEST: Self-Improving GUI Agents through Evidence-Driven Harness Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nImproves GUI-agent execution code by analyzing action effects and repeated-task failures, then validating bounded changes with frozen model weights.\nPlatforms: desktop, windows\nTopics: automation, observability, robustness\nPublished: 2026-10-01\nAuthors: Geyi Yang, Zikun Qu, Xiang Li, Zhiyong Wang, Min Zhang, Shipei Zeng, Zhongxiang Dai\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.00948v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-00948/"}
{"id": "arxiv:2610.01215", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecd40b78c6de41abb442705b993aad8ebf7b35741536a11e00ebbd60140d1f18", "submitted_at": "2026-10-01T07:22:53Z", "updated_at": "2026-10-01T07:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01215v1"}, "authors": ["Cheng Yang", "Yifan Wu", "Yutao Huang", "Zhaohua Zhang", "Beiduo Chen", "Muxi Chen", "Chenchen Zhao", "Hexuan Deng", "Haolin Yang", "Geyuan Zhu", "Sa Zhu", "Jianhuan Zhuo", "Qiuyong Xiao", "Jianhao Ruan", "Yiran Peng", "Jiayi Zhang", "Tian Ye", "Xinlei Yu", "Tianwen Jiang", "Jihong Zhang", "Yuyu Luo"], "canonical_source_url": "https://arxiv.org/abs/2610.01215", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-01215/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ecd40b78c6de41abb442705b993aad8ebf7b35741536a11e00ebbd60140d1f18", "submitted_at": "2026-10-01T07:22:53Z", "updated_at": "2026-10-01T07:22:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01215v1"}}, "environments": ["desktop", "linux", "windows", "macos", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.01215"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01215"}], "name": "AutoGUIWorld: Image Generators as Visual World Models for GUI Agent", "publication": null, "published_at": "2026-10-01", "record_sha256": "747f049a23f78176fe37e5619386c68bc0381a5b7232ac5ad33564bae5e9c625", "relations": [], "resource_id": "arxiv:2610.01215", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.01215v1", "topics": ["world-models", "data-generation", "grounding"], "updated_at": "2026-10-01"}, "text": "AutoGUIWorld: Image Generators as Visual World Models for GUI Agent\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AutoGUIWorld, which combines task planning and image editing to synthesize action-grounded GUI trajectories without running the depicted software. It filters generated transitions and tests whether training on them improves agents operating real desktop and scientific applications.\nPlatforms: desktop, linux, windows, macos, web\nTopics: world-models, data-generation, grounding\nPublished: 2026-10-01\nAuthors: Cheng Yang, Yifan Wu, Yutao Huang, Zhaohua Zhang, Beiduo Chen, Muxi Chen, Chenchen Zhao, Hexuan Deng, Haolin Yang, Geyuan Zhu, Sa Zhu, Jianhuan Zhuo, Qiuyong Xiao, Jianhao Ruan, Yiran Peng, Jiayi Zhang, Tian Ye, Xinlei Yu, Tianwen Jiang, Jihong Zhang, Yuyu Luo\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.01215v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-01215/"}
{"id": "arxiv:2610.01491", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a5b97264b3928c8dc74b955cc4baa7005cd34b0ce9e42cbc48a63bc6dd1f5a7", "submitted_at": "2026-10-01T11:31:37Z", "updated_at": "2026-10-01T11:31:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01491v1"}, "authors": ["Chengguang Gan", "Zimeng He", "Yoshihiro Tsujii", "Ken-ichiro Kobayashi", "Hiroki Itoh", "Kotaro Funakoshi"], "canonical_source_url": "https://arxiv.org/abs/2610.01491", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-01491/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9a5b97264b3928c8dc74b955cc4baa7005cd34b0ce9e42cbc48a63bc6dd1f5a7", "submitted_at": "2026-10-01T11:31:37Z", "updated_at": "2026-10-01T11:31:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01491v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.01491"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01491"}], "name": "Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories", "publication": null, "published_at": "2026-10-01", "record_sha256": "12a69f2459f05a500b80c1b27ff8e4f8e527ac53056d16c569da8936a7f7a7ef", "relations": [], "resource_id": "arxiv:2610.01491", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.01491v1", "topics": ["evaluation", "observability", "memory"], "updated_at": "2026-10-01"}, "text": "Auditing Web Agent Evaluation on WebArena-Lite: Human Review of Outcomes and Trajectories\nResource type: paper\nReview status: curated\nScope: core\n\nAudits WebArena-Lite outcomes and trajectories with human review to identify automatic-evaluator false negatives and the first consequential execution errors. It also studies explicit execution-state memory and procedural guidance, showing why final task scores can miss substantial differences in agent behavior.\nPlatforms: web\nTopics: evaluation, observability, memory\nPublished: 2026-10-01\nAuthors: Chengguang Gan, Zimeng He, Yoshihiro Tsujii, Ken-ichiro Kobayashi, Hiroki Itoh, Kotaro Funakoshi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.01491v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-01491/"}
{"id": "arxiv:2610.01787", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e85a601c31c6d06831ed1c590deef29d1c365e5d4751a308e42e8496ef11b19", "submitted_at": "2026-10-01T14:34:32Z", "updated_at": "2026-10-01T14:34:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01787v1"}, "authors": ["Beining Wu", "Zihao Ding", "Jun Huang"], "canonical_source_url": "https://arxiv.org/abs/2610.01787", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-01787/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7e85a601c31c6d06831ed1c590deef29d1c365e5d4751a308e42e8496ef11b19", "submitted_at": "2026-10-01T14:34:32Z", "updated_at": "2026-10-01T14:34:32Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.01787v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.01787"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.01787"}], "name": "Not All Experience Belongs in the Weights: Component Routing for Self-Improving GUI Agents", "publication": null, "published_at": "2026-10-01", "record_sha256": "374628489ce74adb74aec776be5c005d63e7697cf4f97e2f678dc1a2660af755", "relations": [], "resource_id": "arxiv:2610.01787", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.01787v1", "topics": ["memory", "evaluation", "efficiency"], "updated_at": "2026-10-01"}, "text": "Not All Experience Belongs in the Weights: Component Routing for Self-Improving GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nStudies how to route different parts of GUI experience into model weights or runtime context rather than treating whole trajectories as a single unit. It separates locators, procedures, state facts, and lessons, then predicts their destination from recurrence and dependence on the current state.\nPlatforms: unspecified\nTopics: memory, evaluation, efficiency\nPublished: 2026-10-01\nAuthors: Beining Wu, Zihao Ding, Jun Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.01787v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-01787/"}
{"id": "arxiv:2610.02320", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd66468c2670ec22216e1592719a05e0b56af69606237a50f2067a340172553e", "submitted_at": "2026-10-01T18:00:05Z", "updated_at": "2026-10-01T18:00:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02320v1"}, "authors": ["A. Said Gurbuz", "Ahmed Nassar", "Sunghwan Hong", "Marc Pollefeys", "Peter W. J. Staar"], "canonical_source_url": "https://arxiv.org/abs/2610.02320", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-02320/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "bd66468c2670ec22216e1592719a05e0b56af69606237a50f2067a340172553e", "submitted_at": "2026-10-01T18:00:05Z", "updated_at": "2026-10-01T18:00:05Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02320v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.02320"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02320"}], "name": "DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents", "publication": null, "published_at": "2026-10-01", "record_sha256": "1fe94c54f86fd7515fda151ffde96809bbccfafabf7b10a31279991833023acd", "relations": [], "resource_id": "arxiv:2610.02320", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.02320v1", "topics": ["data-generation", "grounding"], "updated_at": "2026-10-01"}, "text": "DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGenerates dense desktop supervision by varying real application scenes and combining screenshots, accessibility trees, and window geometry.\nPlatforms: desktop\nTopics: data-generation, grounding\nPublished: 2026-10-01\nAuthors: A. Said Gurbuz, Ahmed Nassar, Sunghwan Hong, Marc Pollefeys, Peter W. J. Staar\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.02320v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-02320/"}
{"id": "arxiv:2610.02542", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35ded32e7bac9a875a0fa41efad43fed5fb0aab54fade5cfc2a0628b1e666587", "submitted_at": "2026-10-01T22:30:08Z", "updated_at": "2026-10-01T22:30:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02542v1"}, "authors": ["Dhananjay Ashok", "Shantanu Agarwal", "Vivek Datla", "Jonathan May", "Alfy Samuel"], "canonical_source_url": "https://arxiv.org/abs/2610.02542", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-02542/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "35ded32e7bac9a875a0fa41efad43fed5fb0aab54fade5cfc2a0628b1e666587", "submitted_at": "2026-10-01T22:30:08Z", "updated_at": "2026-10-01T22:30:08Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02542v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.02542v1"], "reason": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56Z", "role": "primary", "sha256": "d8c22c6b3f07d4156c38af9dd7ca7c06a5d0a11889a7350d821a020e21604590", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.02542v1"}], "identifiers": {"arxiv": "2610.02542"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:08:56Z", "url": "https://arxiv.org/html/2610.02542v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02542"}], "name": "How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling", "publication": null, "published_at": "2026-10-01", "record_sha256": "90b920bb231f776b7e70b7c847ae142bc0f865428c59815046334532c19faced", "relations": [], "resource_id": "arxiv:2610.02542", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.02542v1", "topics": ["world-models", "planning", "evaluation", "web-navigation"], "updated_at": "2026-10-01"}, "text": "How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling\nResource type: paper\nReview status: curated\nScope: adjacent\n\nCompares fine-tuned, retrieval-based and hybrid language-model world models for next-observation prediction and action selection in text environments including WebShop.\nPlatforms: web\nTopics: world-models, planning, evaluation, web-navigation\nPublished: 2026-10-01\nAuthors: Dhananjay Ashok, Shantanu Agarwal, Vivek Datla, Jonathan May, Alfy Samuel\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.02542v1\"], \"reason\": \"The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.\"}}}\n\nAdjacent cross-domain study with WebShop as its web evaluation. WebShop uses text-rendered search/product pages and an explicit valid-action menu, not screenshot control or live purchases. A frozen policy proposes five actions; a world model predicts their next observations and the policy scores them before one is executed in the simulator. Utility is normalized environment reward, while fidelity uses a model judge; neither is a universal GUI task-success rate. Fine-tuning and retrieval use collected training experience; the hybrid masks same-episode retrieval, which does not establish general contamination freedom. The retained Atom and v1 HTML abstract wording differs; original bibliographic metadata and its abstract digest are preserved.\n\nPrimary source: https://arxiv.org/abs/2610.02542v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-02542/"}
{"id": "arxiv:2610.02617", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ed117876ece99f55e241e6365d0e785b907f9a7a24cceb01b1fd757191a35c9", "submitted_at": "2026-10-02T00:17:11Z", "updated_at": "2026-10-02T00:17:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02617v1"}, "authors": ["Yun-Yun Tsai", "Yuning Mao", "Shiqi Wang", "Junfeng Yang", "Sinong Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.02617", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-02617/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7ed117876ece99f55e241e6365d0e785b907f9a7a24cceb01b1fd757191a35c9", "submitted_at": "2026-10-02T00:17:11Z", "updated_at": "2026-10-02T00:17:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02617v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.02617"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:31.520084+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02617"}], "name": "WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness", "publication": null, "published_at": "2026-10-02", "record_sha256": "2c417c99610e7c334ae83680644f63eb567e8a94e3d41b39ca02ebf87625c0d1", "relations": [], "resource_id": "arxiv:2610.02617", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.02617v1", "topics": ["evaluation", "web-navigation", "ui-design"], "updated_at": "2026-10-02"}, "text": "WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces an execution-based benchmark for generated web interfaces, with a headless-browser UI-agent harness that checks DOM interactions and functional assertions.\nPlatforms: web\nTopics: evaluation, web-navigation, ui-design\nPublished: 2026-10-02\nAuthors: Yun-Yun Tsai, Yuning Mao, Shiqi Wang, Junfeng Yang, Sinong Wang\nRecorded arXiv version: v1\n\nAdjacent evaluation infrastructure: the target is generated WebUI functionality, not a general score for computer-use agents. This annotation is based on the complete retained abstract; released artifacts and experimental results were not independently inspected or reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.02617v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-02617/"}
{"id": "arxiv:2610.02932", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "923a2e8c6f95b38f84e66e4d2cacdfd54aae8348e17b3fab858d2d7151e82d5b", "submitted_at": "2026-10-02T07:22:46Z", "updated_at": "2026-10-02T07:22:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02932v1"}, "authors": ["Yulong Ming", "Jie Xu", "Zihan Wu", "Xiaohua Jia"], "canonical_source_url": "https://arxiv.org/abs/2610.02932", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-02932/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "923a2e8c6f95b38f84e66e4d2cacdfd54aae8348e17b3fab858d2d7151e82d5b", "submitted_at": "2026-10-02T07:22:46Z", "updated_at": "2026-10-02T07:22:46Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.02932v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.02932"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.02932"}], "name": "When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token Efficiency", "publication": null, "published_at": "2026-10-02", "record_sha256": "dcafdf68bb65d9c44976cd22c5c4fb7b9e4a6a4f57a223ac0a8fd60dbb3a2964", "relations": [], "resource_id": "arxiv:2610.02932", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.02932v1", "topics": ["automation", "efficiency", "workflow"], "updated_at": "2026-10-02"}, "text": "When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token Efficiency\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PACE, a protocol and online decision rule for deciding when repeated GUI procedures are worth compiling into programs. It accounts for failed compilation attempts, measured per-use savings, uncertain reuse, and execution costs under an explicit cumulative budget.\nPlatforms: unspecified\nTopics: automation, efficiency, workflow\nPublished: 2026-10-02\nAuthors: Yulong Ming, Jie Xu, Zihan Wu, Xiaohua Jia\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.02932v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-02932/"}
{"id": "arxiv:2610.03036", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c25f023b6ff66baa38661362d5730fc338e781db5bd23dd91216d3937ae730b0", "submitted_at": "2026-10-02T09:18:28Z", "updated_at": "2026-10-07T13:46:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03036v2"}, "authors": ["Jiangang Han"], "canonical_source_url": "https://arxiv.org/abs/2610.03036", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-03036/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c25f023b6ff66baa38661362d5730fc338e781db5bd23dd91216d3937ae730b0", "submitted_at": "2026-10-02T09:18:28Z", "updated_at": "2026-10-07T13:46:51Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03036v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.03036"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03036"}], "name": "WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites", "publication": null, "published_at": "2026-10-02", "record_sha256": "944fb69679263d02432a0458a7ea70eeb738f3dec32faf27d193498e1a525602", "relations": [], "resource_id": "arxiv:2610.03036", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.03036v2", "topics": ["web-navigation", "robustness", "observability"], "updated_at": "2026-10-07"}, "text": "WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes WebFovea, a vision-based live-web agent whose execution loop checks action parsing, browser effects, feedback accuracy, and observation delivery. The report analyzes failures such as coordinate mismatches and silent control interactions and documents corresponding harness changes.\nPlatforms: web\nTopics: web-navigation, robustness, observability\nPublished: 2026-10-02\nAuthors: Jiangang Han\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.03036v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-03036/"}
{"id": "arxiv:2610.03089", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7469484bb1186dd633b26e41dddc890008e893c303f93d2352f93c9d47fd78c2", "submitted_at": "2026-10-02T10:08:39Z", "updated_at": "2026-10-02T10:08:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.03089v1"}, "authors": ["Giulio Zingrillo", "Hanna Foerster", "Ilia Shumailov", "Yiren Zhao", "Robert Mullins"], "canonical_source_url": "https://arxiv.org/abs/2610.03089", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-03089/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7469484bb1186dd633b26e41dddc890008e893c303f93d2352f93c9d47fd78c2", "submitted_at": "2026-10-02T10:08:39Z", "updated_at": "2026-10-02T10:08:39Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.03089v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.03089"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03089"}], "name": "Securing Computer-Use Agents Against Branch Steering Attacks", "publication": null, "published_at": "2026-10-02", "record_sha256": "d48185139177eed10c60740622732055468091035ade8c4b2ac0feadcbf28517", "relations": [], "resource_id": "arxiv:2610.03089", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.03089v1", "topics": ["safety-security", "planning", "evaluation"], "updated_at": "2026-10-02"}, "text": "Securing Computer-Use Agents Against Branch Steering Attacks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces STEER-Bench to study branch-steering attacks that manipulate untrusted web data toward hazardous paths already permitted by an agent’s plan. Its COBRA architecture combines trusted branching plans with advance constraints on the parameters and destinations each branch can use.\nPlatforms: web\nTopics: safety-security, planning, evaluation\nPublished: 2026-10-02\nAuthors: Giulio Zingrillo, Hanna Foerster, Ilia Shumailov, Yiren Zhao, Robert Mullins\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.03089v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-03089/"}
{"id": "arxiv:2610.03818", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9dd28bdfb8879cbdb68f539c286365a06f4e9d04eba2420bb46f3575fdbfa05", "submitted_at": "2026-10-02T00:31:34Z", "updated_at": "2026-10-12T23:17:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03818v2"}, "authors": ["Maria Sanchez del Rio", "Mikhail Terekhov"], "canonical_source_url": "https://arxiv.org/abs/2610.03818", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-03818/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f9dd28bdfb8879cbdb68f539c286365a06f4e9d04eba2420bb46f3575fdbfa05", "submitted_at": "2026-10-02T00:31:34Z", "updated_at": "2026-10-12T23:17:28Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.03818v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.03818"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.03818"}], "name": "Control OSWorld: An AI Control Environment for GUI Computer Use Agents", "publication": null, "published_at": "2026-10-02", "record_sha256": "ce215215858c1e585eeb179c90718cc98f50609e89a12d95ef4bcedb320d23ec", "relations": [], "resource_id": "arxiv:2610.03818", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.03818v2", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-12"}, "text": "Control OSWorld: An AI Control Environment for GUI Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPairs OSWorld tasks with harmful side objectives to evaluate monitors that detect malicious behavior in computer-use agent trajectories.\nPlatforms: desktop\nTopics: safety-security, evaluation\nPublished: 2026-10-02\nAuthors: Maria Sanchez del Rio, Mikhail Terekhov\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.03818v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-03818/"}
{"id": "arxiv:2610.04116", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97af38af8647b769c129d917b65f8f744035308f6ace88706a4888a96fe345b9", "submitted_at": "2026-10-02T22:43:06Z", "updated_at": "2026-10-02T22:43:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04116v1"}, "authors": ["Yadong Lu", "Theodore Lee", "Yifei Li", "Lawrence Keunho Jang", "Tianci Xue", "Yu Su", "Huan Sun", "Ahmed Hassan Awadallah"], "canonical_source_url": "https://arxiv.org/abs/2610.04116", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-04116/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "97af38af8647b769c129d917b65f8f744035308f6ace88706a4888a96fe345b9", "submitted_at": "2026-10-02T22:43:06Z", "updated_at": "2026-10-02T22:43:06Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04116v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2610.04116"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.04116"}], "name": "CUAWright: A Minimal Unified Interface for Digital Agents", "publication": null, "published_at": "2026-10-02", "record_sha256": "293e7d643f84898576b81549c913c6714b8b31736c422f369fbe8610fff4ac63", "relations": [], "resource_id": "arxiv:2610.04116", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.04116v1", "topics": ["tool-integration", "efficiency"], "updated_at": "2026-10-02"}, "text": "CUAWright: A Minimal Unified Interface for Digital Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nStudies a minimal terminal-based execution framework that lets digital agents construct tools and manage context across browser, desktop, and CAD tasks.\nPlatforms: desktop, web\nTopics: tool-integration, efficiency\nPublished: 2026-10-02\nAuthors: Yadong Lu, Theodore Lee, Yifei Li, Lawrence Keunho Jang, Tianci Xue, Yu Su, Huan Sun, Ahmed Hassan Awadallah\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.04116v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-04116/"}
{"id": "arxiv:2610.04800", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "baacd8b4fad7b146bb9152a54d60a617fdbed5d0142c0c0768dacea2e7f18e60", "submitted_at": "2026-10-03T22:56:09Z", "updated_at": "2026-10-03T22:56:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04800v1"}, "authors": ["Ziyang Long", "Xinqi Li", "Lujing Xing", "Hsin-Jung Yang"], "canonical_source_url": "https://arxiv.org/abs/2610.04800", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-04800/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "baacd8b4fad7b146bb9152a54d60a617fdbed5d0142c0c0768dacea2e7f18e60", "submitted_at": "2026-10-03T22:56:09Z", "updated_at": "2026-10-03T22:56:09Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.04800v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.04800"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.04800"}], "name": "MedImageOSWorld: Benchmarking GUI Agents for Medical Image Consoles", "publication": null, "published_at": "2026-10-03", "record_sha256": "3b741900cbaf735373a247082acb97b1fecba86bad3b1cff352265b5b466b037", "relations": [], "resource_id": "arxiv:2610.04800", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.04800v1", "topics": ["evaluation", "science", "planning"], "updated_at": "2026-10-03"}, "text": "MedImageOSWorld: Benchmarking GUI Agents for Medical Image Consoles\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates GUI agents in simulated medical-imaging consoles, separately checking workflow completion and acquisition outcomes against hidden anatomical targets.\nPlatforms: desktop\nTopics: evaluation, science, planning\nPublished: 2026-10-03\nAuthors: Ziyang Long, Xinqi Li, Lujing Xing, Hsin-Jung Yang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.04800v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-04800/"}
{"id": "arxiv:2610.05185", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d20eb090847d25deb8f69181caa6f964661bbf167223b5572cf3e1cafa52bbd", "submitted_at": "2026-10-04T12:49:01Z", "updated_at": "2026-10-04T12:49:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05185v1"}, "authors": ["Kaiyu Wu", "Beichen Zheng", "Weiyao Huang", "Keze Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.05185", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-05185/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d20eb090847d25deb8f69181caa6f964661bbf167223b5572cf3e1cafa52bbd", "submitted_at": "2026-10-04T12:49:01Z", "updated_at": "2026-10-04T12:49:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05185v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.05185"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05185"}], "name": "Recurrent Latent Visual Search for GUI Grounding", "publication": null, "published_at": "2026-10-04", "record_sha256": "840ddff010982b4b3fcc018770087274289bb24fb3ee2b510dfef1cf479f7d50", "relations": [], "resource_id": "arxiv:2610.05185", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.05185v1", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": "2026-10-04"}, "text": "Recurrent Latent Visual Search for GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ReLaViS, which performs recurrent visual search inside a grounding model by repeatedly querying screenshot tokens and feeding the resulting spatial evidence back into the model. Coarse-to-fine supervision is constructed from existing interface-element annotations.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, efficiency\nPublished: 2026-10-04\nAuthors: Kaiyu Wu, Beichen Zheng, Weiyao Huang, Keze Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.05185v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-05185/"}
{"id": "arxiv:2610.05437", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9e40e7a3404f7aef992270e54986e655995ae88dd0c50db55ccd7f61cc90fb3", "submitted_at": "2026-10-04T18:19:33Z", "updated_at": "2026-10-06T03:17:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.05437v2"}, "authors": ["Justin Chih-Yao Chen", "Elias Stengel-Eskin", "Yan Chen", "Pol Llado", "Scott Counts", "Mohit Bansal", "Benjamin Van Durme", "Harsh Jhamtani", "Gaurav Verma"], "canonical_source_url": "https://arxiv.org/abs/2610.05437", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-05437/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9e40e7a3404f7aef992270e54986e655995ae88dd0c50db55ccd7f61cc90fb3", "submitted_at": "2026-10-04T18:19:33Z", "updated_at": "2026-10-06T03:17:00Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.05437v2"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:20Z", "role": "primary", "sha256": "c420ee40cc87a574ae2554556843f27f2eda43af553f88e95affd6319b1dae33", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2609.34321%2C2609.34769%2C2609.34974%2C2609.35026%2C2609.35671%2C2609.35814%2C2609.35904%2C2609.36201%2C2609.36488%2C2609.36923%2C2609.36927%2C2609.36939%2C2609.37053%2C2609.37176%2C2609.37686%2C2609.38008%2C2609.38397%2C2609.38778%2C2609.38867%2C2609.39306%2C2609.39547%2C2609.39903%2C2609.40253%2C2609.40284%2C2610.00948%2C2610.01215%2C2610.01491%2C2610.01787%2C2610.02320%2C2610.02932%2C2610.03036%2C2610.03089%2C2610.03574%2C2610.03818%2C2610.04116%2C2610.04800%2C2610.04906%2C2610.05185%2C2610.05437%2C2610.05861%2C2610.06829%2C2610.07118%2C2610.07127%2C2610.07444%2C2610.07972%2C2610.08699%2C2610.08773%2C2610.09240%2C2610.09469%2C2610.11050&max_results=50"}], "identifiers": {"arxiv": "2610.05437"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05437"}], "name": "TeleTune: Evolving Agent Skills From Offline Telemetry", "publication": null, "published_at": "2026-10-04", "record_sha256": "aa5bcfdad5ac945ef0fcc2f9ffccd70bb97620634a21e2e04eab860abeb519cd", "relations": [], "resource_id": "arxiv:2610.05437", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.05437v2", "topics": ["memory", "workflow", "efficiency"], "updated_at": "2026-10-06"}, "text": "TeleTune: Evolving Agent Skills From Offline Telemetry\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces TeleTune, which edits a textual skill library using held-out action prediction on offline software-use logs with missing goals or interleaved tasks. The learned workflows also guide retrieval of demonstrations for subsequent web-agent execution.\nPlatforms: web\nTopics: memory, workflow, efficiency\nPublished: 2026-10-04\nAuthors: Justin Chih-Yao Chen, Elias Stengel-Eskin, Yan Chen, Pol Llado, Scott Counts, Mohit Bansal, Benjamin Van Durme, Harsh Jhamtani, Gaurav Verma\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.05437v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-05437/"}
{"id": "arxiv:2610.05861", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43845ad64843a0b019c1988d4ba861eccee3c48a7b2c09d0bae7794e5e86347a", "submitted_at": "2026-10-05T06:28:58Z", "updated_at": "2026-10-05T06:28:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05861v1"}, "authors": ["Yongxin Ning", "Runliang Niu", "Qianli Xing", "Zhiyi Duan", "Qingzu He", "Pan Wang", "Qi Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.05861", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-05861/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "43845ad64843a0b019c1988d4ba861eccee3c48a7b2c09d0bae7794e5e86347a", "submitted_at": "2026-10-05T06:28:58Z", "updated_at": "2026-10-05T06:28:58Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.05861v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.05861"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.05861"}], "name": "Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training", "publication": null, "published_at": "2026-10-05", "record_sha256": "48b3ce5782e18ec1dba9eecbdfa5846c9a3035b2ed50e35e3b55af5a869cf0ce", "relations": [], "resource_id": "arxiv:2610.05861", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.05861v1", "topics": ["world-models", "data-generation", "mobile-navigation"], "updated_at": "2026-10-05"}, "text": "Imagine to Act: High-Fidelity Data Synthesis via Image Editing World Model for Scalable GUI Agent Training\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Infinite-Dreamer, which learns action-conditioned screenshot editing from structured descriptions of GUI changes and uses it to synthesize robustness examples and multi-step trajectories. The synthetic data trains a mobile GUI agent evaluated on AndroidWorld, MobileWorld, and AndroidControl-Curated.\nPlatforms: android\nTopics: world-models, data-generation, mobile-navigation\nPublished: 2026-10-05\nAuthors: Yongxin Ning, Runliang Niu, Qianli Xing, Zhiyi Duan, Qingzu He, Pan Wang, Qi Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.05861v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-05861/"}
{"id": "arxiv:2610.06829", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f3e9860574951cd6b749d47da748e4aec36d4f5afe226f4a3d0d5459fd07935", "submitted_at": "2026-10-05T17:57:51Z", "updated_at": "2026-10-05T17:57:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06829v1"}, "authors": ["Yifan Zhang", "Yutong Dai", "Viraj Prabhu", "Zhiyuan Hu", "Ran Xu", "Zeyuan Chen"], "canonical_source_url": "https://arxiv.org/abs/2610.06829", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-06829/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2f3e9860574951cd6b749d47da748e4aec36d4f5afe226f4a3d0d5459fd07935", "submitted_at": "2026-10-05T17:57:51Z", "updated_at": "2026-10-05T17:57:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06829v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.06829"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.06829"}], "name": "CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling", "publication": null, "published_at": "2026-10-05", "record_sha256": "2820f1402edf18acb05e286e07e8dc5675713de96220efad12a6e61957820816", "relations": [], "resource_id": "arxiv:2610.06829", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.06829v1", "topics": ["reinforcement-learning", "evaluation", "web-navigation"], "updated_at": "2026-10-05"}, "text": "CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CLIFT, which certifies agent-generated verification signals against training-time judges and reuses them for per-step rewards and trajectory selection. At inference, a frozen question bank helps compare candidate web rollouts without querying an external judge.\nPlatforms: web\nTopics: reinforcement-learning, evaluation, web-navigation\nPublished: 2026-10-05\nAuthors: Yifan Zhang, Yutong Dai, Viraj Prabhu, Zhiyuan Hu, Ran Xu, Zeyuan Chen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.06829v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-06829/"}
{"id": "arxiv:2610.06898", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5afea9414462386ee921b9ead9381c0204963c710c9f616a96f4bc47079f059e", "submitted_at": "2026-09-29T03:03:00Z", "updated_at": "2026-09-29T03:03:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06898v1"}, "authors": ["Li Hu", "Kanghua Mo", "Yingbin Jin", "Qingqing Ye", "Haibo Hu"], "canonical_source_url": "https://arxiv.org/abs/2610.06898", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-06898/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5afea9414462386ee921b9ead9381c0204963c710c9f616a96f4bc47079f059e", "submitted_at": "2026-09-29T03:03:00Z", "updated_at": "2026-09-29T03:03:00Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.06898v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:00:31.520084+00:00", "role": "primary", "sha256": "f053c4c65f6443acbda26e25cf36f7b6e8ea45f2916535604229f7c4e62ed679", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22UI+agent%22+OR+ti%3A%22desktop+agent%22+OR+ti%3A%22screen+grounding%22%29+AND+submittedDate%3A%5B200001010000+TO+202610180000%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:10:35.098754+00:00", "role": "primary", "sha256": "73f83c27e58e46086dbd5f32f108df2c5faa9ab2bd968d48062f39d94b2cc14b", "supports": ["summary", "scope", "environments", "notes"], "url": "https://arxiv.org/pdf/2610.06898v1"}], "identifiers": {"arxiv": "2610.06898"}, "kind": "paper", "last_verified_at": "2026-10-18T10:00:31.520084+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.06898"}], "name": "DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections", "publication": null, "published_at": "2026-09-29", "record_sha256": "1e8f461847e13fc8316885ff2cfcb67e4a22538e159ee5c3635fe32cb9beb15b", "relations": [], "resource_id": "arxiv:2610.06898", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.06898v1", "topics": ["evaluation", "safety-security", "mobile-navigation"], "updated_at": "2026-09-29"}, "text": "DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates whether deceptive UI content steers mobile agents away from instruction-consistent choices, using paired benign and injected decision states and decision-integrity metrics.\nPlatforms: android\nTopics: evaluation, safety-security, mobile-navigation\nPublished: 2026-09-29\nAuthors: Li Hu, Kanghua Mo, Yingbin Jin, Qingqing Ye, Haibo Hu\nRecorded arXiv version: v1\n\nThe commercial-app protocol uses fixed screenshots and context for selection decisions, rather than live end-to-end transactions. Android is supported by the paper's emulator setup. Completion and integrity are distinct measures; no reported scores were reproduced. Basis: complete retained abstract and source-reviewed sections of the exact-version paper (3.1 Benchmark components; 3.2 Instance generation; 3.3 Paired protocol/metrics; 5 Limitations/Ethics; Appendix B (Android emulator setup)); not an entire-paper reading or reproduction claim.\n\nPrimary source: https://arxiv.org/abs/2610.06898v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-06898/"}
{"id": "arxiv:2610.07118", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4890856d5961395a9ac8c83c60b988da4adeaab1d99edd511e4046e4f99fd25d", "submitted_at": "2026-10-05T17:11:07Z", "updated_at": "2026-10-05T17:11:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07118v1"}, "authors": ["Chinmay Savadikar", "Zhaoyu Zhang", "Mingyu Zhao", "Shuang Xie", "Han Li", "Tianfu Wu", "Lingyun Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.07118", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-07118/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4890856d5961395a9ac8c83c60b988da4adeaab1d99edd511e4046e4f99fd25d", "submitted_at": "2026-10-05T17:11:07Z", "updated_at": "2026-10-05T17:11:07Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07118v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.07118"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07118"}], "name": "AMBER: Training Long-Horizon Web Agents through Append-Only Memory", "publication": null, "published_at": "2026-10-05", "record_sha256": "40d3e0a766e06404b24a014d754368cdba71c79a3458e9db07445160fbd2f331", "relations": [], "resource_id": "arxiv:2610.07118", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.07118v1", "topics": ["memory", "long-horizon", "reinforcement-learning"], "updated_at": "2026-10-05"}, "text": "AMBER: Training Long-Horizon Web Agents through Append-Only Memory\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AMBER, an append-only memory framework in which a web agent learns to reason, act, and retain evidence using outcome-reward reinforcement learning. It studies whether preserving entries avoids the loss of facts and corrective feedback seen with rewritten memory.\nPlatforms: web\nTopics: memory, long-horizon, reinforcement-learning\nPublished: 2026-10-05\nAuthors: Chinmay Savadikar, Zhaoyu Zhang, Mingyu Zhao, Shuang Xie, Han Li, Tianfu Wu, Lingyun Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.07118v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-07118/"}
{"id": "arxiv:2610.07444", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81342daf25b0fb26f6481fcf836c3d958e26c2358be96eadb51e6914ea87e43c", "submitted_at": "2026-10-05T21:50:50Z", "updated_at": "2026-10-05T21:50:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07444v1"}, "authors": ["Aadi Chauhan", "Arthur Ilyasov"], "canonical_source_url": "https://arxiv.org/abs/2610.07444", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-07444/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "81342daf25b0fb26f6481fcf836c3d958e26c2358be96eadb51e6914ea87e43c", "submitted_at": "2026-10-05T21:50:50Z", "updated_at": "2026-10-05T21:50:50Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07444v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.07444"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07444"}], "name": "Decoupling What from Where: How Should a Small GUI Grounding Model Receive the Action Type?", "publication": null, "published_at": "2026-10-05", "record_sha256": "a1cf39bcf91d0dade35dbe45540140fa51b2764f710576a00658050d81e37ea8", "relations": [], "resource_id": "arxiv:2610.07444", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.07444v1", "topics": ["grounding", "evaluation", "robustness"], "updated_at": "2026-10-05"}, "text": "Decoupling What from Where: How Should a Small GUI Grounding Model Receive the Action Type?\nResource type: paper\nReview status: curated\nScope: core\n\nCompares ways to condition a small GUI grounding model on action type under matched training conditions on Android in the Wild. The study traces apparent gains to a coordinate-serialization choice and reports limited evidence that the gains persist with predicted action types.\nPlatforms: android\nTopics: grounding, evaluation, robustness\nPublished: 2026-10-05\nAuthors: Aadi Chauhan, Arthur Ilyasov\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.07444v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-07444/"}
{"id": "arxiv:2610.07972", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b792cc0355a0947c61014c471c482aef173f761f7396a939c14816ac644abe66", "submitted_at": "2026-10-06T08:38:30Z", "updated_at": "2026-10-06T08:38:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07972v1"}, "authors": ["Yeji Park", "Jaeyun Shim", "Taesik Gong"], "canonical_source_url": "https://arxiv.org/abs/2610.07972", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-07972/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b792cc0355a0947c61014c471c482aef173f761f7396a939c14816ac644abe66", "submitted_at": "2026-10-06T08:38:30Z", "updated_at": "2026-10-06T08:38:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.07972v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.07972"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.07972"}], "name": "Can Agents Work for Everyone? Cross-User Reliability for Mobile GUI Agents in Personalized User Interfaces", "publication": null, "published_at": "2026-10-06", "record_sha256": "9bc235588942e27fe52e4c008eda095420d7ce75f0a62e50c036aab150dae5ab", "relations": [], "resource_id": "arxiv:2610.07972", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.07972v1", "topics": ["evaluation", "robustness", "mobile-navigation", "reinforcement-learning"], "updated_at": "2026-10-06"}, "text": "Can Agents Work for Everyone? Cross-User Reliability for Mobile GUI Agents in Personalized User Interfaces\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PAIR for controlled evaluation of mobile GUI agents across user-conditioned application states and RePAIR for training from cross-user subgoal outcomes. The study examines failures when personalized interfaces change which items an agent must identify.\nPlatforms: unspecified\nTopics: evaluation, robustness, mobile-navigation, reinforcement-learning\nPublished: 2026-10-06\nAuthors: Yeji Park, Jaeyun Shim, Taesik Gong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.07972v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-07972/"}
{"id": "arxiv:2610.08773", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7113025e0dfd784d6994108c24681000625b13598e33ad0e4324e90fd2832c19", "submitted_at": "2026-10-06T17:56:43Z", "updated_at": "2026-10-06T17:56:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.08773v1"}, "authors": ["Sarim Hashmi", "Mukul Ranjan", "Kshitij Mishra", "Mikhail Kuznetsov", "Praneeth Vepakomma", "Nils Lukas"], "canonical_source_url": "https://arxiv.org/abs/2610.08773", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-08773/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7113025e0dfd784d6994108c24681000625b13598e33ad0e4324e90fd2832c19", "submitted_at": "2026-10-06T17:56:43Z", "updated_at": "2026-10-06T17:56:43Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.08773v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.08773"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.08773"}], "name": "AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model", "publication": null, "published_at": "2026-10-06", "record_sha256": "fc4d2859c4744d2cc798d2bea27e4857a0cf8e1e17257d29f2d7b7ae26681065", "relations": [], "resource_id": "arxiv:2610.08773", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.08773v1", "topics": ["safety-security", "world-models", "reinforcement-learning"], "updated_at": "2026-10-06"}, "text": "AdvSim2Real : Training Web Agents Against Adaptive Prompt Injection in a Web World Model\nResource type: paper\nReview status: curated\nScope: core\n\nPresents AdvSim2Real, which jointly adapts web-task curricula, prompt-injection adversaries, and agent policies inside a frozen web world model. It evaluates robustness against an unseen adversary and transfer from simulated interaction to a real browser.\nPlatforms: web\nTopics: safety-security, world-models, reinforcement-learning\nPublished: 2026-10-06\nAuthors: Sarim Hashmi, Mukul Ranjan, Kshitij Mishra, Mikhail Kuznetsov, Praneeth Vepakomma, Nils Lukas\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.08773v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-08773/"}
{"id": "arxiv:2610.09240", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b6ef7e785ad509ebd13fdb3da97fc583ee4c28dd42b4eab842c6cd88d7ef51f", "submitted_at": "2026-10-07T00:05:53Z", "updated_at": "2026-10-07T00:05:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09240v1"}, "authors": ["Wanjing Han", "Levi Taiji Li", "Mu Zhang", "Yue Jiang", "Guanhong Tao"], "canonical_source_url": "https://arxiv.org/abs/2610.09240", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-09240/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3b6ef7e785ad509ebd13fdb3da97fc583ee4c28dd42b4eab842c6cd88d7ef51f", "submitted_at": "2026-10-07T00:05:53Z", "updated_at": "2026-10-07T00:05:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09240v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.09240"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09240"}], "name": "Adversarial Images Hijack Web Agents from Visual Grounding to Browser Execution", "publication": null, "published_at": "2026-10-07", "record_sha256": "d3e5672f35cb77892ba7242306ce3b9c8c0e5f238aa53fa42fc7a1e4323339bc", "relations": [], "resource_id": "arxiv:2610.09240", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.09240v1", "topics": ["safety-security", "grounding", "evaluation"], "updated_at": "2026-10-07"}, "text": "Adversarial Images Hijack Web Agents from Visual Grounding to Browser Execution\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebMirage, a visual red-teaming framework that models webpage element competition and action post-processing to test whether manipulated grounding decisions lead to executed browser actions.\nPlatforms: web\nTopics: safety-security, grounding, evaluation\nPublished: 2026-10-07\nAuthors: Wanjing Han, Levi Taiji Li, Mu Zhang, Yue Jiang, Guanhong Tao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.09240v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-09240/"}
{"id": "arxiv:2610.09469", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f38c9e37ab7fc8356a0c8c72370de673d1c2b6294e15dfdc72c280269f43439c", "submitted_at": "2026-10-07T05:24:40Z", "updated_at": "2026-10-07T05:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09469v1"}, "authors": ["Sarthak Choudhary", "Mihai Christodorescu", "Ashish Hooda", "Somesh Jha", "Tongxin Li", "Damien Octeau"], "canonical_source_url": "https://arxiv.org/abs/2610.09469", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-09469/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f38c9e37ab7fc8356a0c8c72370de673d1c2b6294e15dfdc72c280269f43439c", "submitted_at": "2026-10-07T05:24:40Z", "updated_at": "2026-10-07T05:24:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.09469v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.09469"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09469"}], "name": "Secure-CUA: Controlling Untrusted Influence in Computer-Use Agents", "publication": null, "published_at": "2026-10-07", "record_sha256": "c96353f3d5000ddaf70e3742cd899c6ded2119ce668dada80724c9af2cdba32a", "relations": [], "resource_id": "arxiv:2610.09469", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.09469v1", "topics": ["safety-security", "web-navigation"], "updated_at": "2026-10-07"}, "text": "Secure-CUA: Controlling Untrusted Influence in Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nProposes Secure-CUA, which commits to per-action transactions before reading untrusted page content and separates content queries from GUI target selection. The paper formalizes its protection assumptions and evaluates benign task utility on WebArena.\nPlatforms: web\nTopics: safety-security, web-navigation\nPublished: 2026-10-07\nAuthors: Sarthak Choudhary, Mihai Christodorescu, Ashish Hooda, Somesh Jha, Tongxin Li, Damien Octeau\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.09469v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-09469/"}
{"id": "arxiv:2610.09944", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03aea09f3ed1b75a8a6584efdfa51d45d1cfa9dc53fb3a25efd820c5b72e094a", "submitted_at": "2026-10-07T12:26:24Z", "updated_at": "2026-10-08T09:06:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.09944v2"}, "authors": ["Michael Ofengenden", "Maksym Andriushchenko"], "canonical_source_url": "https://arxiv.org/abs/2610.09944", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-09944/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "03aea09f3ed1b75a8a6584efdfa51d45d1cfa9dc53fb3a25efd820c5b72e094a", "submitted_at": "2026-10-07T12:26:24Z", "updated_at": "2026-10-08T09:06:36Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.09944v2"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:56.488063+00:00", "role": "primary", "sha256": "5424a5f9278ef276c84799a08d4e2e171ee93947a38c1d6670c36fa76dc84fe3", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.09944v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:15:05.368243+00:00", "role": "primary", "sha256": "9bef76f95b41a7ca39c6fa3276d6122e4b27e1e0bd56a777edcfd7e11e7d7342", "supports": ["environments", "notes"], "url": "https://github.com/michaelofengenden/agenttimebench"}], "identifiers": {"arxiv": "2610.09944"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:56.488063+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.09944"}], "name": "AgentTime: Can Agents Estimate and Control Their Own Runtime?", "publication": null, "published_at": "2026-10-07", "record_sha256": "d23546f52b01b5d0afd664ddef166527eb90860eaa7ef9205b8a52c20c9a9878", "relations": [], "resource_id": "arxiv:2610.09944", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.09944v2", "topics": ["evaluation", "observability"], "updated_at": "2026-10-08"}, "text": "AgentTime: Can Agents Estimate and Control Their Own Runtime?\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates duration following, runtime forecasts and elapsed-time estimates in native agent harnesses, including OSWorld desktop tasks, with external timers, native graders and work-versus-wait analysis.\nPlatforms: desktop\nTopics: evaluation, observability\nPublished: 2026-10-07\nAuthors: Michael Ofengenden, Maksym Andriushchenko\nRecorded arXiv version: v2\n\nThis cross-domain suite includes 16 OSWorld 2.0 tasks among 222 tasks from 18 sources. A single run per task and requested duration, provider effects and incomplete native grades limit comparisons; elapsed time is not evidence of productive work. The OSWorld task 092 video judge was skipped without an API key. The inspected repository README identifies desktop-VM timing and states that the current 222-task forecasting runs will be added; its existing forecasting code covers an earlier study, and transcript work/wait labels are not public. No code or benchmark was executed and no license was verified.\n\nPrimary source: https://arxiv.org/abs/2610.09944v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-09944/"}
{"id": "arxiv:2610.11050", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37413f11bf8263f1516a8edca8c594227fcab104c0c767a7daa8eead9645ec34", "submitted_at": "2026-10-08T01:03:48Z", "updated_at": "2026-10-08T01:03:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11050v1"}, "authors": ["Xing Han Lù", "Dheeraj Vattikonda", "Sina Hajimiri", "Fatemeh Pesaran Zadeh", "Parishad BehnamGhader", "Ghazwa Darwiche", "Amirhossein Kazemnejad", "Christopher Pal", "Alexandre Drouin", "Siva Reddy"], "canonical_source_url": "https://arxiv.org/abs/2610.11050", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-11050/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "37413f11bf8263f1516a8edca8c594227fcab104c0c767a7daa8eead9645ec34", "submitted_at": "2026-10-08T01:03:48Z", "updated_at": "2026-10-08T01:03:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11050v1"}}, "environments": ["cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.11050"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11050"}], "name": "AgentHorizon: Evaluating Agentic Judges for Long-Horizon Computer-Use Tasks", "publication": null, "published_at": "2026-10-08", "record_sha256": "6d08414415bade5d8748687df5e5faffe120ea4f6c20aeb42e056fce121347f1", "relations": [], "resource_id": "arxiv:2610.11050", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.11050v1", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": "2026-10-08"}, "text": "AgentHorizon: Evaluating Agentic Judges for Long-Horizon Computer-Use Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces AgentHorizon, a benchmark for judging long computer-use trajectories using closely related instructions paired with recorded actions and screenshots. It tests whether judges can detect constraint violations and distinguish successful tasks from similar but incompatible requests.\nPlatforms: cross-platform\nTopics: evaluation, long-horizon, cross-app\nPublished: 2026-10-08\nAuthors: Xing Han Lù, Dheeraj Vattikonda, Sina Hajimiri, Fatemeh Pesaran Zadeh, Parishad BehnamGhader, Ghazwa Darwiche, Amirhossein Kazemnejad, Christopher Pal, Alexandre Drouin, Siva Reddy\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.11050v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-11050/"}
{"id": "arxiv:2610.11707", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2627960a5698d4c2f245e91ca5033be8c09e8a1ddea02490fcc1ed9f3d42217d", "submitted_at": "2026-10-08T11:12:57Z", "updated_at": "2026-10-08T11:12:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11707v1"}, "authors": ["Yudong Bai", "Yihong Chen", "Quanming Yao", "Yaqing Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.11707", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-11707/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2627960a5698d4c2f245e91ca5033be8c09e8a1ddea02490fcc1ed9f3d42217d", "submitted_at": "2026-10-08T11:12:57Z", "updated_at": "2026-10-08T11:12:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11707v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.11707"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11707"}], "name": "DeltaReplay: Task-Relative Memory Reuse for Mobile GUI Agents", "publication": null, "published_at": "2026-10-08", "record_sha256": "b0e24a1387d9817a7ddd9ea36a11f8d055a578676e79eab09cdf56dd78185215", "relations": [], "resource_id": "arxiv:2610.11707", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.11707v1", "topics": ["memory", "mobile-navigation", "planning"], "updated_at": "2026-10-08"}, "text": "DeltaReplay: Task-Relative Memory Reuse for Mobile GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces DeltaReplay, which stores mobile GUI trajectories as transition graphs and decides at each step whether to reuse an action, substitute task-specific parameters, or defer to the base agent. This supports reuse of partially matching experiences without rewriting stored memory.\nPlatforms: android\nTopics: memory, mobile-navigation, planning\nPublished: 2026-10-08\nAuthors: Yudong Bai, Yihong Chen, Quanming Yao, Yaqing Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.11707v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-11707/"}
{"id": "arxiv:2610.11942", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53120f960c7b98e87210c93ac661a2e397b858cc9f211786b0f8233fa61c6dd3", "submitted_at": "2026-10-08T13:33:23Z", "updated_at": "2026-10-08T13:33:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11942v1"}, "authors": ["Jiaming Zhang", "Xuan Wang", "Fuyao Zhang", "Yang Cao", "Lingjuan Lyu", "Wei Yang Bryan Lim"], "canonical_source_url": "https://arxiv.org/abs/2610.11942", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-11942/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "53120f960c7b98e87210c93ac661a2e397b858cc9f211786b0f8233fa61c6dd3", "submitted_at": "2026-10-08T13:33:23Z", "updated_at": "2026-10-08T13:33:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11942v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.11942"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11942"}], "name": "Right Screen, Wrong Transition: World Models as Verifiers for GUI Agents", "publication": null, "published_at": "2026-10-08", "record_sha256": "1728b3229b5cdf2f5985ee182c5247e2305f80bcb445266ded232afe5fdec8ca", "relations": [], "resource_id": "arxiv:2610.11942", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.11942v1", "topics": ["world-models", "safety-security", "evaluation"], "updated_at": "2026-10-08"}, "text": "Right Screen, Wrong Transition: World Models as Verifiers for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nPresents LGWM, a decoder-free GUI world model that predicts the next screen representation from an action and compares it with the observed state to verify transitions. Its paired diagnostic separates legitimate and hijacked transitions that lead to the same credential screen.\nPlatforms: unspecified\nTopics: world-models, safety-security, evaluation\nPublished: 2026-10-08\nAuthors: Jiaming Zhang, Xuan Wang, Fuyao Zhang, Yang Cao, Lingjuan Lyu, Wei Yang Bryan Lim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.11942v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-11942/"}
{"id": "arxiv:2610.11963", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e77dcf101ba95684aac51084e43b914c0aca86fc25d061612224118004bdb21", "submitted_at": "2026-10-08T13:44:53Z", "updated_at": "2026-10-08T13:44:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11963v1"}, "authors": ["Utku Boran Torun", "Veli Karakaya", "Eray Tüzün"], "canonical_source_url": "https://arxiv.org/abs/2610.11963", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-11963/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e77dcf101ba95684aac51084e43b914c0aca86fc25d061612224118004bdb21", "submitted_at": "2026-10-08T13:44:53Z", "updated_at": "2026-10-08T13:44:53Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.11963v1"}}, "environments": ["web", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:25.581656+00:00", "role": "primary", "sha256": "9fad44164a6f00446d2c91a31e3e77a4aab8145ac3747e9d1ce0165c1aebda95", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.11963v1"}], "identifiers": {"arxiv": "2610.11963"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:25.581656+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.11963"}], "name": "Can LLMs Fix It Without Code? Toward Automated Verification of No-Code Bug Fixes", "publication": null, "published_at": "2026-10-08", "record_sha256": "b37515ee5680cee13d3708d4d494a9c89d794cac3367c3be09b64fdbf5d172d5", "relations": [], "resource_id": "arxiv:2610.11963", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.11963v1", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-10-08"}, "text": "Can LLMs Fix It Without Code? Toward Automated Verification of No-Code Bug Fixes\nResource type: paper\nReview status: curated\nScope: core\n\nTests natural-language no-code fixes through GUI execution in Linux Brave. Scripted fixed/buggy-state gates and issue-specific checkers distinguish fix outcomes from the executor's completion claims.\nPlatforms: web, desktop, linux\nTopics: evaluation, web-navigation\nPublished: 2026-10-08\nAuthors: Utku Boran Torun, Veli Karakaya, Eray Tüzün\nRecorded arXiv version: v1\n\nThe evaluation reuses 322 fixes from 12 generation configurations and is restricted to nine of 51 candidate issues after environment setup and sanity checks. Resolution counts any success in three runs with a 20-step episode limit; executor harnesses and decoding settings differ. Opening the update page triggers a harness-controlled switch to the fixed browser build. Checker behavior, reconstructed starting states and the two-author human reference limit interpretation. The replication package was linked by the paper but was not inspected; no software, agent, model or result was run or reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.11963v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-11963/"}
{"id": "arxiv:2610.12611", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "06b25b25e804683e11305e21dd04feb9131aad27f901998eda70ac91e527c6b9", "submitted_at": "2026-10-08T18:31:16Z", "updated_at": "2026-10-14T08:40:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.12611v2"}, "authors": ["Tian Jin", "Wenhao Wang", "Jing Kang", "Haoting Shi", "Rui Ye", "Yijun Lu", "Siheng Chen"], "canonical_source_url": "https://arxiv.org/abs/2610.12611", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-12611/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "06b25b25e804683e11305e21dd04feb9131aad27f901998eda70ac91e527c6b9", "submitted_at": "2026-10-08T18:31:16Z", "updated_at": "2026-10-14T08:40:10Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.12611v2"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.12611"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12611"}], "name": "SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments", "publication": null, "published_at": "2026-10-08", "record_sha256": "cbe687f0eb4b6d56ef8b29e38ae36ad45650019e9194d36591db189780b854e1", "relations": [], "resource_id": "arxiv:2610.12611", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.12611v2", "topics": ["data-generation", "evaluation", "sandbox"], "updated_at": "2026-10-14"}, "text": "SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SoftwareGym, a pipeline that synthesizes runnable GUI applications and iteratively repairs them using execution checks. It also provides a task benchmark and studies whether training on synthesized application trajectories transfers to other computer-use environments.\nPlatforms: unspecified\nTopics: data-generation, evaluation, sandbox\nPublished: 2026-10-08\nAuthors: Tian Jin, Wenhao Wang, Jing Kang, Haoting Shi, Rui Ye, Yijun Lu, Siheng Chen\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.12611v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-12611/"}
{"id": "arxiv:2610.12823", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6f1521fdfdf6a11a2a195feb05fd915afd6e9bb2332486b870da6375214ac1a", "submitted_at": "2026-10-09T00:03:13Z", "updated_at": "2026-10-09T00:03:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12823v1"}, "authors": ["Yisheng Li", "Dongming Han", "Hongkun Pan", "Bingcheng Mao", "Zijian Wang", "Yunlong Mao", "Wei Tong", "Sheng Zhong"], "canonical_source_url": "https://arxiv.org/abs/2610.12823", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-12823/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6f1521fdfdf6a11a2a195feb05fd915afd6e9bb2332486b870da6375214ac1a", "submitted_at": "2026-10-09T00:03:13Z", "updated_at": "2026-10-09T00:03:13Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12823v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.12823"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12823"}], "name": "When Charts Deceive Web Agents: A Paired Benchmark for Misleading Visualization-Grounded Tasks", "publication": null, "published_at": "2026-10-09", "record_sha256": "42dd71133fcbdabcef410e14f4f2515f96123d4730752ce185ef4d146c19ff2a", "relations": [], "resource_id": "arxiv:2610.12823", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.12823v1", "topics": ["evaluation", "robustness", "screen-understanding"], "updated_at": "2026-10-09"}, "text": "When Charts Deceive Web Agents: A Paired Benchmark for Misleading Visualization-Grounded Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MisVisAgentBench, a paired executable benchmark that changes chart evidence while holding web tasks fixed, to trace how misleading visualizations affect agent decisions and downstream actions.\nPlatforms: web\nTopics: evaluation, robustness, screen-understanding\nPublished: 2026-10-09\nAuthors: Yisheng Li, Dongming Han, Hongkun Pan, Bingcheng Mao, Zijian Wang, Yunlong Mao, Wei Tong, Sheng Zhong\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.12823v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-12823/"}
{"id": "arxiv:2610.12866", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f7effcf5e101cfcc7e229fe4aab40d07401b1dd98cf5ce3b566fef863a42190", "submitted_at": "2026-10-09T01:50:26Z", "updated_at": "2026-10-09T01:50:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12866v1"}, "authors": ["Shangquan Sun", "Ruoyu Chen", "Zhirui Liu", "Xiang Liu", "Chenghao Qian", "Wenqi Ren", "Xiaochun Cao"], "canonical_source_url": "https://arxiv.org/abs/2610.12866", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-12866/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f7effcf5e101cfcc7e229fe4aab40d07401b1dd98cf5ce3b566fef863a42190", "submitted_at": "2026-10-09T01:50:26Z", "updated_at": "2026-10-09T01:50:26Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12866v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.12866"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12866"}], "name": "CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving", "publication": null, "published_at": "2026-10-09", "record_sha256": "04955f0b9cd4f6eb36ef3d3415b25cc2c5914a646c937df37570505934687308", "relations": [], "resource_id": "arxiv:2610.12866", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.12866v1", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": "2026-10-09"}, "text": "CaptchaAgent: Verifiable Visual Reasoning for CAPTCHA Solving\nResource type: paper\nReview status: curated\nScope: core\n\nPresents CaptchaAgent, which builds an explicit visual representation of objects and relations, refines ambiguous regions, and checks counterfactual consistency before acting on interactive CAPTCHAs.\nPlatforms: web\nTopics: grounding, screen-understanding, evaluation\nPublished: 2026-10-09\nAuthors: Shangquan Sun, Ruoyu Chen, Zhirui Liu, Xiang Liu, Chenghao Qian, Wenqi Ren, Xiaochun Cao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.12866v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-12866/"}
{"id": "arxiv:2610.12984", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e90bb713e7034a64821502eb6fc792fa91a5d1e420aff9ff658506d39269f03", "submitted_at": "2026-10-09T04:32:02Z", "updated_at": "2026-10-09T04:32:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12984v1"}, "authors": ["Yuqi Huang", "Hui Li", "Yifei Cao", "Zikun Cai", "Tianci Xie", "Yangfan Gao", "Pinxue Guo", "Chaoyun Zhang", "Bingxuan Li", "Rui Zheng", "Hang Yan", "Ning Liao"], "canonical_source_url": "https://arxiv.org/abs/2610.12984", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-12984/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0e90bb713e7034a64821502eb6fc792fa91a5d1e420aff9ff658506d39269f03", "submitted_at": "2026-10-09T04:32:02Z", "updated_at": "2026-10-09T04:32:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.12984v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.12984"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.12984"}], "name": "Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction", "publication": null, "published_at": "2026-10-09", "record_sha256": "70cf3ce76c30839af419a1e7fbb08d58d5eab2de86eedb365db54ee240e2cd14", "relations": [], "resource_id": "arxiv:2610.12984", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.12984v1", "topics": ["recording", "workflow", "data-generation"], "updated_at": "2026-10-09"}, "text": "Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction\nResource type: paper\nReview status: curated\nScope: core\n\nAligns human screen recordings, input events, and accessibility information to derive reusable workflow skills from desktop demonstrations.\nPlatforms: desktop\nTopics: recording, workflow, data-generation\nPublished: 2026-10-09\nAuthors: Yuqi Huang, Hui Li, Yifei Cao, Zikun Cai, Tianci Xie, Yangfan Gao, Pinxue Guo, Chaoyun Zhang, Bingxuan Li, Rui Zheng, Hang Yan, Ning Liao\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.12984v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-12984/"}
{"id": "arxiv:2610.13092", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb5f0220266a43b74e7a5f547abbe78a54a2ef993fee5d27f67827cb9715f5dc", "submitted_at": "2026-10-09T06:45:23Z", "updated_at": "2026-10-09T06:45:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13092v1"}, "authors": ["Yuchen Ling", "Shengcheng Yu", "Zhenyu Chen", "Chunrong Fang"], "canonical_source_url": "https://arxiv.org/abs/2610.13092", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-13092/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "fb5f0220266a43b74e7a5f547abbe78a54a2ef993fee5d27f67827cb9715f5dc", "submitted_at": "2026-10-09T06:45:23Z", "updated_at": "2026-10-09T06:45:23Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13092v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:54Z", "role": "primary", "sha256": "3db7474953cbe5a3052fb22793038f2383bdd0e7c775af06494a5e8501646d18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.13092v1"}], "identifiers": {"arxiv": "2610.13092"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13092"}], "name": "When Faster Decisions Do Not Make Better Agents: An Empirical Study of Jev in Agent Workflows", "publication": null, "published_at": "2026-10-09", "record_sha256": "74673fb5cd8d0e95570e8f63d786d3c5b17e1cda6564aba31f16dfe17d99a49c", "relations": [], "resource_id": "arxiv:2610.13092", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.13092v1", "topics": ["efficiency", "action-prediction", "evaluation", "mobile-navigation", "self-reflection"], "updated_at": "2026-10-09"}, "text": "When Faster Decisions Do Not Make Better Agents: An Empirical Study of Jev in Agent Workflows\nResource type: paper\nReview status: curated\nScope: core\n\nEvaluates Jev as a bounded action selector inside three AndroidWorld agents and two text-based ALFWorld agents, separating decision savings from task success, recovery and resources per delivered outcome.\nPlatforms: android\nTopics: efficiency, action-prediction, evaluation, mobile-navigation, self-reflection\nPublished: 2026-10-09\nAuthors: Yuchen Ling, Shengcheng Yu, Zhenyu Chen, Chunrong Fang\nRecorded arXiv version: v1\n\nThe mobile evaluation uses text UI observations and one fixed instance for each of 116 AndroidWorld task types; three repeats share seed 0. Five agent families use different budgets, with candidate-interface and tool-call granularity changes in some variants. Reported costs use fixed experimental prices, and outcome-selected trace labels are not population failure rates. Faster selection does not imply faster or more reliable task completion. Selected sections were source-read; no independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.13092v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-13092/"}
{"id": "arxiv:2610.13268", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010e8ebc657dccad670ea6d17985e9c3c1680017f2516eaa86b7ea9db5660990", "submitted_at": "2026-10-09T09:20:02Z", "updated_at": "2026-10-09T09:20:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13268v1"}, "authors": ["Ruotian Zhang", "Yuhui Zhang", "Junzhe Huang", "Yu Xin", "Hehe Fan", "Xiaoxiao Sun"], "canonical_source_url": "https://arxiv.org/abs/2610.13268", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-13268/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "010e8ebc657dccad670ea6d17985e9c3c1680017f2516eaa86b7ea9db5660990", "submitted_at": "2026-10-09T09:20:02Z", "updated_at": "2026-10-09T09:20:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13268v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.13268"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13268"}], "name": "ConClick: Element-Wise Contrastive Learning Improves GUI Grounding", "publication": null, "published_at": "2026-10-09", "record_sha256": "f35985ec664ce929c60e55616ddb88b6329de922807f20a51f720dc70e5834a2", "relations": [], "resource_id": "arxiv:2610.13268", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.13268v1", "topics": ["grounding", "data-generation", "robustness"], "updated_at": "2026-10-09"}, "text": "ConClick: Element-Wise Contrastive Learning Improves GUI Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nAnalyzes recurring distractor clicks in GUI grounding and introduces ConClick, a contrastive training method built from target–distractor pairs. Its ReClick inference procedure revisits candidate regions and compares competing predictions.\nPlatforms: unspecified\nTopics: grounding, data-generation, robustness\nPublished: 2026-10-09\nAuthors: Ruotian Zhang, Yuhui Zhang, Junzhe Huang, Yu Xin, Hehe Fan, Xiaoxiao Sun\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.13268v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-13268/"}
{"id": "arxiv:2610.13416", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cadcf2ab7be51b17d06975eba0e56573c333dca9ec034ade67b84e62d0096903", "submitted_at": "2026-10-09T11:23:40Z", "updated_at": "2026-10-09T11:23:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13416v1"}, "authors": ["Sunghwan Kim", "Deogyong Kim", "Sangam Lee", "Donguk Kwon", "Dongha Lee"], "canonical_source_url": "https://arxiv.org/abs/2610.13416", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-13416/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "cadcf2ab7be51b17d06975eba0e56573c333dca9ec034ade67b84e62d0096903", "submitted_at": "2026-10-09T11:23:40Z", "updated_at": "2026-10-09T11:23:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13416v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.13416"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13416"}], "name": "Web$hell: A Unified Interface for Agentic Web Access", "publication": null, "published_at": "2026-10-09", "record_sha256": "08dfd34b8d3591925d6c4c3b3a07497e68a7155a7617813b6a481081f0249118", "relations": [], "resource_id": "arxiv:2610.13416", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.13416v1", "topics": ["web-navigation", "tool-integration", "memory", "efficiency"], "updated_at": "2026-10-09"}, "text": "Web$hell: A Unified Interface for Agentic Web Access\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Web$hell, an interface that compiles live pages into persistent information and action objects so agents can gather evidence, navigate websites, and reuse acquired content through shared operations.\nPlatforms: web\nTopics: web-navigation, tool-integration, memory, efficiency\nPublished: 2026-10-09\nAuthors: Sunghwan Kim, Deogyong Kim, Sangam Lee, Donguk Kwon, Dongha Lee\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.13416v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-13416/"}
{"id": "arxiv:2610.13703", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "447a4f283746069b2ea51d49c149379ad48a6e0b71f35dd15dedcb19417ab0cd", "submitted_at": "2026-10-09T14:59:24Z", "updated_at": "2026-10-09T14:59:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13703v1"}, "authors": ["Vagner Figueredo de Santana", "Diogo Cortiz"], "canonical_source_url": "https://arxiv.org/abs/2610.13703", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-13703/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "447a4f283746069b2ea51d49c149379ad48a6e0b71f35dd15dedcb19417ab0cd", "submitted_at": "2026-10-09T14:59:24Z", "updated_at": "2026-10-09T14:59:24Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.13703v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.13703"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.13703"}], "name": "A Baseline Assessment for the Agentic Web Technology Stack", "publication": null, "published_at": "2026-10-09", "record_sha256": "f6cf66e1c1b6148e0644e3827cb010514834874468ca1d8a1a715940c2777b0c", "relations": [], "resource_id": "arxiv:2610.13703", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.13703v1", "topics": ["evaluation", "web-navigation"], "updated_at": "2026-10-09"}, "text": "A Baseline Assessment for the Agentic Web Technology Stack\nResource type: paper\nReview status: curated\nScope: core\n\nPresents MASTIF, a framework for comparing combinations of agent models, protocols, and implementation frameworks on web tasks, with an initial assessment using 300 Mind2Web tasks.\nPlatforms: web\nTopics: evaluation, web-navigation\nPublished: 2026-10-09\nAuthors: Vagner Figueredo de Santana, Diogo Cortiz\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.13703v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-13703/"}
{"id": "arxiv:2610.14157", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f7d940d3e341fd8c9f23a4829a5ab1150fc07ee231ccda9b108a3f0d785efb9", "submitted_at": "2026-10-09T17:26:38Z", "updated_at": "2026-10-09T17:26:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14157v1"}, "authors": ["Chenyang Yan", "Yingying Zhang", "Kanzhi Cheng", "Qiushi Sun", "Zhengyuan Pan", "Zheng Ma", "Hang Yan", "Yian Wang", "Nuo Chen", "Jialin Cao", "Xingdong Gong", "Wenpo Song", "Han Chen", "Zichen Ding", "Fangzhi Xu", "Shujian Huang", "Xinyu Dai", "Yichen Liu", "Tiankuo Yao", "Bo Wang", "Ben Kao", "Jianbing Zhang", "Lewei Lu", "Dahua Lin"], "canonical_source_url": "https://arxiv.org/abs/2610.14157", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-14157/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f7d940d3e341fd8c9f23a4829a5ab1150fc07ee231ccda9b108a3f0d785efb9", "submitted_at": "2026-10-09T17:26:38Z", "updated_at": "2026-10-09T17:26:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14157v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}], "identifiers": {"arxiv": "2610.14157"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.14157"}], "name": "OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools", "publication": null, "published_at": "2026-10-09", "record_sha256": "0092343d2e1980fa19087b6c5d4124d80c76a7c87e0f07a86906462b30d97439", "relations": [], "resource_id": "arxiv:2610.14157", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.14157v1", "topics": ["mobile-navigation", "tool-integration", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-09"}, "text": "OpenMobile-2: Building Versatile Mobile Agents with Scalable Environments and App-Native Tools\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces OpenMobile-2, combining simulated mobile apps, emulator tasks, released training data, and policies that can alternate GUI interaction with app-native tools.\nPlatforms: android, web\nTopics: mobile-navigation, tool-integration, data-generation, reinforcement-learning\nPublished: 2026-10-09\nAuthors: Chenyang Yan, Yingying Zhang, Kanzhi Cheng, Qiushi Sun, Zhengyuan Pan, Zheng Ma, Hang Yan, Yian Wang, Nuo Chen, Jialin Cao, Xingdong Gong, Wenpo Song, Han Chen, Zichen Ding, Fangzhi Xu, Shujian Huang, Xinyu Dai, Yichen Liu, Tiankuo Yao, Bo Wang, Ben Kao, Jianbing Zhang, Lewei Lu, Dahua Lin\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.14157v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-14157/"}
{"id": "arxiv:2610.14228", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "191e88c19124ddaddc2fed2f0e3f8e197a4c0c7f8adbc3997c056d3333754042", "submitted_at": "2026-10-09T18:45:11Z", "updated_at": "2026-10-09T18:45:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14228v1"}, "authors": ["Seungjun Yu", "Hojun Choi", "Seojeong Park", "Jaeyo Shin", "Hyunjung Shim"], "canonical_source_url": "https://arxiv.org/abs/2610.14228", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-14228/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "191e88c19124ddaddc2fed2f0e3f8e197a4c0c7f8adbc3997c056d3333754042", "submitted_at": "2026-10-09T18:45:11Z", "updated_at": "2026-10-09T18:45:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.14228v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.14228"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.14228"}], "name": "Latent World Action Model for GUI Agents", "publication": null, "published_at": "2026-10-09", "record_sha256": "3319aad671b2abec7b5d8d9649e78235241e870029cff13d1e248a6e24964f83", "relations": [], "resource_id": "arxiv:2610.14228", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.14228v1", "topics": ["world-models", "planning", "efficiency"], "updated_at": "2026-10-09"}, "text": "Latent World Action Model for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nReplaces future-screenshot generation with a shared latent representation for GUI action selection and action-conditioned transition prediction. An action-aware contrastive objective encourages the transition representation to retain discriminative action-induced changes.\nPlatforms: unspecified\nTopics: world-models, planning, efficiency\nPublished: 2026-10-09\nAuthors: Seungjun Yu, Hojun Choi, Seojeong Park, Jaeyo Shin, Hyunjung Shim\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.14228v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-14228/"}
{"id": "arxiv:2610.15138", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63e19265f1df2c336e28c4a13e11331021bd60866269c0b921d92a3c0fec080d", "submitted_at": "2026-10-10T19:30:31Z", "updated_at": "2026-10-10T19:30:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15138v1"}, "authors": ["Md Imran Hossen"], "canonical_source_url": "https://arxiv.org/abs/2610.15138", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-15138/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "63e19265f1df2c336e28c4a13e11331021bd60866269c0b921d92a3c0fec080d", "submitted_at": "2026-10-10T19:30:31Z", "updated_at": "2026-10-10T19:30:31Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15138v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.15138"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15138"}], "name": "The Agent Clicks Approve: Injected Page Text Makes Browser Agents Authorize Access the Task Never Named", "publication": null, "published_at": "2026-10-10", "record_sha256": "84690570c17617432fe8db07d64447d4fccd6426b2aef54ecd75c0d217ee2522", "relations": [], "resource_id": "arxiv:2610.15138", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.15138v1", "topics": ["safety-security", "evaluation"], "updated_at": "2026-10-10"}, "text": "The Agent Clicks Approve: Injected Page Text Makes Browser Agents Authorize Access the Task Never Named\nResource type: paper\nReview status: curated\nScope: core\n\nStudies prompt-injected account-access approvals in browser agents and proposes a service-side authority envelope that checks authorization requests against the user task before allowing them to complete.\nPlatforms: web\nTopics: safety-security, evaluation\nPublished: 2026-10-10\nAuthors: Md Imran Hossen\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.15138v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-15138/"}
{"id": "arxiv:2610.15502", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff5470e792d14ced2cdff648d97e2606eea4b3c822046b7c10272fd2a5f1695f", "submitted_at": "2026-10-11T10:09:44Z", "updated_at": "2026-10-11T10:09:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15502v1"}, "authors": ["Hyobin Ahn", "Dongjae Lee"], "canonical_source_url": "https://arxiv.org/abs/2610.15502", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-15502/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ff5470e792d14ced2cdff648d97e2606eea4b3c822046b7c10272fd2a5f1695f", "submitted_at": "2026-10-11T10:09:44Z", "updated_at": "2026-10-11T10:09:44Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15502v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:56:03.829575Z", "role": "primary", "sha256": "a8636ac679d2110181d4417d4aef040ddb24464597a7fa6fe6d2c35a2401a3b1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22graphical+user+interface%22+OR+ti%3A%22screen+grounding%22+OR+ti%3A%22UI+grounding%22+OR+ti%3A%22desktop+agents%22+OR+ti%3A%22Android+agent%22+OR+ti%3A%22phone+agent%22+OR+ti%3A%22mobile+GUI%22+OR+ti%3A%22web+navigation%22+OR+ti%3A%22GUI+automation%22+OR+ti%3A%22browser+automation%22%29+AND+submittedDate%3A%5B202608210000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:58.069698Z", "role": "primary", "sha256": "a38b3ddc9d5d6744eefc937f86e174917f9b5de149323632d5d1b20283d79c8e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.15502v1"}], "identifiers": {"arxiv": "2610.15502"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:58.069698Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15502"}], "name": "From Observations to Authorized Actions: Measuring Source Information in Android Agents", "publication": null, "published_at": "2026-10-11", "record_sha256": "a649779145ec2a297e3cde5ed016b1cc5a505241b2523ba45ce95884c8e99adc", "relations": [], "resource_id": "arxiv:2610.15502", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.15502v1", "topics": ["safety-security", "tool-integration", "observability"], "updated_at": "2026-10-11"}, "text": "From Observations to Authorized Actions: Measuring Source Information in Android Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMeasures source-attribution limits in Android accessibility and separately tests application-history and selected-WebView-object checks for authorizing AppFunctions actions.\nPlatforms: android, web\nTopics: safety-security, tool-integration, observability\nPublished: 2026-10-11\nAuthors: Hyobin Ahn, Dongjae Lee\nRecorded arXiv version: v1\n\nThe two prototypes use different workloads and trust assumptions on patched Android/WebView components. Approval is simulated; authenticated browser delivery and atomic capture remain unimplemented. Coarse source history can withhold benign work and persist across tasks. Direct-call checks and limited live-agent trials are separate, and writes are in-memory. No production vulnerability, general defense rate or released artifact is established.\n\nPrimary source: https://arxiv.org/abs/2610.15502v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-15502/"}
{"id": "arxiv:2610.15738", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6c60bb4d3b7bb2e2cba92201924d8da4f5e18ce5553c81fc9d5c6fb62b84751", "submitted_at": "2026-10-11T16:18:11Z", "updated_at": "2026-10-11T16:18:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15738v1"}, "authors": ["Jianlin Chen", "Wenhui Chen", "Chi Man Vong"], "canonical_source_url": "https://arxiv.org/abs/2610.15738", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-15738/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e6c60bb4d3b7bb2e2cba92201924d8da4f5e18ce5553c81fc9d5c6fb62b84751", "submitted_at": "2026-10-11T16:18:11Z", "updated_at": "2026-10-11T16:18:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.15738v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.15738v1"], "reason": "Publisher document declaration only; no code, data or weights license inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:10Z", "role": "primary", "sha256": "6d8c0463b6d50320246021f0394d0e39c9fd839f1ef1f742c8cdfa231872b059", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22+OR+ti%3AGUI+OR+ti%3A%22Android+agent%22+OR+abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:31:34.620763Z", "role": "primary", "sha256": "7ec947ceb462b5fd23a4a67df64e931e427224ef721b681d3fc92c1d4379449f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.15738v1"}], "identifiers": {"arxiv": "2610.15738"}, "kind": "paper", "last_verified_at": "2026-10-20T15:31:34.620763Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:31:34.620763Z", "url": "https://arxiv.org/html/2610.15738v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.15738"}], "name": "Visible but Not Avoided: State-Dependent Re-Proposal of Rejected Candidates in Catalog-Based LLM Optimization Loops", "publication": null, "published_at": "2026-10-11", "record_sha256": "eda979d8047a54ea5167a0f416ec36c7939e89574eb9edbfd2698a6b83dbf1b9", "relations": [], "resource_id": "arxiv:2610.15738", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.15738v1", "topics": ["evaluation", "grounding", "training-free"], "updated_at": "2026-10-11"}, "text": "Visible but Not Avoided: State-Dependent Re-Proposal of Rejected Candidates in Catalog-Based LLM Optimization Loops\nResource type: paper\nReview status: curated\nScope: core\n\nStudies repeated rejected edits while optimizing a screenshot-only GUI agent, contrasting static probes with reached loop states and testing how rejection wording changes subsequent edit proposals.\nPlatforms: unspecified\nTopics: evaluation, grounding, training-free\nPublished: 2026-10-11\nAuthors: Jianlin Chen, Wenhui Chen, Chi Man Vong\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.15738v1\"], \"reason\": \"Publisher document declaration only; no code, data or weights license inferred.\"}}}\n\nThe downstream test uses one screenshot-only agent, nine synthetic apps and 60 cached development items, with TAP-tag or text-answer outputs; this is not demonstrated live-app completion. Two catalogs reword the same 40 edits. Re-proposal effects do not establish internal causal mechanisms or task-success gains. Main serving providers were unpinned; rationale labels came from LLM annotators. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.15738v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-15738/"}
{"id": "arxiv:2610.16234", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa0e0ab6eb19ea55e65ccb354ae515f6552cd03dd688c6f43cdc6e2d59b11824", "submitted_at": "2026-10-12T06:09:02Z", "updated_at": "2026-10-12T06:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16234v1"}, "authors": ["Jingchao Wang", "Dingjiang Huang"], "canonical_source_url": "https://arxiv.org/abs/2610.16234", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-16234/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "aa0e0ab6eb19ea55e65ccb354ae515f6552cd03dd688c6f43cdc6e2d59b11824", "submitted_at": "2026-10-12T06:09:02Z", "updated_at": "2026-10-12T06:09:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16234v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.16234"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16234"}], "name": "Seg2Click: Learning Pixel-Level Action Masks for GUI Visual Grounding", "publication": null, "published_at": "2026-10-12", "record_sha256": "ac9505631e9fd147f355fdad80a6af0f03a9b6c13a6bc2b5fbe40238037c1fae", "relations": [], "resource_id": "arxiv:2610.16234", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.16234v1", "topics": ["grounding", "screen-understanding"], "updated_at": "2026-10-12"}, "text": "Seg2Click: Learning Pixel-Level Action Masks for GUI Visual Grounding\nResource type: paper\nReview status: curated\nScope: core\n\nRecasts GUI grounding as predicting a mask of valid click locations from bounding-box supervision, then derives a click point with a deterministic selection rule.\nPlatforms: unspecified\nTopics: grounding, screen-understanding\nPublished: 2026-10-12\nAuthors: Jingchao Wang, Dingjiang Huang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.16234v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-16234/"}
{"id": "arxiv:2610.16319", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4596cc60436c4fa6c2aa0951b4699b804edb73aa1b1a46900b1d14f901d20f3b", "submitted_at": "2026-10-12T07:42:11Z", "updated_at": "2026-10-12T07:42:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16319v1"}, "authors": ["Zijian Cai", "Haowei Liu", "Kailai Chen", "Lin Peng", "Cong Wan", "Ziqi Zhang", "Chunfeng Yuan", "Bing Li", "Weiming Hu", "Haiyang Xu"], "canonical_source_url": "https://arxiv.org/abs/2610.16319", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-16319/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "4596cc60436c4fa6c2aa0951b4699b804edb73aa1b1a46900b1d14f901d20f3b", "submitted_at": "2026-10-12T07:42:11Z", "updated_at": "2026-10-12T07:42:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16319v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.16319"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16319"}], "name": "HyWe: Unifying Search and Operation for Real-World Hybrid Web Tasks", "publication": null, "published_at": "2026-10-12", "record_sha256": "353f6b6c8588db2444f9be8df66828d1007763fb2d0f6d85ebfda4c99abf7b37", "relations": [], "resource_id": "arxiv:2610.16319", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.16319v1", "topics": ["web-navigation", "planning", "data-generation"], "updated_at": "2026-10-12"}, "text": "HyWe: Unifying Search and Operation for Real-World Hybrid Web Tasks\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces HyWe, which switches between information search and browser operation within one trajectory, alongside a demonstration pipeline and a benchmark of hybrid web tasks.\nPlatforms: web\nTopics: web-navigation, planning, data-generation\nPublished: 2026-10-12\nAuthors: Zijian Cai, Haowei Liu, Kailai Chen, Lin Peng, Cong Wan, Ziqi Zhang, Chunfeng Yuan, Bing Li, Weiming Hu, Haiyang Xu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.16319v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-16319/"}
{"id": "arxiv:2610.16720", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73073d2d533307cd71b05afef2441a8a7b8f198f804f11f83536d4297db2491a", "submitted_at": "2026-10-12T12:52:21Z", "updated_at": "2026-10-12T12:52:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16720v1"}, "authors": ["Jiahao Lu", "Yize Shen", "Chuanyuan Tan", "Mianqiu Huang", "Shicheng Fang", "Xinchi Chen", "Yining Zheng", "Yuxin Wang", "Xipeng Qiu"], "canonical_source_url": "https://arxiv.org/abs/2610.16720", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-16720/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "73073d2d533307cd71b05afef2441a8a7b8f198f804f11f83536d4297db2491a", "submitted_at": "2026-10-12T12:52:21Z", "updated_at": "2026-10-12T12:52:21Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16720v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.16720"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16720"}], "name": "Learn from Move: the Next Step for GUI Agents", "publication": null, "published_at": "2026-10-12", "record_sha256": "a9c7b4ff7db5994861c0d0181bedd85728499be94b947d681dd83ef71e52ce30", "relations": [], "resource_id": "arxiv:2610.16720", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.16720v1", "topics": ["evaluation", "grounding", "reinforcement-learning"], "updated_at": "2026-10-12"}, "text": "Learn from Move: the Next Step for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces LatentGUIWorld to test learning interface semantics and dynamics through interaction, and studies finer movement actions with intermediate visual feedback and exploration demonstrations.\nPlatforms: unspecified\nTopics: evaluation, grounding, reinforcement-learning\nPublished: 2026-10-12\nAuthors: Jiahao Lu, Yize Shen, Chuanyuan Tan, Mianqiu Huang, Shicheng Fang, Xinchi Chen, Yining Zheng, Yuxin Wang, Xipeng Qiu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.16720v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-16720/"}
{"id": "arxiv:2610.16778", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5a2743a18bae1deeed13fbd1dd7379c3f21c67488627057ccb466ffe51d09dc", "submitted_at": "2026-10-12T13:34:52Z", "updated_at": "2026-10-12T13:34:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16778v1"}, "authors": ["Zian Yuan", "Wei Li"], "canonical_source_url": "https://arxiv.org/abs/2610.16778", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-16778/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "c5a2743a18bae1deeed13fbd1dd7379c3f21c67488627057ccb466ffe51d09dc", "submitted_at": "2026-10-12T13:34:52Z", "updated_at": "2026-10-12T13:34:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.16778v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.16778"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.16778"}], "name": "PrivacyGuard: Source-Linked Sanitization and Locally Resolvable PII References for Cloud GUI Agents", "publication": null, "published_at": "2026-10-12", "record_sha256": "87e8aee502a11fff19f7e41ab56a239f9c23ab5b15ddf512b57ffb1e4ab45f29", "relations": [], "resource_id": "arxiv:2610.16778", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.16778v1", "topics": ["safety-security", "mobile-navigation", "evaluation"], "updated_at": "2026-10-12"}, "text": "PrivacyGuard: Source-Linked Sanitization and Locally Resolvable PII References for Cloud GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces PrivacyGuard, which sanitizes source-linked personal information and maintains locally resolvable references across mobile-agent turns to reduce exposure to remote planners.\nPlatforms: android\nTopics: safety-security, mobile-navigation, evaluation\nPublished: 2026-10-12\nAuthors: Zian Yuan, Wei Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.16778v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-16778/"}
{"id": "arxiv:2610.17457", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247cd4c9634dea78780a2230138c728adf53871e8b65614697f76ba53dea858c", "submitted_at": "2026-10-12T22:17:51Z", "updated_at": "2026-10-12T22:17:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17457v1"}, "authors": ["Yuetian Chen", "Kaiyuan Zhang", "Yuntao Du", "Thomas Li", "Mark Tenenholtz", "Kyle Polley", "Jerry Ma", "Denis Yarats", "Ninghui Li"], "canonical_source_url": "https://arxiv.org/abs/2610.17457", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-17457/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "247cd4c9634dea78780a2230138c728adf53871e8b65614697f76ba53dea858c", "submitted_at": "2026-10-12T22:17:51Z", "updated_at": "2026-10-12T22:17:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17457v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.17457"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17457"}], "name": "BrowseSafe++: Benchmarking Prompt Injection Detection Across Web Contexts, Languages, and Browser Views", "publication": null, "published_at": "2026-10-12", "record_sha256": "a84fd0d21d72ca85605c114cc22c8044911ae9e96138213325a0bbbe798068af", "relations": [], "resource_id": "arxiv:2610.17457", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.17457v1", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": "2026-10-12"}, "text": "BrowseSafe++: Benchmarking Prompt Injection Detection Across Web Contexts, Languages, and Browser Views\nResource type: paper\nReview status: curated\nScope: core\n\nExtends BrowseSafe with paired attacked and clean webpages in several representations, and studies text and screenshot detectors plus feedback-adapted prompt injections.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\nPublished: 2026-10-12\nAuthors: Yuetian Chen, Kaiyuan Zhang, Yuntao Du, Thomas Li, Mark Tenenholtz, Kyle Polley, Jerry Ma, Denis Yarats, Ninghui Li\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.17457v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-17457/"}
{"id": "arxiv:2610.17710", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46ea5a1a350ccfcc12d22ab3783e05e3c580719962831561b01aeb3fcc6f6c28", "submitted_at": "2026-10-13T04:37:18Z", "updated_at": "2026-10-13T04:37:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17710v1"}, "authors": ["Zirui Shang", "Yang Liu", "Xin Shu", "Zhi Gao", "Xinxiao Wu", "Lifeng Fan"], "canonical_source_url": "https://arxiv.org/abs/2610.17710", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-17710/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "46ea5a1a350ccfcc12d22ab3783e05e3c580719962831561b01aeb3fcc6f6c28", "submitted_at": "2026-10-13T04:37:18Z", "updated_at": "2026-10-13T04:37:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17710v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.17710"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17710"}], "name": "Adaptive-Group Process Rewards for GUI Agents", "publication": null, "published_at": "2026-10-13", "record_sha256": "9ec24b476c71c66ae544612e0703db9555bb4478656bbe4951924c9f33760a31", "relations": [], "resource_id": "arxiv:2610.17710", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.17710v1", "topics": ["reinforcement-learning", "long-horizon", "planning"], "updated_at": "2026-10-13"}, "text": "Adaptive-Group Process Rewards for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nGroups GUI trajectories by subgoal and combines action-level and group-level assessments into intermediate rewards for training agents with sparse task outcomes.\nPlatforms: unspecified\nTopics: reinforcement-learning, long-horizon, planning\nPublished: 2026-10-13\nAuthors: Zirui Shang, Yang Liu, Xin Shu, Zhi Gao, Xinxiao Wu, Lifeng Fan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.17710v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-17710/"}
{"id": "arxiv:2610.17792", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b1786bba4f48ca54166d49e3d98d55aaef52eb6e2233f282d8d94d54dfb5ed2c", "submitted_at": "2026-10-13T06:20:38Z", "updated_at": "2026-10-13T06:20:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17792v1"}, "authors": ["Minghua Chen", "Xingming Liao", "Weiqiang Zhong", "Zhuowei Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.17792", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-17792/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b1786bba4f48ca54166d49e3d98d55aaef52eb6e2233f282d8d94d54dfb5ed2c", "submitted_at": "2026-10-13T06:20:38Z", "updated_at": "2026-10-13T06:20:38Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17792v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.17792"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17792"}], "name": "MHVF: Training-Free GUI Grounding via Verification-Guided Multi-Hypothesis Search", "publication": null, "published_at": "2026-10-13", "record_sha256": "a55ea737c724e6ef168e818393c5ece2a105eda5dfd50b6ccf6d030c15359b6a", "relations": [], "resource_id": "arxiv:2610.17792", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.17792v1", "topics": ["grounding", "screen-understanding", "robustness"], "updated_at": "2026-10-13"}, "text": "MHVF: Training-Free GUI Grounding via Verification-Guided Multi-Hypothesis Search\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces MHVF, a training-free grounding method that explores several candidate screen regions, verifies them in local views, and uses those checks to guide further search.\nPlatforms: web\nTopics: grounding, screen-understanding, robustness\nPublished: 2026-10-13\nAuthors: Minghua Chen, Xingming Liao, Weiqiang Zhong, Zhuowei Wang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.17792v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-17792/"}
{"id": "arxiv:2610.17895", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14d2128155d1540955e54c3bdf9010724f27cf2e101079650315602856c98efa", "submitted_at": "2026-10-13T07:49:50Z", "updated_at": "2026-10-14T06:37:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.17895v2"}, "authors": ["Wayne Chi", "I-Hung Hsu", "Li Chen", "Wei-Lin Chiang", "Ion Stoica", "Anastasios Nikolas Angelopoulos", "Ameet Talwalkar", "Chris Donahue", "Cho-Jui Hsieh"], "canonical_source_url": "https://arxiv.org/abs/2610.17895", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-17895/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "14d2128155d1540955e54c3bdf9010724f27cf2e101079650315602856c98efa", "submitted_at": "2026-10-13T07:49:50Z", "updated_at": "2026-10-14T06:37:47Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.17895v2"}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.17895"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17895"}], "name": "Interaction-as-a-Judge", "publication": null, "published_at": "2026-10-13", "record_sha256": "e2dba726b933a302f1baeaf7436d28206cd40cbe0cca7861f83530e26ab2bc63", "relations": [], "resource_id": "arxiv:2610.17895", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.17895v2", "topics": ["evaluation", "data-generation", "automation"], "updated_at": "2026-10-14"}, "text": "Interaction-as-a-Judge\nResource type: paper\nReview status: curated\nScope: core\n\nUses computer-use agents to interact with generated applications, studying whether interaction improves preference judging, requirement checks, and feedback for application generation.\nPlatforms: web, desktop\nTopics: evaluation, data-generation, automation\nPublished: 2026-10-13\nAuthors: Wayne Chi, I-Hung Hsu, Li Chen, Wei-Lin Chiang, Ion Stoica, Anastasios Nikolas Angelopoulos, Ameet Talwalkar, Chris Donahue, Cho-Jui Hsieh\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.17895v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-17895/"}
{"id": "arxiv:2610.17928", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "551f198c0976c75693c6298400800cc66ce1b1b5ee304965155a6bb3ca9d68fe", "submitted_at": "2026-10-13T08:15:52Z", "updated_at": "2026-10-13T08:15:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17928v1"}, "authors": ["Taeyun Roh", "Suhyeong Park", "Gijin Kim", "Wonjune Jang", "Jaewoo Kang"], "canonical_source_url": "https://arxiv.org/abs/2610.17928", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-17928/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "551f198c0976c75693c6298400800cc66ce1b1b5ee304965155a6bb3ca9d68fe", "submitted_at": "2026-10-13T08:15:52Z", "updated_at": "2026-10-13T08:15:52Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.17928v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "3fb5aabfb07df620955f0b74347f9d5926a1cd3d506b1098eccba9477627269f", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/dmis-lab/SwitchMem/HEAD/README.md"}], "identifiers": {"arxiv": "2610.17928"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.17928"}], "name": "Learning How to Remember: Memory Modality Selection for GUI Agents", "publication": null, "published_at": "2026-10-13", "record_sha256": "16724f2f84295c4e691cd14ac046abe3bdb3fd96a922dca1a78099b6d19df721", "relations": [], "resource_id": "arxiv:2610.17928", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.17928v1", "topics": ["memory", "mobile-navigation"], "updated_at": "2026-10-13"}, "text": "Learning How to Remember: Memory Modality Selection for GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SwitchMem, which chooses a textual or visual representation of retrieved experience at each GUI interaction step instead of fixing one memory modality for a task.\nPlatforms: android\nTopics: memory, mobile-navigation\nPublished: 2026-10-13\nAuthors: Taeyun Roh, Suhyeong Park, Gijin Kim, Wonjune Jang, Jaewoo Kang\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. Android platform tagging is additionally supported by the primary project documentation.\n\nPrimary source: https://arxiv.org/abs/2610.17928v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-17928/"}
{"id": "arxiv:2610.18067", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6add363f25f87ea1814c7b79ebf6b51f7431b3745a7d2d910960b78e2d97d13f", "submitted_at": "2026-10-13T09:52:02Z", "updated_at": "2026-10-13T09:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18067v1"}, "authors": ["Hang He", "Linting Wang", "Guanming Liu", "Bo Mao", "Jinxiu Liu", "Yiran Suo", "Maosen Zhou", "Zhiheng Xi", "Guanting Dong", "Zhou Shao", "Rui Zheng", "Jie Zhou", "Chengcheng Wan", "Tao Gui", "Qi Zhang", "Xipeng Qiu", "Xuanjing Huang"], "canonical_source_url": "https://arxiv.org/abs/2610.18067", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-18067/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6add363f25f87ea1814c7b79ebf6b51f7431b3745a7d2d910960b78e2d97d13f", "submitted_at": "2026-10-13T09:52:02Z", "updated_at": "2026-10-13T09:52:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18067v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.18067v1"], "reason": "Document declaration only; code, data and weights permissions are not inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:20Z", "role": "primary", "sha256": "f1fbdd2ecb2a54b7723372118e2b0ac818420f3a3dd280c2486c8e064cf486d6", "supports": ["name", "kind", "summary", "scope", "environments", "topics", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.18067v1"}], "identifiers": {"arxiv": "2610.18067"}, "kind": "paper", "last_verified_at": "2026-10-20T19:02:20Z", "licenses": {"code": null, "data": null, "document": {"basis": "The exact-v1 arXiv HTML infobox explicitly declares CC BY 4.0 for this paper.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:20Z", "url": "https://arxiv.org/html/2610.18067v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18067"}], "name": "WorkRover: Scaling Long-Horizon Cowork Agent Training Via Recursive Environment Evolution", "publication": null, "published_at": "2026-10-13", "record_sha256": "e6d1e58019c365332004dee67faba7b4c25d5ab5cd5e3ffa3c10281c7206dfff", "relations": [], "resource_id": "arxiv:2610.18067", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.18067v1", "topics": ["data-generation", "long-horizon", "evaluation", "reward-models", "reinforcement-learning", "workflow"], "updated_at": "2026-10-13"}, "text": "WorkRover: Scaling Long-Horizon Cowork Agent Training Via Recursive Environment Evolution\nResource type: paper\nReview status: curated\nScope: core\n\nSynthesizes resettable workspaces with files, application observations and skills; repairs environment incompatibilities and trains long-task agents using verifier-linked turn credit.\nPlatforms: unspecified\nTopics: data-generation, long-horizon, evaluation, reward-models, reinforcement-learning, workflow\nPublished: 2026-10-13\nAuthors: Hang He, Linting Wang, Guanming Liu, Bo Mao, Jinxiu Liu, Yiran Suo, Maosen Zhou, Zhiheng Xi, Guanting Dong, Zhou Shao, Rui Zheng, Jie Zhou, Chengcheng Wan, Tao Gui, Qi Zhang, Xipeng Qiu, Xuanjing Huang\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.18067v1\"], \"reason\": \"Document declaration only; code, data and weights permissions are not inferred.\"}}}\n\nApp WorkEnvs expose UI observations and interaction interfaces alongside file and tool workflows; no specific operating system or pixel-control success is established. Verifier shaping depends on deterministic-check and dependency coverage. External reference scores use differing harnesses or subsets; GDPval Elo is not comparable to percentage metrics. A successful-trace analysis subset is separate from online RL outcomes. No implementation release or independent reproduction was verified.\n\nPrimary source: https://arxiv.org/abs/2610.18067v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-18067/"}
{"id": "arxiv:2610.18508", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5565d38dd40b02b4158d7edcc2c5c27e7e285dbedefe1c7101a31c4f8d17b659", "submitted_at": "2026-10-13T15:03:33Z", "updated_at": "2026-10-13T15:03:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18508v1"}, "authors": ["Dominik Seip", "Bartol Markovinović", "Matthias Hein"], "canonical_source_url": "https://arxiv.org/abs/2610.18508", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-18508/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5565d38dd40b02b4158d7edcc2c5c27e7e285dbedefe1c7101a31c4f8d17b659", "submitted_at": "2026-10-13T15:03:33Z", "updated_at": "2026-10-13T15:03:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18508v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.18508"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18508"}], "name": "Towards Safe Autonomy for Computer Use", "publication": null, "published_at": "2026-10-13", "record_sha256": "9817023e583f5d02bf636f4ffdfd798cf166010cb35540dd2c527874f399cf29", "relations": [], "resource_id": "arxiv:2610.18508", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.18508v1", "topics": ["safety-security", "screen-understanding"], "updated_at": "2026-10-13"}, "text": "Towards Safe Autonomy for Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nPresents SA-CUA, separating trusted and untrusted visual content using accessibility-derived masking rules and a controlled channel for verifying information before endorsement.\nPlatforms: unspecified\nTopics: safety-security, screen-understanding\nPublished: 2026-10-13\nAuthors: Dominik Seip, Bartol Markovinović, Matthias Hein\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.18508v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-18508/"}
{"id": "arxiv:2610.18579", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "070f02f1ea798369817f471df5fe1d002b68d544135a1ac8e3a2aed2830e8c01", "submitted_at": "2026-10-13T15:43:30Z", "updated_at": "2026-10-13T15:43:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18579v1"}, "authors": ["Yubo Li", "Ramayya Krishnan", "Rema Padman"], "canonical_source_url": "https://arxiv.org/abs/2610.18579", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-18579/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "070f02f1ea798369817f471df5fe1d002b68d544135a1ac8e3a2aed2830e8c01", "submitted_at": "2026-10-13T15:43:30Z", "updated_at": "2026-10-13T15:43:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.18579v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.18579v1"], "reason": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses."}}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T15:06:37Z", "role": "primary", "sha256": "a3d753815e82190ee2de6f322cd816c32f2dd391f862845305fc7e57300303a6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22desktop+agent%22+OR+abs%3A%22computer+interaction%22+OR+abs%3A%22GUI+automation%22+OR+abs%3A%22web+navigation%22+OR+abs%3A%22browser+automation%22+OR+abs%3A%22visual+grounding%22+OR+abs%3A%22screen+understanding%22+OR+abs%3A%22mobile+agents%22%29+AND+submittedDate%3A%5B202609200000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:08:55Z", "role": "primary", "sha256": "bcbc640d1e732ba0415a9755bce120a811a13a2ba7fba3e58a167b3a9111c2a6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.18579v1"}], "identifiers": {"arxiv": "2610.18579"}, "kind": "paper", "last_verified_at": "2026-10-20T15:08:55Z", "licenses": {"code": null, "data": null, "document": {"basis": "The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T15:08:55Z", "url": "https://arxiv.org/html/2610.18579v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18579"}], "name": "Reading Right, Acting Wrong: Diagnosing Decision Failures of System One Models in Agent Loops", "publication": null, "published_at": "2026-10-13", "record_sha256": "8d4c86ebe2f0a523b1513316a600f57aafa48df8fa0fc2c251dcfdab6ddee418", "relations": [], "resource_id": "arxiv:2610.18579", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.18579v1", "topics": ["evaluation", "web-navigation", "interaction-history"], "updated_at": "2026-10-13"}, "text": "Reading Right, Acting Wrong: Diagnosing Decision Failures of System One Models in Agent Loops\nResource type: paper\nReview status: curated\nScope: adjacent\n\nPairs factual probes with next-action choices on frozen agent traces, including Mind2Web, to diagnose reading-decision gaps without closed-loop execution.\nPlatforms: web\nTopics: evaluation, web-navigation, interaction-history\nPublished: 2026-10-13\nAuthors: Yubo Li, Ramayya Krishnan, Rema Padman\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.18579v1\"], \"reason\": \"The retained version-specific arXiv HTML infobox explicitly labels this paper License: CC BY 4.0. This document declaration does not establish code, training-data or model-weight licenses.\"}}}\n\nAdjacent diagnostic research across simulators, policies, tools, coding and an offline Mind2Web subset. Models choose among supplied admissible actions; they do not generate actions or execute a closed-loop browser policy in the reported evaluation. The paired probes measure selected visible facts, not complete understanding or internal causation. Pooled results are not a web-only leaderboard, and the paper reports coverage limits and label-sensitivity analyses. No benchmark release, code, data or weights were independently inspected.\n\nPrimary source: https://arxiv.org/abs/2610.18579v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-18579/"}
{"id": "arxiv:2610.18784", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4a9b6fae506d4dd35798c1930d5df4ec7fc7a2c452ab8bd5a5714cf385d505", "submitted_at": "2026-10-13T17:35:27Z", "updated_at": "2026-10-14T04:46:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.18784v2"}, "authors": ["Yuchen Lu", "Julian McAuley", "Lei Zhang"], "canonical_source_url": "https://arxiv.org/abs/2610.18784", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-18784/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6f4a9b6fae506d4dd35798c1930d5df4ec7fc7a2c452ab8bd5a5714cf385d505", "submitted_at": "2026-10-13T17:35:27Z", "updated_at": "2026-10-14T04:46:42Z", "version": 2, "versioned_url": "https://arxiv.org/abs/2610.18784v2"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.18784"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.18784"}], "name": "Pack Light, CarryOn: Building Object-Oriented Context for Long-Horizon Agents", "publication": null, "published_at": "2026-10-13", "record_sha256": "846063969bef68d29d43c8ffe19dc630fd373454d52435074007edf9fcf2ed1d", "relations": [], "resource_id": "arxiv:2610.18784", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.18784v2", "topics": ["memory", "long-horizon", "efficiency"], "updated_at": "2026-10-14"}, "text": "Pack Light, CarryOn: Building Object-Oriented Context for Long-Horizon Agents\nResource type: paper\nReview status: curated\nScope: adjacent\n\nIntroduces CarryOn, which maintains evidence-linked task objects and remaining-work context across history compression, with evaluations on desktop tasks and web research.\nPlatforms: desktop, web\nTopics: memory, long-horizon, efficiency\nPublished: 2026-10-13\nAuthors: Yuchen Lu, Julian McAuley, Lei Zhang\nRecorded arXiv version: v2\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.18784v2\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-18784/"}
{"id": "arxiv:2610.19100", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08d90bcec690b238be96207602ca694759c8672fecbed85e30372dd314e4006e", "submitted_at": "2026-10-13T19:37:49Z", "updated_at": "2026-10-13T19:37:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19100v1"}, "authors": ["Qingsong Xie", "Jiahao Tian", "Jiacheng Tang", "Zhifeng Li", "Madalina Ciobanu", "Navan Preet Singh", "Qingqing Mao", "Ritankar Das"], "canonical_source_url": "https://arxiv.org/abs/2610.19100", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19100/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08d90bcec690b238be96207602ca694759c8672fecbed85e30372dd314e4006e", "submitted_at": "2026-10-13T19:37:49Z", "updated_at": "2026-10-13T19:37:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19100v1"}}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.19100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19100"}], "name": "GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning", "publication": null, "published_at": "2026-10-13", "record_sha256": "b584716ffc914fe421b09670ad190dd27c92997bc46ed1e915f7a98d01b0099e", "relations": [], "resource_id": "arxiv:2610.19100", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.19100v1", "topics": ["grounding", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-13"}, "text": "GroundGUI: GUI Grounding by High-Quality Data and Distillation-to-Reinforcement Learning\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces ScreenRef screenshot grounding data and a training sequence from supervised learning through teacher-guided distillation to reinforcement learning for GroundGUI-8B.\nPlatforms: desktop, web, cross-platform\nTopics: grounding, data-generation, reinforcement-learning\nPublished: 2026-10-13\nAuthors: Qingsong Xie, Jiahao Tian, Jiacheng Tang, Zhifeng Li, Madalina Ciobanu, Navan Preet Singh, Qingqing Mao, Ritankar Das\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.19100v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19100/"}
{"id": "arxiv:2610.19128", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f72958f9a70e7714f17b0eb51c5453e08ea8ad4e8e7d4caa5e3170231088251", "submitted_at": "2026-10-13T20:01:18Z", "updated_at": "2026-10-13T20:01:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19128v1"}, "authors": ["Yuxiang Lai", "Jike Zhong", "Chengzhi Liu", "Ming Li", "Haonan Wang", "Xiaofeng Yang", "Liang Zhao", "Youbao Tang", "Ruixi Fan"], "canonical_source_url": "https://arxiv.org/abs/2610.19128", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19128/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "5f72958f9a70e7714f17b0eb51c5453e08ea8ad4e8e7d4caa5e3170231088251", "submitted_at": "2026-10-13T20:01:18Z", "updated_at": "2026-10-13T20:01:18Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19128v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.19128"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19128"}], "name": "CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation Pipeline", "publication": null, "published_at": "2026-10-13", "record_sha256": "dbd63ff8da39b5e4a86ee20634f734f5747c92b6f70d7904839c8c15e4c401d7", "relations": [], "resource_id": "arxiv:2610.19128", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.19128v1", "topics": ["evaluation", "robustness", "workflow"], "updated_at": "2026-10-13"}, "text": "CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation Pipeline\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces CLASHarness, generating verifiable computer-use tasks with conflicting agent reports, shared-state constraints, and coordination requirements to evaluate conflict detection and resolution.\nPlatforms: unspecified\nTopics: evaluation, robustness, workflow\nPublished: 2026-10-13\nAuthors: Yuxiang Lai, Jike Zhong, Chengzhi Liu, Ming Li, Haonan Wang, Xiaofeng Yang, Liang Zhao, Youbao Tang, Ruixi Fan\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.19128v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19128/"}
{"id": "arxiv:2610.19366", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f1a48aa4b89fb97e8e1a186ba7345826a8b9efc9aa7b82e0dd1f5ce17d42b73", "submitted_at": "2026-10-14T01:11:57Z", "updated_at": "2026-10-14T01:11:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19366v1"}, "authors": ["Minghao Guo", "Qingyue Jiao", "Xiujin Liu", "Jianhong Pang", "Ruimeng Yang", "Naihao Xu", "Liwei Che", "Haochen Xue", "Ziqi Wang", "Xinying Cai", "Jiaxuan Luo", "Jiaxin Deng", "Xuesong Ye", "Jintao Qu", "Wenyao Gao", "Jingwen Gu", "Chenyang Zhao", "Shichun Liu", "Guibin Zhang", "Zefang Liu", "Siyin Wang", "Wujiang Xu", "Danrui Li", "Zeru Shi", "Woongyeong Yeo", "Shuang Liu", "Wei-Chieh Huang", "Weizhi Zhang", "Tianyu Yang", "Kangsan Kim", "Yebowen Hu", "Bowen Wen", "Bowen Jiang", "Haohao Su", "Ming Yin", "Sheldon Yu", "Mian Zhang", "Yuyan Chen", "Hanrong Zhang", "Kai Mei", "Zihao Zhang", "Mingyu Jin", "Xi Zhu", "Hengyi Wang", "Tianxin Wei", "Linyong Nan", "Zhenwen Liang", "Zhendong Chu", "Ziqi Huang", "Yongye Su", "Sixun Dong", "Shujian Liu", "Kaiqu Liang", "Jiaojiao Han", "Zhaoye Fei", "Kun Wang", "Shihan Dou", "Zhichun Guo", "Xingjian Diao", "Zeqi Gu", "Heng Wang", "Tao Feng", "Cheng Qian", "Hamid Palangi", "Silei Xu", "Xi Yun", "Yi Zhu", "Shilong Liu", "Qingsong Wen", "Sung Ju Hwang", "Zuxuan Wu", "Yu-Gang Jiang", "Yiyu Shi", "Dimitris N. Metaxas", "Jiaxuan You", "Julian McAuley", "Mengdi Wang", "Jiawei Han", "Philip S. Yu", "Ziyi Ye"], "canonical_source_url": "https://arxiv.org/abs/2610.19366", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19366/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7f1a48aa4b89fb97e8e1a186ba7345826a8b9efc9aa7b82e0dd1f5ce17d42b73", "submitted_at": "2026-10-14T01:11:57Z", "updated_at": "2026-10-14T01:11:57Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19366v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.19366v1"], "reason": "Paper-specific document declaration only; no code, data or weights permissions are inferred."}}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-17T12:31:39Z", "role": "primary", "sha256": "c52e4e283e0dae88ef8b3ce67b7424b1d4b48e8f41015d872f4da69f98782eaa", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2609.39924%2C2610.00346%2C2610.00583%2C2610.01939%2C2610.05099%2C2610.09127%2C2610.09484%2C2610.09633%2C2610.11251%2C2610.11922%2C2610.13098%2C2610.13166%2C2610.13361%2C2610.13623%2C2610.16117%2C2610.17953%2C2610.18067%2C2610.19366%2C2610.20424&max_results=19"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:22.864948+00:00", "role": "primary", "sha256": "308796868bf9b1cb042bc6008e1e6f4ba6e776a2d4e13d40ad5684b477861563", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.19366v1"}], "identifiers": {"arxiv": "2610.19366"}, "kind": "survey", "last_verified_at": "2026-10-20T19:02:22.864948+00:00", "licenses": {"code": null, "data": null, "document": {"basis": "The retained exact-v1 arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:22.864948+00:00", "url": "https://arxiv.org/html/2610.19366v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19366"}], "name": "Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory", "publication": null, "published_at": "2026-10-14", "record_sha256": "6f036d006c6e5125154eddb65956c28c68c72f765c197da7a710606492dd6a2a", "relations": [], "resource_id": "arxiv:2610.19366", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.19366v1", "topics": ["survey", "memory", "evaluation", "interaction-history"], "updated_at": "2026-10-14"}, "text": "Seeing, Maintaining, and Learning: The Evolution of Multimodal Agent Memory\nResource type: survey\nReview status: curated\nScope: adjacent\n\nSurveys multimodal memory using representation and management axes, distinguishing fixed-policy updates from policy evolution and comparing GUI memory evaluation with other modalities.\nPlatforms: unspecified\nTopics: survey, memory, evaluation, interaction-history\nPublished: 2026-10-14\nAuthors: Minghao Guo, Qingyue Jiao, Xiujin Liu, Jianhong Pang, Ruimeng Yang, Naihao Xu, Liwei Che, Haochen Xue, Ziqi Wang, Xinying Cai, Jiaxuan Luo, Jiaxin Deng, Xuesong Ye, Jintao Qu, Wenyao Gao, Jingwen Gu, Chenyang Zhao, Shichun Liu, Guibin Zhang, Zefang Liu, Siyin Wang, Wujiang Xu, Danrui Li, Zeru Shi, Woongyeong Yeo, Shuang Liu, Wei-Chieh Huang, Weizhi Zhang, Tianyu Yang, Kangsan Kim, Yebowen Hu, Bowen Wen, Bowen Jiang, Haohao Su, Ming Yin, Sheldon Yu, Mian Zhang, Yuyan Chen, Hanrong Zhang, Kai Mei, Zihao Zhang, Mingyu Jin, Xi Zhu, Hengyi Wang, Tianxin Wei, Linyong Nan, Zhenwen Liang, Zhendong Chu, Ziqi Huang, Yongye Su, Sixun Dong, Shujian Liu, Kaiqu Liang, Jiaojiao Han, Zhaoye Fei, Kun Wang, Shihan Dou, Zhichun Guo, Xingjian Diao, Zeqi Gu, Heng Wang, Tao Feng, Cheng Qian, Hamid Palangi, Silei Xu, Xi Yun, Yi Zhu, Shilong Liu, Qingsong Wen, Sung Ju Hwang, Zuxuan Wu, Yu-Gang Jiang, Yiyu Shi, Dimitris N. Metaxas, Jiaxuan You, Julian McAuley, Mengdi Wang, Jiawei Han, Philip S. Yu, Ziyi Ye\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.19366v1\"], \"reason\": \"Paper-specific document declaration only; no code, data or weights permissions are inferred.\"}}}\n\nA broad literature synthesis spanning images, video, audio, GUI and embodied settings. Its GUI section compares interface-state retention, action outcomes, cross-session reuse and visual evidence, but the paper does not provide a new GUI-agent experiment. Taxonomy corpus counts and benchmark classifications are authors' annotations, not independently audited field coverage. Updating stored content alone does not establish the survey's self-evolving-policy tier; endpoint task success does not isolate memory use. Lifecycle, resource-budget and provenance recommendations are proposed evaluation directions rather than demonstrated guarantees. Platform facets stay unspecified; cited systems and releases were not independently verified.\n\nPrimary source: https://arxiv.org/abs/2610.19366v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19366/"}
{"id": "arxiv:2610.19384", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da334d943d98b4bf9915c6ca28aab734d736f376e67cbab950aa1af65f540ebf", "submitted_at": "2026-10-14T01:30:03Z", "updated_at": "2026-10-14T01:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19384v1"}, "authors": ["Yuke Wang", "Benlei Cui", "Shen Pang", "Xuemei Dong", "Longtao Huang", "Hui Xue", "Yuwen Zhai", "Junjie Li", "Jingqun Tang", "Haiwen Hong"], "canonical_source_url": "https://arxiv.org/abs/2610.19384", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19384/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "da334d943d98b4bf9915c6ca28aab734d736f376e67cbab950aa1af65f540ebf", "submitted_at": "2026-10-14T01:30:03Z", "updated_at": "2026-10-14T01:30:03Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19384v1"}, "license_review": {"document": {"evidence_urls": ["https://arxiv.org/html/2610.19384v1"], "reason": "Paper-specific document declaration only; no code, data or weights permissions are inferred."}}}, "environments": ["web", "desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "role": "primary", "sha256": "ca995ec330f62a43d3d4fbbf28dfe83c5fa8b05f5c1e446929493409760268d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://arxiv.org/html/2610.19384v1"}], "identifiers": {"arxiv": "2610.19384"}, "kind": "paper", "last_verified_at": "2026-10-20T19:02:16.329923+00:00", "licenses": {"code": null, "data": null, "document": {"basis": "The retained exact-v1 arXiv HTML infobox explicitly declares License: CC BY 4.0 for this document.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-20T19:02:16.329923+00:00", "url": "https://arxiv.org/html/2610.19384v1"}, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19384"}], "name": "MMA-SafetyBench: A Benchmark for Multimodal Agent Safety Evaluation", "publication": null, "published_at": "2026-10-14", "record_sha256": "e9a2b4a3ad3141ddfcd7980320e3e1205448ddd20c9789f8a8a4638c92fc09be", "relations": [], "resource_id": "arxiv:2610.19384", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.19384v1", "topics": ["evaluation", "safety-security", "grounding"], "updated_at": "2026-10-14"}, "text": "MMA-SafetyBench: A Benchmark for Multimodal Agent Safety Evaluation\nResource type: paper\nReview status: curated\nScope: core\n\nBuilds trajectory-guided visual-injection tests for web, desktop and mobile observations, separating target-matching action predictions from broader semantic compromise.\nPlatforms: web, desktop\nTopics: evaluation, safety-security, grounding\nPublished: 2026-10-14\nAuthors: Yuke Wang, Benlei Cui, Shen Pang, Xuemei Dong, Longtao Huang, Hui Xue, Yuwen Zhai, Junjie Li, Jingqun Tang, Haiwen Hong\nRecorded arXiv version: v1\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://arxiv.org/html/2610.19384v1\"], \"reason\": \"Paper-specific document declaration only; no code, data or weights permissions are inferred.\"}}}\n\nASR measures predicted coordinates inside an annotated target or compromised generated content; actual clicks, environment changes and downstream harm are not verified. SCR additionally counts adopted adversarial intent or task/schema disruption. Clean-task and attacked-trajectory denominators differ, including 42 clean desktop tasks versus 209 adversarial desktop trajectories. Payloads and injection points are fixed, and neutral prompts change both compliance cues and instruction-priority guidance. GPT-5 judging has a 300-trajectory human-label check, with borderline disagreements. Android-style mobile dialogs do not establish an Android runtime. Code/data links were identified but not inspected; no independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.19384v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19384/"}
{"id": "arxiv:2610.19834", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d154a14c5ac85fb352721c203fe7bd709eaba4287ef7376e6a6f5fc395708f1", "submitted_at": "2026-10-14T09:06:49Z", "updated_at": "2026-10-14T09:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19834v1"}, "authors": ["Md Farhan Ishmam", "Kenneth Marino"], "canonical_source_url": "https://arxiv.org/abs/2610.19834", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19834/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6d154a14c5ac85fb352721c203fe7bd709eaba4287ef7376e6a6f5fc395708f1", "submitted_at": "2026-10-14T09:06:49Z", "updated_at": "2026-10-14T09:06:49Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19834v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.19834"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19834"}], "name": "Teaching Agents in their Own Style: Blackbox Teacher Distillation for Web Agents", "publication": null, "published_at": "2026-10-14", "record_sha256": "2c338ed927abf171af9fea4029bce79dca536436b6ed990cbb3de32ff76f40ef", "relations": [], "resource_id": "arxiv:2610.19834", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.19834v1", "topics": ["data-generation", "web-navigation", "efficiency"], "updated_at": "2026-10-14"}, "text": "Teaching Agents in their Own Style: Blackbox Teacher Distillation for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces WebKD, which repairs failed student web-agent trajectories with minimal teacher edits and trains on corrections that preserve the student's reasoning style.\nPlatforms: web\nTopics: data-generation, web-navigation, efficiency\nPublished: 2026-10-14\nAuthors: Md Farhan Ishmam, Kenneth Marino\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.19834v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19834/"}
{"id": "arxiv:2610.19900", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50d5cb78e6715f0749c5973e900745ecdb3672febc2c8e090da94cb708127352", "submitted_at": "2026-10-14T09:58:40Z", "updated_at": "2026-10-14T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19900v1"}, "authors": ["Ridong Zhang", "Shuxian Liang", "Chaoqun Wan", "Zhen Huang", "Xian-Sheng Hua"], "canonical_source_url": "https://arxiv.org/abs/2610.19900", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-19900/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "50d5cb78e6715f0749c5973e900745ecdb3672febc2c8e090da94cb708127352", "submitted_at": "2026-10-14T09:58:40Z", "updated_at": "2026-10-14T09:58:40Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.19900v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "652e8369ef11334d76c0eb67afeaa5c9f7841fd31ca02243b0dfb1493316849a", "supports": ["environments", "notes"], "url": "https://iaie-physicalai.github.io/GraceBench"}], "identifiers": {"arxiv": "2610.19900"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.19900"}], "name": "How GUI Agents Fail: GraceBench for Execution Stability Diagnosis", "publication": null, "published_at": "2026-10-14", "record_sha256": "f77ececf1e48102dc69ee36330d1b670cd9c8b4786cc28462c75ee3c774d0e34", "relations": [], "resource_id": "arxiv:2610.19900", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.19900v1", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": "2026-10-14"}, "text": "How GUI Agents Fail: GraceBench for Execution Stability Diagnosis\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces GraceBench, using application transition graphs and recorded interaction evidence to diagnose deviations, loops, and recovery that final task-success scores can hide.\nPlatforms: android\nTopics: evaluation, mobile-navigation, robustness\nPublished: 2026-10-14\nAuthors: Ridong Zhang, Shuxian Liang, Chaoqun Wan, Zhen Huang, Xian-Sheng Hua\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced. Android platform tagging is additionally supported by the primary project documentation.\n\nPrimary source: https://arxiv.org/abs/2610.19900v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-19900/"}
{"id": "arxiv:2610.20100", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "602f4eb7aacc05748b83b9c32b28439098005c77eafb7624804f820995521a79", "submitted_at": "2026-10-14T12:35:59Z", "updated_at": "2026-10-14T12:35:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20100v1"}, "authors": ["Wenkui Yang", "Yiming Hao", "Kai Zhou", "Chengyu Pu", "Yuzhe Zhang", "Jie Cao", "Jin Huang", "Shuai Ma", "Feng Tian"], "canonical_source_url": "https://arxiv.org/abs/2610.20100", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-20100/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "602f4eb7aacc05748b83b9c32b28439098005c77eafb7624804f820995521a79", "submitted_at": "2026-10-14T12:35:59Z", "updated_at": "2026-10-14T12:35:59Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20100v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.20100"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20100"}], "name": "Task-State Governance: A Virtual Machine for Verifiable and Recoverable GUI Agent Execution", "publication": null, "published_at": "2026-10-14", "record_sha256": "cbee41fa5964b21c6c0e648118c78cc5679cfa0260e1471c4502ccf9c666f777", "relations": [], "resource_id": "arxiv:2610.20100", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.20100v1", "topics": ["human-in-the-loop", "safety-security", "workflow"], "updated_at": "2026-10-14"}, "text": "Task-State Governance: A Virtual Machine for Verifiable and Recoverable GUI Agent Execution\nResource type: paper\nReview status: curated\nScope: core\n\nPresents TaskVM, an interface for specifying desired and preserved application state, checking agent actions, and exposing whether changes can be reversed or compensated.\nPlatforms: unspecified\nTopics: human-in-the-loop, safety-security, workflow\nPublished: 2026-10-14\nAuthors: Wenkui Yang, Yiming Hao, Kai Zhou, Chengyu Pu, Yuzhe Zhang, Jie Cao, Jin Huang, Shuai Ma, Feng Tian\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.20100v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-20100/"}
{"id": "arxiv:2610.20107", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47488d3b5450753b4e1a4b0991e50dc92fbab197a9b54bd8d4e3e1a3d762ee96", "submitted_at": "2026-10-14T12:39:51Z", "updated_at": "2026-10-14T12:39:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20107v1"}, "authors": ["Chengguang Gan", "Yunhao Liang", "Hanjun Wei", "Qinghao Zhang", "Zhixi Cai", "Shiwen Ni"], "canonical_source_url": "https://arxiv.org/abs/2610.20107", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-20107/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "47488d3b5450753b4e1a4b0991e50dc92fbab197a9b54bd8d4e3e1a3d762ee96", "submitted_at": "2026-10-14T12:39:51Z", "updated_at": "2026-10-14T12:39:51Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20107v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.20107"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20107"}], "name": "Why Generate a Click? Plug-in Decision Heads for Web Agents", "publication": null, "published_at": "2026-10-14", "record_sha256": "b02ecda79446b00742ef27ab2893e0e9c1ea707ae701a1345e0045897eb58e44", "relations": [], "resource_id": "arxiv:2610.20107", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.20107v1", "topics": ["efficiency", "grounding", "web-navigation"], "updated_at": "2026-10-14"}, "text": "Why Generate a Click? Plug-in Decision Heads for Web Agents\nResource type: paper\nReview status: curated\nScope: core\n\nAdds decision heads to a frozen vision-language generator for selecting web actions and targets, with a confidence gate that falls back to text generation.\nPlatforms: web\nTopics: efficiency, grounding, web-navigation\nPublished: 2026-10-14\nAuthors: Chengguang Gan, Yunhao Liang, Hanjun Wei, Qinghao Zhang, Zhixi Cai, Shiwen Ni\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.20107v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-20107/"}
{"id": "arxiv:2610.20390", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f9d3ce7dffa8c726ce8129083c8d6c580449d1a32d87c7cd3aaade4fe6cf59b", "submitted_at": "2026-10-14T15:53:48Z", "updated_at": "2026-10-14T15:53:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20390v1"}, "authors": ["Shuo Wang", "Sifan Zhou", "Xiaobo Wang", "Tianyu Fu", "Chenxu Zhao", "Anyang Su", "Hanning Wang", "Ping Jiang", "Minghui Wu"], "canonical_source_url": "https://arxiv.org/abs/2610.20390", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-20390/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f9d3ce7dffa8c726ce8129083c8d6c580449d1a32d87c7cd3aaade4fe6cf59b", "submitted_at": "2026-10-14T15:53:48Z", "updated_at": "2026-10-14T15:53:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20390v1"}}, "environments": ["web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.20390"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20390"}], "name": "STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents", "publication": null, "published_at": "2026-10-14", "record_sha256": "80498aa029306c30caa146eda1bc9da2035d592d5c55fa036aa2ceae8129c435", "relations": [], "resource_id": "arxiv:2610.20390", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.20390v1", "topics": ["efficiency", "grounding", "screen-understanding"], "updated_at": "2026-10-14"}, "text": "STP-GUI: Spatially-Aware Token Pruning for Efficient Web-GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces STP-GUI, which selects spatially coherent screenshot windows conditioned on the instruction to reduce visual tokens while retaining useful interface structure.\nPlatforms: web, android\nTopics: efficiency, grounding, screen-understanding\nPublished: 2026-10-14\nAuthors: Shuo Wang, Sifan Zhou, Xiaobo Wang, Tianyu Fu, Chenxu Zhao, Anyang Su, Hanning Wang, Ping Jiang, Minghui Wu\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.20390v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-20390/"}
{"id": "arxiv:2610.20584", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67b539468bd2bbf0d3b49a5a7eed4791a8ddca4b56ad9515d23a29204d1f069f", "submitted_at": "2026-10-14T17:57:12Z", "updated_at": "2026-10-14T17:57:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20584v1"}, "authors": ["Mingyu Ouyang", "Yiran Wang", "Siyuan Hu", "Xingyilang Yin", "Huiqiang Sun", "Guangzhi Wang", "Junfu Pu", "Cheng Cheng", "Honghao Chen", "Ying Shan", "Mike Zheng Shou", "Hwee Tou Ng"], "canonical_source_url": "https://arxiv.org/abs/2610.20584", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-20584/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "67b539468bd2bbf0d3b49a5a7eed4791a8ddca4b56ad9515d23a29204d1f069f", "submitted_at": "2026-10-14T17:57:12Z", "updated_at": "2026-10-14T17:57:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20584v1"}}, "environments": ["unspecified"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:33:24Z", "role": "primary", "sha256": "9641755401e0e8ec9613f7c4509029d968ab4f10751ba4d43cab57e7d27c1da1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2610.11655%2C2610.11707%2C2610.11942%2C2610.12611%2C2610.12823%2C2610.12866%2C2610.12984%2C2610.13268%2C2610.13416%2C2610.13703%2C2610.14157%2C2610.14228%2C2610.15138%2C2610.15765%2C2610.16234%2C2610.16319%2C2610.16720%2C2610.16778%2C2610.17457%2C2610.17710%2C2610.17792%2C2610.17895%2C2610.17928%2C2610.18508%2C2610.18784%2C2610.18868%2C2610.19100%2C2610.19128%2C2610.20584&max_results=29"}], "identifiers": {"arxiv": "2610.20584"}, "kind": "paper", "last_verified_at": "2026-10-16T06:33:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20584"}], "name": "Player-1: Learning to Play from Synthetic Experience", "publication": null, "published_at": "2026-10-14", "record_sha256": "fe884fead09a04832c2871bab3047637fe45de4ba3ade71fb73223a3aef1b2ef", "relations": [], "resource_id": "arxiv:2610.20584", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.20584v1", "topics": ["data-generation", "reinforcement-learning", "evaluation"], "updated_at": "2026-10-14"}, "text": "Player-1: Learning to Play from Synthetic Experience\nResource type: paper\nReview status: curated\nScope: adjacent\n\nExtends computer-use policies to game control through synthetic training environments and richer keyboard and pointer actions. Its evaluation concerns games rather than ordinary productivity workflows.\nPlatforms: unspecified\nTopics: data-generation, reinforcement-learning, evaluation\nPublished: 2026-10-14\nAuthors: Mingyu Ouyang, Yiran Wang, Siyuan Hu, Xingyilang Yin, Huiqiang Sun, Guangzhi Wang, Junfu Pu, Cheng Cheng, Honghao Chen, Ying Shan, Mike Zheng Shou, Hwee Tou Ng\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.20584v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-20584/"}
{"id": "arxiv:2610.20939", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5f4eda178718aac211eddc7281d572dc1d189dfd0cd130d4eba9f30f5a05ca2", "submitted_at": "2026-10-14T21:06:01Z", "updated_at": "2026-10-14T21:06:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20939v1"}, "authors": ["Panrong Tong", "Quyu Kong", "Pengxiang Li", "Tianyu Xia", "Hanzhang Zhou", "Lei Wang", "Chenglin Cai", "Long Li", "Long Chen", "Jianan Zhang", "Zhengwen Zeng", "Xu Zhang", "Gongjie Zhang", "Yue Wang", "Steven Hoi"], "canonical_source_url": "https://arxiv.org/abs/2610.20939", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-20939/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "a5f4eda178718aac211eddc7281d572dc1d189dfd0cd130d4eba9f30f5a05ca2", "submitted_at": "2026-10-14T21:06:01Z", "updated_at": "2026-10-14T21:06:01Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.20939v1"}}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.20939"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.20939"}], "name": "MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments", "publication": null, "published_at": "2026-10-14", "record_sha256": "c8de030460fa026deab4d80e7e45614930d5e9506b5ed00f653dd98e80a4f951", "relations": [], "resource_id": "arxiv:2610.20939", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.20939v1", "topics": ["sandbox", "evaluation", "reinforcement-learning"], "updated_at": "2026-10-14"}, "text": "MAI-Gym: An Open All-in-One Platform for GUI Agent Evaluation and RL Training across Mobile and Computer Environments\nResource type: paper\nReview status: curated\nScope: core\n\nUnifies GUI evaluation, trajectory collection, environment synthesis, and online reinforcement learning across mobile, computer, and web tasks.\nPlatforms: cross-platform, desktop, web\nTopics: sandbox, evaluation, reinforcement-learning\nPublished: 2026-10-14\nAuthors: Panrong Tong, Quyu Kong, Pengxiang Li, Tianyu Xia, Hanzhang Zhou, Lei Wang, Chenglin Cai, Long Li, Long Chen, Jianan Zhang, Zhengwen Zeng, Xu Zhang, Gongjie Zhang, Yue Wang, Steven Hoi\nRecorded arXiv version: v1\n\nThe annotation and scope were reviewed against the primary paper abstract. Experimental results and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.20939v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-20939/"}
{"id": "arxiv:2610.22200", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beab15a2bf8a6aca19d219dbe0a047e8c6d6bd84a580df33633b5d366c175a0e", "submitted_at": "2026-10-15T16:51:35Z", "updated_at": "2026-10-15T16:51:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22200v1"}, "authors": ["Hanzhang Zhou", "Long Li", "Quyu Kong", "Lei Wang", "Panrong Tong", "Long Chen", "Tianyu Xia", "Gongjie Zhang", "Xu Zhang", "Chen Liu", "Jianan Zhang", "Chenglin Cai", "Zhengwen Zeng", "Steven Hoi", "Yue Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.22200", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22200/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "beab15a2bf8a6aca19d219dbe0a047e8c6d6bd84a580df33633b5d366c175a0e", "submitted_at": "2026-10-15T16:51:35Z", "updated_at": "2026-10-15T16:51:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22200v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.22200"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22200"}], "name": "OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use", "publication": null, "published_at": "2026-10-15", "record_sha256": "198ddcea51a71e3c7dd7c819fb4e0e72be0f660dadde96beec9b68cd1552515b", "relations": [], "resource_id": "arxiv:2610.22200", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22200v1", "topics": ["tool-integration", "planning"], "updated_at": "2026-10-15"}, "text": "OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces an open generalist computer-use stack combining visual and programmatic interaction, with model-controlled observation and task-state decisions.\nPlatforms: desktop\nTopics: tool-integration, planning\nPublished: 2026-10-15\nAuthors: Hanzhang Zhou, Long Li, Quyu Kong, Lei Wang, Panrong Tong, Long Chen, Tianyu Xia, Gongjie Zhang, Xu Zhang, Chen Liu, Jianan Zhang, Chenglin Cai, Zhengwen Zeng, Steven Hoi, Yue Wang\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.22200v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22200/"}
{"id": "arxiv:2610.22234", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b7f3542ca2faa63df81e25721a01928bfb35e9e3de458e8be27fd3f9010adc8", "submitted_at": "2026-10-15T17:14:02Z", "updated_at": "2026-10-15T17:14:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22234v1"}, "authors": ["Bowen Yang", "Kaiming Jin", "Zichen Ding", "Zhenyu Wu", "Zehao Li", "Zhoumianze Liu", "Kanzhi Cheng", "Fangzhi Xu", "Yanan Sun", "Qiushi Sun", "Yu Qiao"], "canonical_source_url": "https://arxiv.org/abs/2610.22234", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22234/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "2b7f3542ca2faa63df81e25721a01928bfb35e9e3de458e8be27fd3f9010adc8", "submitted_at": "2026-10-15T17:14:02Z", "updated_at": "2026-10-15T17:14:02Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22234v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.22234"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22234"}], "name": "OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents", "publication": null, "published_at": "2026-10-15", "record_sha256": "e4b7db0bfaec9fa809f7019ef0a32abeb23869bfe3d5624b7e65d4fc46bb05e0", "relations": [], "resource_id": "arxiv:2610.22234", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22234v1", "topics": ["tool-integration", "reinforcement-learning"], "updated_at": "2026-10-15"}, "text": "OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents\nResource type: paper\nReview status: curated\nScope: core\n\nTrains native computer-use models to coordinate GUI actions and code execution using hybrid demonstrations and online reinforcement learning.\nPlatforms: desktop, linux\nTopics: tool-integration, reinforcement-learning\nPublished: 2026-10-15\nAuthors: Bowen Yang, Kaiming Jin, Zichen Ding, Zhenyu Wu, Zehao Li, Zhoumianze Liu, Kanzhi Cheng, Fangzhi Xu, Yanan Sun, Qiushi Sun, Yu Qiao\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.22234v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22234/"}
{"id": "arxiv:2610.22280", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a249a56a61ebb14b456500f9b245efe7b72997af5d5f515e5cc7707fe68b018", "submitted_at": "2026-10-15T17:40:12Z", "updated_at": "2026-10-15T17:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22280v1"}, "authors": ["Qingni Wang", "Changdae Oh", "Yuzhe Yang", "Sharon Li", "Jiachen Yang", "Ang Li", "Xin Eric Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.22280", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22280/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6a249a56a61ebb14b456500f9b245efe7b72997af5d5f515e5cc7707fe68b018", "submitted_at": "2026-10-15T17:40:12Z", "updated_at": "2026-10-15T17:40:12Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22280v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {"arxiv": "2610.22280"}, "kind": "paper", "last_verified_at": "2026-10-16T06:36:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22280"}], "name": "LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents", "publication": null, "published_at": "2026-10-15", "record_sha256": "df30c62303ae1501c99ebf1bf4eb23dfeb8a4a37af4df8b67728244ad946a5df", "relations": [], "resource_id": "arxiv:2610.22280", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22280v1", "topics": ["efficiency", "planning"], "updated_at": "2026-10-15"}, "text": "LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nUses weak-agent uncertainty to decide when to escalate to a stronger model and monitor its subsequent task execution.\nPlatforms: desktop\nTopics: efficiency, planning\nPublished: 2026-10-15\nAuthors: Qingni Wang, Changdae Oh, Yuzhe Yang, Sharon Li, Jiachen Yang, Ang Li, Xin Eric Wang\nRecorded arXiv version: v1\n\nEditorial summary based on the primary paper abstract; experiments and released artifacts have not been independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.22280v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22280/"}
{"id": "arxiv:2610.22408", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "025c7018c86ae4de79a659242070b2f91c5913333e9ecd4bd7fd357d13d27a34", "submitted_at": "2026-10-14T14:39:41Z", "updated_at": "2026-10-14T14:39:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22408v1"}, "authors": ["Tengda Guo", "Jiawang Hu", "Yifei Tao", "Yin Wang", "Haotian Hu", "Fanyi Wang", "Liujuan Tang"], "canonical_source_url": "https://arxiv.org/abs/2610.22408", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22408/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "025c7018c86ae4de79a659242070b2f91c5913333e9ecd4bd7fd357d13d27a34", "submitted_at": "2026-10-14T14:39:41Z", "updated_at": "2026-10-14T14:39:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22408v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:20.217517Z", "role": "primary", "sha256": "f275c47a8b6d349a70e39a9e55eea5a09d99d5f35cc896100f3bcfd5b0923e0b", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28abs%3A%22computer-use+agents%22+OR+abs%3A%22GUI+agents%22+OR+abs%3A%22browser+agents%22+OR+abs%3A%22mobile+GUI%22%29+AND+submittedDate%3A%5B202609200000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:54.814500Z", "role": "primary", "sha256": "0389d96a8f6c4975129f5c526e8df9ca4690f31364a2f4e66aaea4a75f0eabbc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22408v1"}], "identifiers": {"arxiv": "2610.22408"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:54.814500Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22408"}], "name": "ECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation", "publication": null, "published_at": "2026-10-14", "record_sha256": "7c6d98c211bed8b533e9d9965cdfca32e442c546b6a49298bb57bb7b02eb316d", "relations": [], "resource_id": "arxiv:2610.22408", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22408v1", "topics": ["memory", "interaction-history", "state-tracking", "long-horizon", "imitation-learning", "reinforcement-learning", "efficiency", "evaluation"], "updated_at": "2026-10-14"}, "text": "ECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation\nResource type: paper\nReview status: curated\nScope: core\n\nECHOES learns recursive textual task memory through privileged screenshot-history distillation and GRPO, then predicts GUI actions from the current screenshot and memory.\nPlatforms: android\nTopics: memory, interaction-history, state-tracking, long-horizon, imitation-learning, reinforcement-learning, efficiency, evaluation\nPublished: 2026-10-14\nAuthors: Tengda Guo, Jiawang Hu, Yifei Tao, Yin Wang, Haotian Hu, Fanyi Wang, Liujuan Tang\nRecorded arXiv version: v1\n\nPrivileged screenshot history is used during training, while deployment retains self-generated textual state. GUI results use action/step metrics and model-specific history settings; online Sokoban transfer is separate. GRPO rewards action correctness, valid output format and compression length. Diagnostic memory replay is distinct from the validation gap used to schedule training. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.22408v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22408/"}
{"id": "arxiv:2610.22539", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f064032a0b0d673273d90a9250c76a6b823b1728761635b8a555a0bd96b888a", "submitted_at": "2026-10-15T18:37:41Z", "updated_at": "2026-10-15T18:37:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22539v1"}, "authors": ["Md Farhan Ishmam", "Minh Pham Dinh", "Kenneth Marino"], "canonical_source_url": "https://arxiv.org/abs/2610.22539", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22539/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0f064032a0b0d673273d90a9250c76a6b823b1728761635b8a555a0bd96b888a", "submitted_at": "2026-10-15T18:37:41Z", "updated_at": "2026-10-15T18:37:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22539v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:34:07.486792+00:00", "role": "primary", "sha256": "0c95ddb42bbfc91aabf25e2dc2d2393a9977b6d979a2c891c5bcf80798dbfbc3", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22539v1"}], "identifiers": {"arxiv": "2610.22539"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22539"}], "name": "Don't Retrain, Remix: Adaptation of Web Agents using Mixture Hypernetworks", "publication": null, "published_at": "2026-10-15", "record_sha256": "e6cd3c84b79af5161c35c26c9ddfde45afa90770e0f702b07d6dfa6cf53ab542", "relations": [], "resource_id": "arxiv:2610.22539", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22539v1", "topics": ["web-navigation", "imitation-learning", "robustness"], "updated_at": "2026-10-15"}, "text": "Don't Retrain, Remix: Adaptation of Web Agents using Mixture Hypernetworks\nResource type: paper\nReview status: curated\nScope: core\n\nWebMix uses a scout-produced interface manual to route mixtures of source-trained LoRA adapters for web agents, adapting to target interfaces without target-time gradient updates.\nPlatforms: web\nTopics: web-navigation, imitation-learning, robustness\nPublished: 2026-10-15\nAuthors: Md Farhan Ishmam, Minh Pham Dinh, Kenneth Marino\nRecorded arXiv version: v1\n\nReviewed the complete retained abstract and selected exact-v1 method, evaluation and analysis sections. Adapter-bank and router training use source demonstrations. Domain scouting accesses target training-task intents and answer-stripped reference procedures; version scouting is task-free. No target-time gradients does not mean no target task information. No independent reproduction or artifact/license verification.\n\nPrimary source: https://arxiv.org/abs/2610.22539v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22539/"}
{"id": "arxiv:2610.22672", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "caaf24ef4d8b86c8aaebc852235d0fe947a5ed859dc1ba8329ce07a2f7be9aa1", "submitted_at": "2026-10-15T21:40:34Z", "updated_at": "2026-10-15T21:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22672v1"}, "authors": ["Adamo Orsini", "Seyed Mehdi Ayyoubzadeh", "Sadman Rakib Pinon", "Hannuo Zhang", "Koosha Gholipour Baradari", "Yuanhao Yu", "Xinxin Zuo", "Yang Wang"], "canonical_source_url": "https://arxiv.org/abs/2610.22672", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22672/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "caaf24ef4d8b86c8aaebc852235d0fe947a5ed859dc1ba8329ce07a2f7be9aa1", "submitted_at": "2026-10-15T21:40:34Z", "updated_at": "2026-10-15T21:40:34Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22672v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:56:56.464343+00:00", "role": "primary", "sha256": "5fd9d7926e99d068fa8fb7238a9e7811e3f8c55d00e6b89d5672754cb3e619bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22672v1"}], "identifiers": {"arxiv": "2610.22672"}, "kind": "paper", "last_verified_at": "2026-10-20T12:56:56.464343+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22672"}], "name": "Not All Graphs Are Maps: A Task-Agnostic Exploration Framework for Mobile GUI Navigation", "publication": null, "published_at": "2026-10-15", "record_sha256": "d9fd88704ed3c0b3431c25ba49ca57f6d8087da59ea077942da28051ae91da40", "relations": [], "resource_id": "arxiv:2610.22672", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22672v1", "topics": ["mobile-navigation", "memory", "evaluation"], "updated_at": "2026-10-15"}, "text": "Not All Graphs Are Maps: A Task-Agnostic Exploration Framework for Mobile GUI Navigation\nResource type: paper\nReview status: curated\nScope: core\n\nTAG-Nav supplies Android GUI agents with destination and route memory distilled from offline app exploration. DestBench measures arrival at a reference pre-action screen, separately from downstream task execution.\nPlatforms: android\nTopics: mobile-navigation, memory, evaluation\nPublished: 2026-10-15\nAuthors: Adamo Orsini, Seyed Mehdi Ayyoubzadeh, Sadman Rakib Pinon, Hannuo Zhang, Koosha Gholipour Baradari, Yuanhao Yu, Xinxin Zuo, Yang Wang\nRecorded arXiv version: v1\n\nThe controlled cohort has 84 tasks across six apps; the broader inventory spans 20 apps. Main results use human destination labels. DroidTask and SPA-Bench are converted to navigation-only protocols. Some comparison cells received another graph condition’s memory, with sensitivity analyses reported. Cross-family runs differ in app versions, state and denominators. A small stale-graph probe does not establish robustness. Models, code, artifacts and performance were not independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.22672v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22672/"}
{"id": "arxiv:2610.22697", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "077dd623c980813e83197a38b6545161a3d2fc842d99fe291c53372b9770acd2", "submitted_at": "2026-10-15T22:02:45Z", "updated_at": "2026-10-15T22:02:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22697v1"}, "authors": ["Umar Iqbal"], "canonical_source_url": "https://arxiv.org/abs/2610.22697", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22697/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "077dd623c980813e83197a38b6545161a3d2fc842d99fe291c53372b9770acd2", "submitted_at": "2026-10-15T22:02:45Z", "updated_at": "2026-10-15T22:02:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22697v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:38:46.444001Z", "role": "primary", "sha256": "cb6a8a9c0cc47be21b36b3d1831b26c6189d8f2e4af0d7e26d643c9c2a65452f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22697v1"}], "identifiers": {"arxiv": "2610.22697"}, "kind": "paper", "last_verified_at": "2026-10-20T13:38:46.444001Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22697"}], "name": "What It Is, Not What It Says: Securing Agentic Browsing Interactions with Web Semantics", "publication": null, "published_at": "2026-10-15", "record_sha256": "fbee6e48d73d94a345f438a8c7507357c4bc6cf663b5a75c3af8770037d81217", "relations": [], "resource_id": "arxiv:2610.22697", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22697v1", "topics": ["safety-security", "web-navigation", "planning", "human-in-the-loop"], "updated_at": "2026-10-15"}, "text": "What It Is, Not What It Says: Securing Agentic Browsing Interactions with Web Semantics\nResource type: paper\nReview status: curated\nScope: core\n\nIntact records webpage element provenance and capabilities, isolates planning per element, and uses a browser harness to constrain actions and request consent for third-party interactions.\nPlatforms: web\nTopics: safety-security, web-navigation, planning, human-in-the-loop\nPublished: 2026-10-15\nAuthors: Umar Iqbal\nRecorded arXiv version: v1\n\nEvaluation instructs the model to simulate compromise and measures planner outputs and permission appearance. Utility uses plan matching and LLM-judged element suitability; no steps are executed in that assessment. Assumes an honest first party, trusted browser/harness and untampered instrumentation; multi-page and availability attacks are outside scope. Parallel latency reduction is an unimplemented estimate. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.22697v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22697/"}
{"id": "arxiv:2610.22990", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f613ef7ada260b494a411058bc733c7a5e0b074cdb01906c38132acf7275b9b1", "submitted_at": "2026-10-16T06:14:55Z", "updated_at": "2026-10-16T06:14:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22990v1"}, "authors": ["Haoming Li", "Defu Cao", "Xu Hu", "Yingtao Luo", "Yiyang Wang", "Xianghe Xu", "Jiashu He", "Wanting Jiang", "Yifan Zhang", "Yunhan Zhao", "Siyang Wu", "Chen Yang", "Haixin Wang", "Zheyuan Deng", "Taoran Li", "Jingwen Zhang", "Zhejian Peng", "Shushu Liang", "Yue Zhao", "Bingzhe Li", "Feng Chen", "Yifan Wang", "Jintao Huang", "Yijun Wang", "Yuexing Hao", "Xiaomin Li"], "canonical_source_url": "https://arxiv.org/abs/2610.22990", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-22990/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "f613ef7ada260b494a411058bc733c7a5e0b074cdb01906c38132acf7275b9b1", "submitted_at": "2026-10-16T06:14:55Z", "updated_at": "2026-10-16T06:14:55Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.22990v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:42:06.870916+00:00", "role": "primary", "sha256": "bf9066e8b3472100f120e21534f92655f4d0814f2521b9d23d778e5239c7a2a4", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.22990v1"}], "identifiers": {"arxiv": "2610.22990"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.22990"}], "name": "PersonaCUABench: Benchmarking Computer-Use Agents Serving Diverse Personas", "publication": null, "published_at": "2026-10-16", "record_sha256": "558e4c6d9046468403751cb900543e21ce89df5afe9bb18d791931cfccefe422", "relations": [], "resource_id": "arxiv:2610.22990", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.22990v1", "topics": ["evaluation", "web-navigation", "human-in-the-loop"], "updated_at": "2026-10-16"}, "text": "PersonaCUABench: Benchmarking Computer-Use Agents Serving Diverse Personas\nResource type: paper\nReview status: curated\nScope: core\n\nPersonaCUABench varies simulated user requests and clarification dialogue over fixed web tasks, comparing task completion with requirement handling and persona-agent satisfaction ratings.\nPlatforms: web\nTopics: evaluation, web-navigation, human-in-the-loop\nPublished: 2026-10-16\nAuthors: Haoming Li, Defu Cao, Xu Hu, Yingtao Luo, Yiyang Wang, Xianghe Xu, Jiashu He, Wanting Jiang, Yifan Zhang, Yunhan Zhao, Siyang Wu, Chen Yang, Haixin Wang, Zheyuan Deng, Taoran Li, Jingwen Zhang, Zhejian Peng, Shushu Liang, Yue Zhao, Bingzhe Li, Feng Chen, Yifan Wang, Jintao Huang, Yijun Wang, Yuexing Hao, Xiaomin Li\nRecorded arXiv version: v1\n\nReviewed the complete retained abstract and selected exact-v1 construction, evaluation and limitations sections. Main interactions and satisfaction ratings use LLM-simulated personas; human annotators validate selected outputs. The harness observes page text, with screenshots only for auditing. Results are not evidence about real demographic populations or desktop agents. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.22990v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-22990/"}
{"id": "arxiv:2610.23012", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9585696d256bacc710dce57bfd7cda6a8037ce5725a30f959028352bc07e97c", "submitted_at": "2026-10-16T06:44:56Z", "updated_at": "2026-10-16T06:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23012v1"}, "authors": ["Chulhoon Yoon", "Kyungri Park", "Jinyeong Yun", "Daeyun Jeong", "Sukmin Yun", "Woohwan Jung"], "canonical_source_url": "https://arxiv.org/abs/2610.23012", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-23012/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "b9585696d256bacc710dce57bfd7cda6a8037ce5725a30f959028352bc07e97c", "submitted_at": "2026-10-16T06:44:56Z", "updated_at": "2026-10-16T06:44:56Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23012v1"}}, "environments": ["desktop"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-19T04:42:05Z", "role": "primary", "sha256": "ffbb3f28a3a0fc8888f30ea8b2cdaf441239685aaacc2d80f311581f4b5a1d5c", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610050000+TO+202610190000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:34:04.380045+00:00", "role": "primary", "sha256": "75c025a862a046b8bcbdd84b800a37132edcc1ec271d40c57c18ec984652a0b0", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23012v1"}], "identifiers": {"arxiv": "2610.23012"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23012"}], "name": "Adaptation of Computer Use Agents to New Software with User Manuals", "publication": null, "published_at": "2026-10-16", "record_sha256": "90bb88f618dea8a3f46b7e99f74891ee90a26fff9183d97147a2d4cbc697da7d", "relations": [], "resource_id": "arxiv:2610.23012", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.23012v1", "topics": ["planning", "data-generation", "reinforcement-learning"], "updated_at": "2026-10-16"}, "text": "Adaptation of Computer Use Agents to New Software with User Manuals\nResource type: paper\nReview status: curated\nScope: core\n\nMAGE retrieves software-manual sections into fixed execution plans and generates setup/main task-plan pairs whose GUI rollouts support application-specific policy adaptation with round-based GRPO.\nPlatforms: desktop\nTopics: planning, data-generation, reinforcement-learning\nPublished: 2026-10-16\nAuthors: Chulhoon Yoon, Kyungri Park, Jinyeong Yun, Daeyun Jeong, Sukmin Yun, Woohwan Jung\nRecorded arXiv version: v1\n\nReviewed the complete retained abstract and selected exact-v1 methods, evaluation and validation sections. Execution plans stay fixed per episode; policy training is separate and application-specific. Screenshot-based training/Slack judgments can miss task conditions, and OSWorld tasks may overlap backbone pretraining. No independent reproduction, safe-deployment guarantee or artifact-release verification.\n\nPrimary source: https://arxiv.org/abs/2610.23012v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-23012/"}
{"id": "arxiv:2610.23303", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea753c1cb0666950d89f2781511f26872e3d5b1c2fd2434896248214160fa22c", "submitted_at": "2026-10-16T11:06:37Z", "updated_at": "2026-10-16T11:06:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23303v1"}, "authors": ["Songping Wang", "Borui Li", "Qinglong Liu", "Yueming Lyu", "Yi Huang", "Fengting Li", "Chao Feng", "Caifeng Shan"], "canonical_source_url": "https://arxiv.org/abs/2610.23303", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-23303/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "ea753c1cb0666950d89f2781511f26872e3d5b1c2fd2434896248214160fa22c", "submitted_at": "2026-10-16T11:06:37Z", "updated_at": "2026-10-16T11:06:37Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23303v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T10:42:09.967883+00:00", "role": "primary", "sha256": "ec9d647818be346ddbf7807be87b4789cd53e50c5d3d7a7904ad4f49bc71c38e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23303v1"}], "identifiers": {"arxiv": "2610.23303"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23303"}], "name": "MobileJailBench: Red Queen Jailbreak Benchmarking for Mobile Agents", "publication": null, "published_at": "2026-10-16", "record_sha256": "f166ba6707b02edb436eba9ef814ccf35ae2200ff20ee5d50541e5869bfae924", "relations": [], "resource_id": "arxiv:2610.23303", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.23303v1", "topics": ["safety-security", "mobile-navigation", "evaluation", "observability", "cross-app"], "updated_at": "2026-10-16"}, "text": "MobileJailBench: Red Queen Jailbreak Benchmarking for Mobile Agents\nResource type: paper\nReview status: curated\nScope: core\n\nMobileJailBench evaluates mobile-agent jailbreaks in a browser-based Android-like simulator, combining feedback-driven attack evolution, separately admitted task expansions and evidence-grounded outcome adjudication.\nPlatforms: web\nTopics: safety-security, mobile-navigation, evaluation, observability, cross-app\nPublished: 2026-10-16\nAuthors: Songping Wang, Borui Li, Qinglong Liu, Yueming Lyu, Yi Huang, Fengting Li, Chao Feng, Caifeng Shan\nRecorded arXiv version: v1\n\nReviewed the complete retained abstract and selected exact-v1 threat-model, construction and evaluation sections. Execution uses a browser-based Android-like simulator, not real Android isolation or live services. The attacker controls instructions and designated carriers, not privileged simulator state/APIs or system/model/harness internals. Rates depend on release, budget and adjudication coverage. SAF labels track observable expressions and actions, not latent reasoning; unresolved episodes are excluded from rate denominators. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.23303v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-23303/"}
{"id": "arxiv:2610.23307", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fc22fbeda19ac2947b00af73801f0674c550e0d7c78919f811cae754b06128c", "submitted_at": "2026-10-16T11:10:30Z", "updated_at": "2026-10-16T11:10:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23307v1"}, "authors": ["Songping Wang", "Peiyang Li", "Yi Huang", "Fengting Li", "Xuhao Ran", "Yueming Lyu", "Chao Feng", "Caifeng Shan"], "canonical_source_url": "https://arxiv.org/abs/2610.23307", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-23307/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "3fc22fbeda19ac2947b00af73801f0674c550e0d7c78919f811cae754b06128c", "submitted_at": "2026-10-16T11:10:30Z", "updated_at": "2026-10-16T11:10:30Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23307v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:01Z", "role": "primary", "sha256": "63e37cd26bfe69455a3fa835f3d6ba4cfcc5bb7dc91f90505920d37b1e08f0b6", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.19384%2C2610.22672%2C2610.22697%2C2610.23303%2C2610.23307%2C2610.24729&max_results=6"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:36.223826+00:00", "role": "primary", "sha256": "3803984a185b6258276b87b87ac1991583878fb41dc22618ff743f6f69c7bc7e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23307v1"}], "identifiers": {"arxiv": "2610.23307"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23307"}], "name": "MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection", "publication": null, "published_at": "2026-10-16", "record_sha256": "b29843dded17e0ea3877396efac840020150e711a651ad99db83bba8524492d1", "relations": [], "resource_id": "arxiv:2610.23307", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.23307v1", "topics": ["safety-security", "mobile-navigation", "evaluation", "observability", "cross-app"], "updated_at": "2026-10-16"}, "text": "MobileInject: Evaluating Reasoning-to-Action Security of MLLM-Based Mobile Agents under Cross-Modal Consistency Injection\nResource type: paper\nReview status: curated\nScope: core\n\nMobileInject evaluates visual/textual observation injection in stateful Android workflows using resettable emulators, benign/adversarial state oracles and trace-based diagnosis of emitted rationales and actions.\nPlatforms: android\nTopics: safety-security, mobile-navigation, evaluation, observability, cross-app\nPublished: 2026-10-16\nAuthors: Songping Wang, Peiyang Li, Yi Huang, Fengting Li, Xuhao Ran, Yueming Lyu, Chao Feng, Caifeng Shan\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 construction, threat-model and evaluation sections. The attacker modifies observation surfaces, not system instructions, model parameters or the OS. State oracles are distinct from emitted-rationale diagnoses, whose labels include outcome-dependent constraints. Those diagnostics are not independent evidence of hidden-reasoning causality. No reproduction or release verification.\n\nPrimary source: https://arxiv.org/abs/2610.23307v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-23307/"}
{"id": "arxiv:2610.23986", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dfa0c59ec1262cebcf2a4e61ab1e0b58c647513d8e6eedb2ce7e6a1a675558c", "submitted_at": "2026-10-14T12:46:54Z", "updated_at": "2026-10-14T12:46:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23986v1"}, "authors": ["Xinyu Fang", "Yuchi Wang", "Weihao Yu", "Rongyao Fang", "Zhaohai Li", "Zhixun Li", "Tiancheng Gu", "Xilin Wei", "Yuhang Zang", "Hongsheng Li", "Dahua Lin", "Guofeng Zhang", "Shuai Bai", "Zhibo Yang", "Dayiheng Liu"], "canonical_source_url": "https://arxiv.org/abs/2610.23986", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-23986/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "6dfa0c59ec1262cebcf2a4e61ab1e0b58c647513d8e6eedb2ce7e6a1a675558c", "submitted_at": "2026-10-14T12:46:54Z", "updated_at": "2026-10-14T12:46:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.23986v1"}}, "environments": ["desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T13:48:11Z", "role": "primary", "sha256": "89e110b6fa882639fb73936cb27e5373d11b632938c31968f563152534bf5670", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?id_list=2610.14966%2C2610.23986&max_results=2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T15:09:58.148629Z", "role": "primary", "sha256": "bef4399f087e484b62ff5b8dc2c283f92e0357b94c93b464b86fd97fe7b069cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.23986v1"}], "identifiers": {"arxiv": "2610.23986"}, "kind": "paper", "last_verified_at": "2026-10-20T15:09:58.148629Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.23986"}], "name": "BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References", "publication": null, "published_at": "2026-10-14", "record_sha256": "8b6894ad6a14969a508949101932781132913213f8947d5d1f7de0bdf9a5ea77", "relations": [], "resource_id": "arxiv:2610.23986", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://arxiv.org/abs/2610.23986v1", "topics": ["evaluation", "tool-integration", "sandbox"], "updated_at": "2026-10-14"}, "text": "BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References\nResource type: paper\nReview status: curated\nScope: adjacent\n\nBenchmarks agents creating editable Blender projects through shell, Python and MCP tools, judging delivered scenes with multi-view renders, task-specific rubrics and visual-quality ratings.\nPlatforms: desktop, linux\nTopics: evaluation, tool-integration, sandbox\nPublished: 2026-10-14\nAuthors: Xinyu Fang, Yuchi Wang, Weihao Yu, Rongyao Fang, Zhaohai Li, Zhixun Li, Tiancheng Gu, Xilin Wei, Yuhang Zang, Hongsheng Li, Dahua Lin, Guofeng Zhang, Shuai Bai, Zhibo Yang, Dayiheng Liu\nRecorded arXiv version: v1\n\nThe evaluated agents use Linux shell/Python/MCP tools and can inspect viewport screenshots; graphical CUA control and Windows/macOS are future work. Each agent is run once per task, with different Codex/Claude Code harnesses across systems. Scores combine weighted rubric satisfaction and model-rated visual quality. Tool counts include unsuccessful calls and do not establish matched compute or successful actions. No independent reproduction or artifact-release verification.\n\nPrimary source: https://arxiv.org/abs/2610.23986v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-23986/"}
{"id": "arxiv:2610.24277", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e675a3082f571da547fb5b6cb89e4f68a034bfa57236886e1218cb2bb2cdf1c3", "submitted_at": "2026-10-16T22:12:45Z", "updated_at": "2026-10-16T22:12:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24277v1"}, "authors": ["Egor Zverev", "Juhee Kim", "Leon Trampert", "Boris Köpf", "Aashish Kolluri", "Shruti Tople", "Santiago Zanella-Béguelin"], "canonical_source_url": "https://arxiv.org/abs/2610.24277", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-24277/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e675a3082f571da547fb5b6cb89e4f68a034bfa57236886e1218cb2bb2cdf1c3", "submitted_at": "2026-10-16T22:12:45Z", "updated_at": "2026-10-16T22:12:45Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24277v1"}}, "environments": ["web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:33.132448+00:00", "role": "primary", "sha256": "daceefecb2e6b910a6c7f2ffe35e89c7070f6282f4f35ee89e885a2e77aa94c4", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.24277v1"}], "identifiers": {"arxiv": "2610.24277"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.24277"}], "name": "SPIDES: Information-Flow Control for Browser Agents", "publication": null, "published_at": "2026-10-16", "record_sha256": "9e77704cedecd1f3a72e307d997d9a087d48a69a377a2968df53f2cc0cc7cfab", "relations": [], "resource_id": "arxiv:2610.24277", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.24277v1", "topics": ["safety-security", "web-navigation", "human-in-the-loop", "evaluation"], "updated_at": "2026-10-16"}, "text": "SPIDES: Information-Flow Control for Browser Agents\nResource type: paper\nReview status: curated\nScope: core\n\nSPIDES propagates HTML integrity and confidentiality labels into agent context and mediates browser actions with information-flow policies and approval gates, evaluated in the labeled MicroNet ecosystem.\nPlatforms: web\nTopics: safety-security, web-navigation, human-in-the-loop, evaluation\nPublished: 2026-10-16\nAuthors: Egor Zverev, Juhee Kim, Leon Trampert, Boris Köpf, Aashish Kolluri, Shruti Tople, Santiago Zanella-Béguelin\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 threat-model, policy, evaluation and limitations sections. Guarantees assume correct labels/trusted origins and a trusted browser/harness mediating actions; evaluation uses idealized approvals. Origin-level reader sets can miss recipient distinctions within one origin. No unrestricted prompt-injection immunity or independent reproduction is claimed.\n\nPrimary source: https://arxiv.org/abs/2610.24277v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-24277/"}
{"id": "arxiv:2610.24372", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d6808cf26a609b50b7835ae3ecece0a13580aafcd589823f0b1eeff54732f69", "submitted_at": "2026-10-17T02:03:20Z", "updated_at": "2026-10-17T02:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24372v1"}, "authors": ["Sizhe Tang", "Hanhan Zhou", "Jeffrey Svajlenko", "Rashmi Gangadharaiah"], "canonical_source_url": "https://arxiv.org/abs/2610.24372", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-24372/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9d6808cf26a609b50b7835ae3ecece0a13580aafcd589823f0b1eeff54732f69", "submitted_at": "2026-10-17T02:03:20Z", "updated_at": "2026-10-17T02:03:20Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.24372v1"}}, "environments": ["android", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T12:58:15.922053Z", "role": "primary", "sha256": "b379e14eefa272f62641224fe1d839b7dfbc414d957124b8abeb103444a843ef", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3AGUI+OR+ti%3A%22computer-using%22+OR+ti%3A%22agentic+browsing%22+OR+ti%3A%22visual+grounding%22+OR+ti%3A%22UI+automation%22%29+AND+submittedDate%3A%5B202608210000+TO+202610201255%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:51.579373Z", "role": "primary", "sha256": "9dff18935a1faa7ec16366bac6a6e9e41f0381d561568c3aba39fe783efcfbd3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.24372v1"}], "identifiers": {"arxiv": "2610.24372"}, "kind": "paper", "last_verified_at": "2026-10-20T13:12:51.579373Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.24372"}], "name": "Beyond Next-Screen Prediction: The Decision Value of GUI World Models", "publication": null, "published_at": "2026-10-17", "record_sha256": "fc6ea3e37947652146547472b741a509f66d0639ed97d5b8a21c2ec50336f648", "relations": [], "resource_id": "arxiv:2610.24372", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.24372v1", "topics": ["evaluation", "world-models", "action-prediction", "imitation-learning"], "updated_at": "2026-10-17"}, "text": "Beyond Next-Screen Prediction: The Decision Value of GUI World Models\nResource type: paper\nReview status: curated\nScope: core\n\nSeparates next-screen prediction from action evaluation in GUI lookahead and trains Direct Action Ranking on reference actions, comparing prediction-augmented scoring under matched offline supervision.\nPlatforms: android, web\nTopics: evaluation, world-models, action-prediction, imitation-learning\nPublished: 2026-10-17\nAuthors: Sizhe Tang, Hanhan Zhou, Jeffrey Svajlenko, Rashmi Gangadharaiah\nRecorded arXiv version: v1\n\nThe matched augmentation experiment measures offline ranking; AndroidWorld task-success comparisons use differently trained evaluators and do not isolate prediction value. A separate web evaluator is trained on Mind2Web, and web results do not uniformly favor direct scoring. The main study concerns one-step selection. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.24372v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-24372/"}
{"id": "arxiv:2610.26071", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a297401f7ad4dcd2f485ade6bab295e365280bb803dcecc111aebf0693dd203", "submitted_at": "2026-10-19T09:30:54Z", "updated_at": "2026-10-19T09:30:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26071v1"}, "authors": ["Zhanhui Zhou", "Weichen Zhang", "Haoran Liu", "Lingjie Chen", "Tianneng Shi", "Kevin Lin", "Zhengyuan Yang", "Lijuan Wang", "Dawn Song"], "canonical_source_url": "https://arxiv.org/abs/2610.26071", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26071/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "0a297401f7ad4dcd2f485ade6bab295e365280bb803dcecc111aebf0693dd203", "submitted_at": "2026-10-19T09:30:54Z", "updated_at": "2026-10-19T09:30:54Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26071v1"}}, "environments": ["desktop", "linux", "windows", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:28:55.967592+00:00", "role": "primary", "sha256": "d4231fa9677e916ca8cc27493dfbbe9fb34c7430207150e6b9be10c2bad43bfb", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26071v1"}], "identifiers": {"arxiv": "2610.26071"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26071"}], "name": "CUA-Lite: An Open Platform for Computer Use Agents", "publication": null, "published_at": "2026-10-19", "record_sha256": "207082e4f7d4ba5e77309f0dc0928f3b02236703386cce4b3c1612a75799d1b3", "relations": [], "resource_id": "arxiv:2610.26071", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26071v1", "topics": ["tool-integration", "data-generation", "evaluation", "imitation-learning", "reinforcement-learning", "sandbox"], "updated_at": "2026-10-19"}, "text": "CUA-Lite: An Open Platform for Computer Use Agents\nResource type: paper\nReview status: curated\nScope: core\n\nCUA-Lite unifies environment interfaces, supervised data formats and per-model harnesses for evaluation, reinforcement learning and supervised fine-tuning across desktop, browser and mobile tasks.\nPlatforms: desktop, linux, windows, web, android\nTopics: tool-integration, data-generation, evaluation, imitation-learning, reinforcement-learning, sandbox\nPublished: 2026-10-19\nAuthors: Zhanhui Zhou, Weichen Zhang, Haoran Liu, Lingjie Chen, Tianneng Shi, Kevin Lin, Zhengyuan Yang, Lijuan Wang, Dawn Song\nRecorded arXiv version: v1\n\nReviewed the complete official abstract and selected v1 methods, training and limitations sections. WebVoyager uses offline mirrors; VM-free sandboxes are Linux-only, and reported desktop/mobile gains do not establish unseen-app transfer. No results or released artifacts were independently reproduced.\n\nPrimary source: https://arxiv.org/abs/2610.26071v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26071/"}
{"id": "arxiv:2610.26203", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41178d19a23653c0f49423b8f725584bded7ab8819a1a46e83a062794460b4cb", "submitted_at": "2026-10-19T11:30:48Z", "updated_at": "2026-10-19T11:30:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26203v1"}, "authors": ["Fuyao Zhang", "Wenwen He", "Jiaming Zhang", "Yurong Hao", "Xuan Wang", "Guowei Guan", "Che Wang", "Longtao Huang", "Wei Yang Bryan Lim"], "canonical_source_url": "https://arxiv.org/abs/2610.26203", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26203/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "41178d19a23653c0f49423b8f725584bded7ab8819a1a46e83a062794460b4cb", "submitted_at": "2026-10-19T11:30:48Z", "updated_at": "2026-10-19T11:30:48Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26203v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:23.872790+00:00", "role": "primary", "sha256": "28848f8f597b2a8865ad98f5aa1e4401e15a49f92b33f7bce4d4f044c168ada0", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26203v1"}], "identifiers": {"arxiv": "2610.26203"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26203"}], "name": "Focused Omni-Modal GUI Agents", "publication": null, "published_at": "2026-10-19", "record_sha256": "56e2489fc51f4392818415cd700e8792ddfc59df2f0dd17821d239972ed55d5a", "relations": [], "resource_id": "arxiv:2610.26203", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26203v1", "topics": ["screen-understanding", "grounding", "imitation-learning", "interaction-history", "mobile-navigation"], "updated_at": "2026-10-19"}, "text": "Focused Omni-Modal GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nFocusGUI learns step-specific weights over screenshot, audio, video and interaction-context channels, injecting a compact residual into a frozen decoder without pruning its inputs.\nPlatforms: android\nTopics: screen-understanding, grounding, imitation-learning, interaction-history, mobile-navigation\nPublished: 2026-10-19\nAuthors: Fuyao Zhang, Wenwen He, Jiaming Zhang, Yurong Hao, Xuan Wang, Guowei Guan, Che Wang, Longtao Huang, Wei Yang Bryan Lim\nRecorded arXiv version: v1\n\nReviewed the complete retained abstract and selected v1 method/evaluation sections. OmniGUI uses episode-disjoint reference-action evaluation; Type Match, Exact Match and Goal Progress are not presented as independently demonstrated live task-success gains. The Android role is explicit in Appendix D.1. No reproduction or artifact-release verification.\n\nPrimary source: https://arxiv.org/abs/2610.26203v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26203/"}
{"id": "arxiv:2610.26469", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ef7c9cdda652f60d3f2d4ab89e7218ad5a2b85f5c76f0277c289db06fd2796", "submitted_at": "2026-10-19T14:17:33Z", "updated_at": "2026-10-19T14:17:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26469v1"}, "authors": ["Zhizhou Wang", "Yaxi Lu", "Zhong Zhang", "Zikun Hu", "Xin Cong", "Yankai Lin", "Fangming Liu"], "canonical_source_url": "https://arxiv.org/abs/2610.26469", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26469/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "08ef7c9cdda652f60d3f2d4ab89e7218ad5a2b85f5c76f0277c289db06fd2796", "submitted_at": "2026-10-19T14:17:33Z", "updated_at": "2026-10-19T14:17:33Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26469v1"}}, "environments": ["desktop", "linux", "web", "android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:28:59.073204+00:00", "role": "primary", "sha256": "f13f0225c721d5aa430bb8aea1d1e7a7da5973760c4c3b82f26ba648aade82ba", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26469v1"}], "identifiers": {"arxiv": "2610.26469"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26469"}], "name": "VeriCUA: Learning Proactive Verification for Reliable GUI Agent Rewards", "publication": null, "published_at": "2026-10-19", "record_sha256": "75155e47fa52f9c3746f1eb46eaf85a78a4ad17860320c58d9b259910346747a", "relations": [], "resource_id": "arxiv:2610.26469", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26469v1", "topics": ["reward-models", "reinforcement-learning", "evaluation", "efficiency"], "updated_at": "2026-10-19"}, "text": "VeriCUA: Learning Proactive Verification for Reliable GUI Agent Rewards\nResource type: paper\nReview status: curated\nScope: core\n\nVeriCUA trains a Judge to select verification checks and stopping decisions while a frozen Probe inspects the post-task GUI. Single-attempt reinforcement learning balances verdict correctness and Probe use.\nPlatforms: desktop, linux, web, android\nTopics: reward-models, reinforcement-learning, evaluation, efficiency\nPublished: 2026-10-19\nAuthors: Zhizhou Wang, Yaxi Lu, Zhong Zhang, Zikun Hu, Xin Cong, Yankai Lin, Fangming Liu\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 method/evaluation sections. The Judge is trained on OSWorld and frozen for MobileWorld/WebArena transfer; downstream Actor RL is separate. Evaluation uses screened terminal outcomes, and reported cost counts Probe tokens. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.26469v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26469/"}
{"id": "arxiv:2610.26503", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7dc2ceb9aadbf7506efc068df99326fea5b60fdc6feb340c490841b88aa19c30", "submitted_at": "2026-10-19T14:38:41Z", "updated_at": "2026-10-19T14:38:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26503v1"}, "authors": ["Zhicheng Zheng", "Cheng Zhang", "Shengping Yu", "Tianxiang Zhao"], "canonical_source_url": "https://arxiv.org/abs/2610.26503", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26503/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "7dc2ceb9aadbf7506efc068df99326fea5b60fdc6feb340c490841b88aa19c30", "submitted_at": "2026-10-19T14:38:41Z", "updated_at": "2026-10-19T14:38:41Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26503v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:29:02.225995+00:00", "role": "primary", "sha256": "0554aa531a5f94f51cba6c51bd1b13fb05b0f462497afa5d1dae7ec6fc6e375a", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26503v1"}], "identifiers": {"arxiv": "2610.26503"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26503"}], "name": "ScopeMem: Latent Memory That Learns Where to Revise Frozen GUI Agents", "publication": null, "published_at": "2026-10-19", "record_sha256": "d1b5eb369cc28574e96691855c0bba3e3c39f397c86229a98b5b6ca26b47a14c", "relations": [], "resource_id": "arxiv:2610.26503", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26503v1", "topics": ["memory", "interaction-history", "imitation-learning", "robustness", "evaluation"], "updated_at": "2026-10-19"}, "text": "ScopeMem: Latent Memory That Learns Where to Revise Frozen GUI Agents\nResource type: paper\nReview status: curated\nScope: core\n\nScopeMem conditions latent-memory prefixes on a frozen GUI agent's initial action and trains them to correct errors while preserving correct decisions. Its evaluation separates action repairs from harmful revisions.\nPlatforms: android\nTopics: memory, interaction-history, imitation-learning, robustness, evaluation\nPublished: 2026-10-19\nAuthors: Zhicheng Zheng, Cheng Zhang, Shengping Yu, Tianxiang Zhao\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 controlled-study, method and limitations sections. Results are offline, step-level action matching; online task success is explicitly left untested. The frozen actor is queried twice per step. No independent reproduction or redistribution-license claim.\n\nPrimary source: https://arxiv.org/abs/2610.26503v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26503/"}
{"id": "arxiv:2610.26660", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e91c0f0ab28f7f2da89bdff5f9ce3dc592469a8606f35137144f514e70c96f78", "submitted_at": "2026-10-19T16:18:16Z", "updated_at": "2026-10-19T16:18:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26660v1"}, "authors": ["Zhuomin Chen", "Jingchao Ni", "Xu Zheng", "Yong Chen", "Mo Sha", "Wei Cheng", "Dongsheng Luo"], "canonical_source_url": "https://arxiv.org/abs/2610.26660", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26660/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "e91c0f0ab28f7f2da89bdff5f9ce3dc592469a8606f35137144f514e70c96f78", "submitted_at": "2026-10-19T16:18:16Z", "updated_at": "2026-10-19T16:18:16Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26660v1"}}, "environments": ["android"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:26.960480+00:00", "role": "primary", "sha256": "7cb553f8230cfb203df446f08e29d48b676addfe9b076d35171d09970d1970bf", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26660v1"}], "identifiers": {"arxiv": "2610.26660"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26660"}], "name": "Tandem: Serving Hybrid Linear-Attention GUI Agents on Mobile NPUs", "publication": null, "published_at": "2026-10-19", "record_sha256": "374cc133393af2253958f3a09ab8f85a2d2a177d3cba1ff3eee0323a28af6ee8", "relations": [], "resource_id": "arxiv:2610.26660", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26660v1", "topics": ["efficiency", "mobile-navigation"], "updated_at": "2026-10-19"}, "text": "Tandem: Serving Hybrid Linear-Attention GUI Agents on Mobile NPUs\nResource type: paper\nReview status: curated\nScope: core\n\nTandem serves hybrid GUI-agent models on Android NPUs using divergence-point state checkpoints, rollback-aware speculative decoding and idle power release between request bursts.\nPlatforms: android\nTopics: efficiency, mobile-navigation\nPublished: 2026-10-19\nAuthors: Zhuomin Chen, Jingchao Ni, Xu Zheng, Yong Chen, Mo Sha, Wei Cheng, Dongsheng Luo\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 system/evaluation sections. Main latency measurements replay recorded requests and exclude app execution; the paper also reports instrumented real tasks. These are serving-time, energy and action-fidelity results, not a claimed task-success improvement. Hardware, cooling and model configurations matter. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.26660v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26660/"}
{"id": "arxiv:2610.26703", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9858b5ec15ee953504bd00e26b455c1e81104f80429c2dd112de17f9d23977ab", "submitted_at": "2026-10-19T16:39:11Z", "updated_at": "2026-10-19T16:39:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26703v1"}, "authors": ["Ananya Gubbi Mohanbabu", "Rosiana Natalie", "Anhong Guo", "Amy Pavel"], "canonical_source_url": "https://arxiv.org/abs/2610.26703", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26703/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "9858b5ec15ee953504bd00e26b455c1e81104f80429c2dd112de17f9d23977ab", "submitted_at": "2026-10-19T16:39:11Z", "updated_at": "2026-10-19T16:39:11Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26703v1"}}, "environments": ["desktop", "web"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T09:17:30.049627+00:00", "role": "primary", "sha256": "9a964cbf50123887ad20bf156e69c0bbe4567d6b2c28e4cf202d5538eab8294c", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26703v1"}], "identifiers": {"arxiv": "2610.26703"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26703"}], "name": "Comparing Computer Use Behavior Across Multiple Levels of Abstraction", "publication": null, "published_at": "2026-10-19", "record_sha256": "55ef0c05222ef380bba09a9a795e20bd8552c723a260e273cd6e44560f826bf2", "relations": [], "resource_id": "arxiv:2610.26703", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26703v1", "topics": ["evaluation", "observability"], "updated_at": "2026-10-19"}, "text": "Comparing Computer Use Behavior Across Multiple Levels of Abstraction\nResource type: paper\nReview status: curated\nScope: core\n\nMAP annotates computer-use traces at linked operation, function/consequence, method and goal levels, then compares human and agent workflows through sequence, pattern and effort metrics.\nPlatforms: desktop, web\nTopics: evaluation, observability\nPublished: 2026-10-19\nAuthors: Ananya Gubbi Mohanbabu, Rosiana Natalie, Anhong Guo, Amy Pavel\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 taxonomy, annotation, evaluation and limitations sections. MAP analyzes existing human/agent traces rather than deploying a new acting policy. Annotation and timing coverage depend on the source datasets; manual agreement reflects a particular labeling scheme. Behavioral findings do not establish a privacy guarantee. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.26703v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26703/"}
{"id": "arxiv:2610.26828", "metadata": {"aliases": [], "arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f1ff33859352b61f017709fff91fd8e32f08636e9423fb044e7ca8e28f06905", "submitted_at": "2026-10-19T17:55:35Z", "updated_at": "2026-10-19T17:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26828v1"}, "authors": ["Zhaoyang Wang", "Qianhui Wu", "Haorui Wang", "Simon Yu", "Baolin Peng", "Canwen Xu", "Hao Cheng", "Wenlin Yao", "Boyi Liu", "Yite Wang", "Jianfeng Gao", "Huaxiu Yao"], "canonical_source_url": "https://arxiv.org/abs/2610.26828", "citation_url": "https://awesomecua.github.io/resources/arxiv-2610-26828/", "details": {"arxiv_snapshot": {"abstract_hash_method": "sha256-utf8-whitespace-collapsed-v1", "abstract_sha256": "1f1ff33859352b61f017709fff91fd8e32f08636e9423fb044e7ca8e28f06905", "submitted_at": "2026-10-19T17:55:35Z", "updated_at": "2026-10-19T17:55:35Z", "version": 1, "versioned_url": "https://arxiv.org/abs/2610.26828v1"}}, "environments": ["web", "desktop", "linux"], "evidence": [{"method": "arxiv_api", "retrieved_at": "2026-10-20T05:52:00Z", "role": "primary", "sha256": "af310824eea97aac8209023ff53bad7bea77778816586b10bb94110148347af9", "supports": ["name", "authors", "published_at", "updated_at", "identifiers.arxiv", "details.arxiv_snapshot"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B202610060000+TO+202610200000%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:27:26.236797+00:00", "role": "primary", "sha256": "80bed80786f939973d0d13b4f892300fee02a6110e326247131fc83d257796de", "supports": ["name", "kind", "summary", "topics", "scope"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agent%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22+OR+ti%3A%22browser+agents%22+OR+ti%3A%22mobile+agent%22+OR+ti%3A%22mobile+agents%22+OR+ti%3A%22desktop+agent%22%29+AND+submittedDate%3A%5B202610130000+TO+202610192359%5D&start=0&max_results=200&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T08:29:05.337543+00:00", "role": "primary", "sha256": "b680e07a5a883e7bb0210a1734a628be6c47d91b7df0487a2044d56d1b7e227e", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://arxiv.org/html/2610.26828v1"}], "identifiers": {"arxiv": "2610.26828"}, "kind": "paper", "last_verified_at": "2026-10-20T08:27:26.236797+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://arxiv.org/abs/2610.26828"}], "name": "Agent World Model in the Wild: Computer-Use Agent Environment Synthesis", "publication": null, "published_at": "2026-10-19", "record_sha256": "8c42922798a1f2f18d7a58608e8efaf714ca361d82473ba842a9f8c9e878437e", "relations": [], "resource_id": "arxiv:2610.26828", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://arxiv.org/abs/2610.26828v1", "topics": ["data-generation", "sandbox", "tool-integration", "long-horizon", "imitation-learning", "reinforcement-learning"], "updated_at": "2026-10-19"}, "text": "Agent World Model in the Wild: Computer-Use Agent Environment Synthesis\nResource type: paper\nReview status: curated\nScope: core\n\nWildAWM synthesizes resettable GUI/MCP/CLI environments and cross-interface tasks, using execution-based checks of solvability, difficulty and verifier consistency before supervised and reinforcement learning.\nPlatforms: web, desktop, linux\nTopics: data-generation, sandbox, tool-integration, long-horizon, imitation-learning, reinforcement-learning\nPublished: 2026-10-19\nAuthors: Zhaoyang Wang, Qianhui Wu, Haorui Wang, Simon Yu, Baolin Peng, Canwen Xu, Hao Cheng, Wenlin Yao, Boyi Liu, Yite Wang, Jianfeng Gao, Huaxiu Yao\nRecorded arXiv version: v1\n\nReviewed the complete abstract and selected v1 synthesis, training and evaluation sections. The system uses hybrid GUI/MCP/CLI interfaces. Evaluation includes benchmark subsets and model-based judgments, so reported scores are not treated as directly comparable GUI-only results. No independent reproduction.\n\nPrimary source: https://arxiv.org/abs/2610.26828v1\nCatalog record: https://awesomecua.github.io/resources/arxiv-2610-26828/"}
{"id": "component:google-research/android_control", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research/google-research/tree/master/android_control", "citation_url": "https://awesomecua.github.io/resources/component-google-research-android-control/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "population": "AndroidControl collection", "qualifier": "more-than", "unit": "demonstrations", "value": 15000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "population": "AndroidControl collection", "qualifier": "exact", "unit": "applications", "value": 833}], "evidence_urls": ["https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md"], "formats": ["GZIP-compressed TFRecord"], "modalities": ["screenshots", "accessibility trees", "instructions", "JSON actions"], "notes": "Step instructions can be repeated across multiple actions. Each episode has one more screenshot than action; do not count repeated step instructions as independent task goals.", "splits": ["publisher split file"], "supervision": ["interaction trajectories", "high- and low-level action instruction"]}, "modalities": ["screenshot", "accessibility-tree", "instruction", "action"], "reported_scale": {"comparison": ">", "unit": "demonstrations", "value": 15000}, "sample_count": null, "sample_unit": "demonstrations"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "e4f8baa15441fb63cad07220cc5a4de177a1f785d4fe1fed5cfcbf3c3867faf1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.reported_scale", "details.modalities", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/google-research/tree/master/android_control"}], "name": "AndroidControl", "publication": null, "published_at": null, "record_sha256": "9cfdd5459117d09e14d4894257453b24f85b76735b4e1638aa60586018c562ac", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md", "target_id": "arxiv:2406.03679", "type": "introduced_by"}], "resource_id": "component:google-research/android_control", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research/google-research/tree/master/android_control", "topics": ["data-generation", "mobile-navigation"], "updated_at": null}, "text": "AndroidControl\nResource type: dataset\nReview status: curated\nScope: core\n\nAndroid demonstrations with high-level goals, step instructions, screenshots, and accessibility trees.\nPlatforms: android\nTopics: data-generation, mobile-navigation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md\", \"population\": \"AndroidControl collection\", \"qualifier\": \"more-than\", \"unit\": \"demonstrations\", \"value\": 15000}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md\", \"population\": \"AndroidControl collection\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 833}], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research/google-research/master/android_control/README.md\"], \"formats\": [\"GZIP-compressed TFRecord\"], \"modalities\": [\"screenshots\", \"accessibility trees\", \"instructions\", \"JSON actions\"], \"notes\": \"Step instructions can be repeated across multiple actions. Each episode has one more screenshot than action; do not count repeated step instructions as independent task goals.\", \"splits\": [\"publisher split file\"], \"supervision\": [\"interaction trajectories\", \"high- and low-level action instruction\"]}, \"modalities\": [\"screenshot\", \"accessibility-tree\", \"instruction\", \"action\"], \"reported_scale\": {\"comparison\": \">\", \"unit\": \"demonstrations\", \"value\": 15000}, \"sample_unit\": \"demonstrations\"}\n\nPrimary source: https://github.com/google-research/google-research/tree/master/android_control\nCatalog record: https://awesomecua.github.io/resources/component-google-research-android-control/"}
{"id": "component:google-research/android_in_the_wild", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research/google-research/tree/master/android_in_the_wild", "citation_url": "https://awesomecua.github.io/resources/component-google-research-android-in-the-wild/", "details": {"access": "public download referenced by publisher", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "population": "sum of the five published AitW subsets", "qualifier": "exact", "unit": "episodes", "value": 715142}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "population": "sum of the five published AitW subsets", "qualifier": "exact", "unit": "action examples", "value": 5689993}], "evidence_urls": ["https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md"], "formats": ["GZIP-compressed TFRecord"], "modalities": ["screenshots", "instructions", "actions"], "notes": "Episodes and action examples are different units. The collection combines multi-step subsets and hindsight-labeled single-step tasks; preserve subset and split choices.", "splits": [], "supervision": ["interaction trajectories", "action prediction"]}, "download_url": "https://console.cloud.google.com/storage/browser/gresearch/android-in-the-wild", "example_count": 5689993, "modalities": ["screenshot", "action", "instruction"], "sample_count": 715142, "sample_unit": "episodes"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "6d8964c95db41dada1d5867e50a5ec61f6cb90595965d0e72e010b8e4a12dcb2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.example_count", "details.modalities", "details.access", "details.download_url", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/google-research/google-research/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/google-research/tree/master/android_in_the_wild"}], "name": "Android in the Wild", "publication": null, "published_at": null, "record_sha256": "af22c311bacb2f59f169aa6e5a0391767ca08a96e73cbc7c40d3c87588f3822f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md", "target_id": "arxiv:2307.10088", "type": "introduced_by"}], "resource_id": "component:google-research/android_in_the_wild", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research/google-research/tree/master/android_in_the_wild", "topics": ["data-generation", "mobile-navigation"], "updated_at": null}, "text": "Android in the Wild\nResource type: dataset\nReview status: curated\nScope: core\n\nHuman demonstrations of Android tasks with screenshots and touch actions.\nPlatforms: android\nTopics: data-generation, mobile-navigation\n\nRecorded specifications: {\"access\": \"public download referenced by publisher\", \"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md\", \"population\": \"sum of the five published AitW subsets\", \"qualifier\": \"exact\", \"unit\": \"episodes\", \"value\": 715142}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md\", \"population\": \"sum of the five published AitW subsets\", \"qualifier\": \"exact\", \"unit\": \"action examples\", \"value\": 5689993}], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research/google-research/master/android_in_the_wild/README.md\"], \"formats\": [\"GZIP-compressed TFRecord\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\"], \"notes\": \"Episodes and action examples are different units. The collection combines multi-step subsets and hindsight-labeled single-step tasks; preserve subset and split choices.\", \"splits\": [], \"supervision\": [\"interaction trajectories\", \"action prediction\"]}, \"download_url\": \"https://console.cloud.google.com/storage/browser/gresearch/android-in-the-wild\", \"example_count\": 5689993, \"modalities\": [\"screenshot\", \"action\", \"instruction\"], \"sample_count\": 715142, \"sample_unit\": \"episodes\"}\n\nPrimary source: https://github.com/google-research/google-research/tree/master/android_in_the_wild\nCatalog record: https://awesomecua.github.io/resources/component-google-research-android-in-the-wild/"}
{"id": "component:tencent/selfevolvingagent/webevolver", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver", "citation_url": "https://awesomecua.github.io/resources/component-tencent-selfevolvingagent-webevolver/", "details": {"component_path": "WebEvolver", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "a82450e1b5e15070bee79ee8b79f8d9a97cb85f7", "repository_url": "https://github.com/Tencent/SelfEvolvingAgent"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "447dfa1e52789f7f3f111e9e0d453fb1f9fd8bf2d7b858f28fbe21a2239b4b2e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "4214af42bcadf1ea52a46c2008f8e741f0aa955a58f346d9880858589be7de1d", "supports": ["identity"], "url": "https://api.github.com/repos/Tencent/SelfEvolvingAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:25Z", "role": "primary", "sha256": "ad0f4ae4b58b5a0e84feeb5df717929b97ea37cb945c17c98510ba54509c99e7", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/Tencent/SelfEvolvingAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "873a78a46309fa3892d94b72cf73fd55542bae0e0ba8235afb26a848a64a8255", "supports": ["identity"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "1d7a5369af7ec336bc173dba6cfee9ad0898e527f96c5b7a10cee930806c67a3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver"}], "name": "WebEvolver code component", "publication": null, "published_at": null, "record_sha256": "02d402b223d91d566402eee5307347e223fdcf6830d736358be7aa543b30ca35", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md", "target_id": "arxiv:2504.21024", "type": "documented_by"}], "resource_id": "component:tencent/selfevolvingagent/webevolver", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver", "topics": ["web-navigation", "world-models", "data-generation", "self-reflection"], "updated_at": null}, "text": "WebEvolver code component\nResource type: framework\nReview status: curated\nScope: core\n\nComponent for web-agent trajectory collection, supervised-data preparation and world-model-based trajectory synthesis.\nPlatforms: web\nTopics: web-navigation, world-models, data-generation, self-reflection\n\nRecorded specifications: {\"component_path\": \"WebEvolver\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/License.txt\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"a82450e1b5e15070bee79ee8b79f8d9a97cb85f7\", \"repository_url\": \"https://github.com/Tencent/SelfEvolvingAgent\"}\n\nScoped to the WebEvolver directory; the repository also hosts WebCoT and VScan, which are not merged into this resource. Training and model services require separate setup. Root License.txt explicitly declares MIT for SelfEvolvingAgent and separately identifies CogKernel’s Apache-2.0 dependency terms.\n\nPrimary source: https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver\nCatalog record: https://awesomecua.github.io/resources/component-tencent-selfevolvingagent-webevolver/"}
{"id": "component:x-plug/mobileagent/mobile-agent-e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E", "citation_url": "https://awesomecua.github.io/resources/component-x-plug-mobileagent-mobile-agent-e/", "details": {"component_path": "Mobile-Agent-E", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "11cea575561fb7800b5fb6b6cafa56f7a91de11f", "repository_url": "https://github.com/X-PLUG/MobileAgent"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "f8515fbce73ea0ef3d1d558d4567bb949244e6384e61c956ba3ddbc27294f0ae", "supports": ["name", "kind", "environments", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:03Z", "role": "primary", "sha256": "d403251020124bebcd9f180f800c8ea9f18a3f117c82fcf04c3cbe9ffe0e35f9", "supports": ["identity"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:21Z", "role": "primary", "sha256": "7293efadde7c96b64ff8483b2157241e17d0b0fbfff4839de9c7b11485b07cbb", "supports": ["identity", "release_status", "details.repository_tree_sha", "relations"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["summary", "topics"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E"}], "name": "Mobile-Agent-E code component", "publication": null, "published_at": null, "record_sha256": "af97921114d5a88efd4118df10b6296315fb56b9066a93bf009d3ab47d97508e", "relations": [{"evidence_url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1", "target_id": "github:x-plug/mobileagent", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "target_id": "arxiv:2501.11733", "type": "documented_by"}], "resource_id": "component:x-plug/mobileagent/mobile-agent-e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E", "topics": ["planning", "memory", "self-reflection", "mobile-navigation"], "updated_at": null}, "text": "Mobile-Agent-E code component\nResource type: agent\nReview status: curated\nScope: core\n\nAndroid automation component with hierarchical planning, reflection and persistent tips/shortcuts learned across task sequences.\nPlatforms: android\nTopics: planning, memory, self-reflection, mobile-navigation\n\nRecorded specifications: {\"component_path\": \"Mobile-Agent-E\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"11cea575561fb7800b5fb6b6cafa56f7a91de11f\", \"repository_url\": \"https://github.com/X-PLUG/MobileAgent\"}\n\nThe inspected README explicitly supports Android only, with experiments on Samsung Galaxy A15; host-machine macOS/Linux instructions do not imply control of those operating systems. Agent code, launch scripts and benchmark definitions are present. The root MIT code license does not establish terms for the separately published task dataset or external models.\n\nPrimary source: https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E\nCatalog record: https://awesomecua.github.io/resources/component-x-plug-mobileagent-mobile-agent-e/"}
{"id": "component:x-plug/mobileagent/pc-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent", "citation_url": "https://awesomecua.github.io/resources/component-x-plug-mobileagent-pc-agent/", "details": {"component_path": "PC-Agent", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "11cea575561fb7800b5fb6b6cafa56f7a91de11f", "repository_url": "https://github.com/X-PLUG/MobileAgent"}, "environments": ["desktop", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "ca5b488f10c229b79cc6f9e1f32007d7d84c34366492a10b90d0d66622e77273", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/PC-Agent/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:03Z", "role": "primary", "sha256": "d403251020124bebcd9f180f800c8ea9f18a3f117c82fcf04c3cbe9ffe0e35f9", "supports": ["identity"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:21Z", "role": "primary", "sha256": "7293efadde7c96b64ff8483b2157241e17d0b0fbfff4839de9c7b11485b07cbb", "supports": ["identity", "release_status", "details.repository_tree_sha", "relations"], "url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent"}], "name": "PC-Agent (X-PLUG component)", "publication": null, "published_at": null, "record_sha256": "b72b222b303e9735b74b6cf597639fae8546eaa761bbd63620d7714e1b5c8459", "relations": [{"evidence_url": "https://api.github.com/repos/X-PLUG/MobileAgent/git/trees/main?recursive=1", "target_id": "github:x-plug/mobileagent", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/PC-Agent/README.md", "target_id": "arxiv:2502.14282", "type": "documented_by"}], "resource_id": "component:x-plug/mobileagent/pc-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent", "topics": ["planning", "cross-app", "self-reflection"], "updated_at": null}, "text": "PC-Agent (X-PLUG component)\nResource type: agent\nReview status: curated\nScope: core\n\nMulti-agent desktop automation component for Windows and macOS, combining active perception, task decomposition and reflection for workflows spanning applications.\nPlatforms: desktop, windows, macos\nTopics: planning, cross-app, self-reflection\n\nRecorded specifications: {\"component_path\": \"PC-Agent\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"11cea575561fb7800b5fb6b6cafa56f7a91de11f\", \"repository_url\": \"https://github.com/X-PLUG/MobileAgent\"}\n\nScoped to PC-Agent in the MobileAgent monorepo and distinct from GAIR-NLP/PC-Agent. The source tree contains current and v1 implementations; the README states a March 2025 code update. Root MIT covers repository code; the separately linked PC-Eval tasks and application software are not assigned that license by this record.\n\nPrimary source: https://github.com/X-PLUG/MobileAgent/tree/main/PC-Agent\nCatalog record: https://awesomecua.github.io/resources/component-x-plug-mobileagent-pc-agent/"}
{"id": "component:xlang-ai/openagents/web_agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent", "citation_url": "https://awesomecua.github.io/resources/component-xlang-ai-openagents-web-agent/", "details": {"availability": "Public Web Agent source and Chrome-extension archive manifest with local setup instructions.", "component_path": "real_agents/web_agent", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"], "reason": "OpenAgents source license, scoped here to the reviewed Web Agent component; bundled third-party components retain their notices."}}, "repository_revision": "ff2e46440699af1324eb25655b622c4a131265bb"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "fb6ef7fdb133612467e27ffa642ccbe73cfb4bd8fb0c63c1bc629ad686694e1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "b843b0519dc74bf6dbf693e0274e8ef8eb1ade94a656d3ab23faac2df4f7c9fa", "supports": ["identity"], "url": "https://api.github.com/repos/xlang-ai/OpenAgents"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "cf5a65d5a6b7ba2a9cf950421f6902534d523f7e8d850d1b23bd5395fd3d04c8", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/xlang-ai/OpenAgents/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "bcd2232e1ec4ad9ba0628cddd30031c364d62c6f112d8ecc2c24b9a4d35696af", "supports": ["release_status", "notes", "environments"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/frontend/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:45:23Z", "role": "primary", "sha256": "5b7084cfe6149602740cf4cd85e958142aed3aea0a0a0345a82ec1618ceb7b19", "supports": ["summary", "environments", "scope"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/real_agents/web_agent/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "a86eea1027b8d92a68ded4231671379388df757992fa69d3e8fc8f8115d124fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": {"basis": "OpenAgents source license, scoped here to the reviewed Web Agent component; bundled third-party components retain their notices.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:09Z", "url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent"}], "name": "OpenAgents Web Agent", "publication": null, "published_at": null, "record_sha256": "00fa9ec55b55149205a1b467afd9cd281bb98add116104656a46b61ca613a218", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/README.md", "target_id": "arxiv:2310.10634", "type": "documented_by"}], "resource_id": "component:xlang-ai/openagents/web_agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent", "topics": ["web-navigation", "automation", "tool-integration"], "updated_at": null}, "text": "OpenAgents Web Agent\nResource type: agent\nReview status: curated\nScope: core\n\nThe OpenAgents browser-control component, connecting a Chrome extension to a web agent and local chat platform for navigating sites and completing interface tasks.\nPlatforms: web\nTopics: web-navigation, automation, tool-integration\n\nRecorded specifications: {\"availability\": \"Public Web Agent source and Chrome-extension archive manifest with local setup instructions.\", \"component_path\": \"real_agents/web_agent\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/xlang-ai/OpenAgents/HEAD/LICENSE\"], \"reason\": \"OpenAgents source license, scoped here to the reviewed Web Agent component; bundled third-party components retain their notices.\"}}, \"repository_revision\": \"ff2e46440699af1324eb25655b622c4a131265bb\"}\n\nThis component record covers real_agents/web_agent and the documented WeBot Chrome-extension integration. The broader Data Agent and Plugins Agent are not treated as GUI-control assets. The source manifest includes the agent implementation and frontend/webot_extension.zip; the archive was not downloaded or executed. The README describes Docker deployment as under development and recommends source deployment. Apache-2.0 applies to project source; third-party extension/dependency notices remain separate.\n\nPrimary source: https://github.com/xlang-ai/OpenAgents/tree/main/real_agents/web_agent\nCatalog record: https://awesomecua.github.io/resources/component-xlang-ai-openagents-web-agent/"}
{"id": "github:aaronlab/browsertrace", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/aaronlab/browsertrace", "citation_url": "https://awesomecua.github.io/resources/github-aaronlab-browsertrace/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:05Z", "role": "primary", "sha256": "86d7fbe30596de4a1f1a1df2a986f5f247351a91c0c593dc4c29ba9c70d28c8f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "7b4ea96686420061c9d0e1ad4984d7c5efb7e1a62ce6dffd47862f02a7260867", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/aaronlab/browsertrace/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/aaronlab/browsertrace"}], "name": "BrowserTrace", "publication": null, "published_at": null, "record_sha256": "31ef3efbb93f552a02e1bde7016fd1784bc99af944fdb6d9979fea528767a94d", "relations": [], "resource_id": "github:aaronlab/browsertrace", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/aaronlab/browsertrace", "topics": ["observability"], "updated_at": null}, "text": "BrowserTrace\nResource type: tool\nReview status: curated\nScope: core\n\nLocal, cloud-free replay debugger that records each Browser Use step's screenshot, URL, action and model I/O as a timeline and reports the step at which a failed run diverges from a known-good one.\nPlatforms: web\nTopics: observability\n\nPrimary source: https://github.com/aaronlab/browsertrace\nCatalog record: https://awesomecua.github.io/resources/github-aaronlab-browsertrace/"}
{"id": "github:aburns4/motif", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/aburns4/MoTIF", "citation_url": "https://awesomecua.github.io/resources/github-aburns4-motif/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md"], "formats": ["JSON metadata", "screen images"], "modalities": ["screenshots", "view hierarchies", "language", "interaction sequences", "feasibility labels"], "notes": "The raw collection mixes action, language, exploration, and unusable/other traces. These groups have different annotation coverage; not every raw trace is a valid task demonstration.", "splits": [], "supervision": ["interaction trajectories", "task feasibility"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4b2f5a23dc28e96b671272380b953e7fabaf34f4b1e83f3513643846718e3389", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/aburns4/MoTIF"}], "name": "MoTIF", "publication": null, "published_at": null, "record_sha256": "21b42d7d215815c47ecff7c40e08dd25de86c2f6acf32d050ddd01f3f6a04605", "relations": [{"evidence_url": "https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md", "target_id": "arxiv:2202.02312", "type": "documented_by"}], "resource_id": "github:aburns4/motif", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/aburns4/MoTIF", "topics": ["mobile-navigation"], "updated_at": null}, "text": "MoTIF\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile task demonstrations and natural-language instructions.\nPlatforms: android\nTopics: mobile-navigation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://raw.githubusercontent.com/aburns4/MoTIF/HEAD/README.md\"], \"formats\": [\"JSON metadata\", \"screen images\"], \"modalities\": [\"screenshots\", \"view hierarchies\", \"language\", \"interaction sequences\", \"feasibility labels\"], \"notes\": \"The raw collection mixes action, language, exploration, and unusable/other traces. These groups have different annotation coverage; not every raw trace is a valid task demonstration.\", \"splits\": [], \"supervision\": [\"interaction trajectories\", \"task feasibility\"]}}\n\nPrimary source: https://github.com/aburns4/MoTIF\nCatalog record: https://awesomecua.github.io/resources/github-aburns4-motif/"}
{"id": "github:acrab-agentic-labs/appsim", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Acrab-Agentic-Labs/AppSim", "citation_url": "https://awesomecua.github.io/resources/github-acrab-agentic-labs-appsim/", "details": {"benchmark_protocol": {"actions": ["mobile GUI operations"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "AppSim-Bench paper-described release", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2609.07712", "unit": "tasks", "value": 557}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "population": "published simulated app set", "qualifier": "exact", "unit": "applications", "value": 17}], "evaluation": "Controlled backend data and outcome-based task verification.", "evidence_urls": ["https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["mobile task completion"], "notes": "Use the simulated apps and controlled backend state supplied by the benchmark. Desktop host instructions are not the platform being evaluated.", "observations": ["mobile interfaces"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "12dc2afe339c1ed66e4a6a4ccf5ae13d27ce32f452960341529a4b9016f69059", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:32Z", "role": "primary", "sha256": "13d3fb5811d203f2e6b80eb45d30ebc0b3cfefd88e2034ffa3b1b80687bef285", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Acrab-Agentic-Labs/AppSim"}], "name": "AppSim-Bench", "publication": null, "published_at": null, "record_sha256": "5092c7a3932b577194aa19f275740d315b00c20fd0abc8504f49405c20b2383a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md", "target_id": "arxiv:2609.07712", "type": "documented_by"}], "resource_id": "github:acrab-agentic-labs/appsim", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Acrab-Agentic-Labs/AppSim", "topics": ["evaluation", "sandbox"], "updated_at": null}, "text": "AppSim-Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\nSimulated mobile applications and evaluation tooling for reproducible GUI-agent tasks with controlled application state.\nPlatforms: android\nTopics: evaluation, sandbox\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"mobile GUI operations\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending\", \"population\": \"AppSim-Bench paper-described release\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2609.07712\", \"unit\": \"tasks\", \"value\": 557}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md\", \"population\": \"published simulated app set\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 17}], \"evaluation\": \"Controlled backend data and outcome-based task verification.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/Acrab-Agentic-Labs/AppSim/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=100&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"focus\": [\"mobile task completion\"], \"notes\": \"Use the simulated apps and controlled backend state supplied by the benchmark. Desktop host instructions are not the platform being evaluated.\", \"observations\": [\"mobile interfaces\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/Acrab-Agentic-Labs/AppSim\nCatalog record: https://awesomecua.github.io/resources/github-acrab-agentic-labs-appsim/"}
{"id": "github:agent-infra/sandbox", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/agent-infra/sandbox", "citation_url": "https://awesomecua.github.io/resources/github-agent-infra-sandbox/", "details": {}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "0fe34cd605e492c4c93f760e9caefd194f9c8ec7cfc3d4d4a927a1ac1613e51c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/agent-infra/sandbox/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agent-infra/sandbox"}], "name": "Agent Infra Sandbox", "publication": null, "published_at": null, "record_sha256": "84b6d05e32999efcc1ceeba35a21a9a6f6891ea596756d48268c29ad1768f37a", "relations": [], "resource_id": "github:agent-infra/sandbox", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/agent-infra/sandbox", "topics": ["sandbox"], "updated_at": null}, "text": "Agent Infra Sandbox\nResource type: infrastructure\nReview status: curated\nScope: core\n\nSandbox environments supporting computer and browser interaction.\nPlatforms: desktop, web\nTopics: sandbox\n\nPrimary source: https://github.com/agent-infra/sandbox\nCatalog record: https://awesomecua.github.io/resources/github-agent-infra-sandbox/"}
{"id": "github:agentsea/agentdesk", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/agentsea/agentdesk", "citation_url": "https://awesomecua.github.io/resources/github-agentsea-agentdesk/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "3ee6212ef10c39943546d20a3beeed412262f4a8a910120becb5409626569b0f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "40f2354eb662ebb7b99d874b45f639e1265bec931073cda9ff21c6cf2cb0c499", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/agentsea/agentdesk/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agentsea/agentdesk"}], "name": "AgentDesk", "publication": null, "published_at": null, "record_sha256": "e98709631b19e39df38b711b97b5c981fe9b2fe7d35c7fafed2a2623e20636a9", "relations": [], "resource_id": "github:agentsea/agentdesk", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/agentsea/agentdesk", "topics": ["sandbox"], "updated_at": null}, "text": "AgentDesk\nResource type: infrastructure\nReview status: curated\nScope: core\n\nDesktop environments that AI agents control programmatically, run locally with Docker or remotely on Kubernetes, built on the agentd REST daemon and offering a coordinate-grid overlay for LMM input.\nPlatforms: desktop\nTopics: sandbox\n\nPrimary source: https://github.com/agentsea/agentdesk\nCatalog record: https://awesomecua.github.io/resources/github-agentsea-agentdesk/"}
{"id": "github:agentsea/surfkit", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/agentsea/surfkit", "citation_url": "https://awesomecua.github.io/resources/github-agentsea-surfkit/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "32420db547a889f0d86090689d0bd0c93d1776a70346dfd4029db7d99ae4bd9f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "40f2354eb662ebb7b99d874b45f639e1265bec931073cda9ff21c6cf2cb0c499", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/agentsea/surfkit/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/agentsea/surfkit"}], "name": "Surfkit", "publication": null, "published_at": null, "record_sha256": "682a98135fdd9a08520e4a17a98c677ea35562abe1a2b467735e7eff2fa54535", "relations": [], "resource_id": "github:agentsea/surfkit", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/agentsea/surfkit", "topics": ["workflow"], "updated_at": null}, "text": "Surfkit\nResource type: framework\nReview status: curated\nScope: core\n\nPython toolkit and CLI for building, publishing and running multimodal agents that operate devices such as Ubuntu desktops, locally or on Kubernetes, with task trackers and an agent hub.\nPlatforms: desktop\nTopics: workflow\n\nPrimary source: https://github.com/agentsea/surfkit\nCatalog record: https://awesomecua.github.io/resources/github-agentsea-surfkit/"}
{"id": "github:ai-agents-2030/spa-bench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ai-agents-2030/SPA-Bench", "citation_url": "https://awesomecua.github.io/resources/github-ai-agents-2030-spa-bench/", "details": {"benchmark_protocol": {"actions": ["agent-specific Android actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "complete README task collection", "qualifier": "exact", "unit": "tasks", "value": 340}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "single-app subset", "qualifier": "exact", "unit": "tasks", "value": 300}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "cross-app subset", "qualifier": "exact", "unit": "tasks", "value": 40}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "population": "README total including third-party, Google, and system apps", "qualifier": "exact", "unit": "applications", "value": 66}], "evaluation": "A multimodal evaluator and annotated key-component matching assess completion; the framework also reports resource-use metrics.", "evidence_urls": ["https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/Documentation.md"], "focus": ["mobile task completion", "resource use", "cross-app workflows"], "notes": "The two task subsets partition the 340 tasks. The integrated agent and evaluator modes must accompany results. The preconfigured emulator snapshot is still marked coming soon in the inspected documentation.", "observations": ["screenshot trajectories"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "5294a67728072a3658cb2073bdf9a084fe3c00d87a848923dff68adb35213072", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "5294a67728072a3658cb2073bdf9a084fe3c00d87a848923dff68adb35213072", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "a55290d62f8823097c73d74bac4eb08853a3caff724382e60a6a9364e7ced987", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/Documentation.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "6962bfc8ee16c6e985e4fb680b8fe46d5b34d30f7cb8a1e61fd4c55c93120ddf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ai-agents-2030/SPA-Bench"}], "name": "SPA-Bench", "publication": null, "published_at": null, "record_sha256": "4217c62c20199971d28b0d75d65c0cca3d323cdc5e9e0125b73c4d269ce61160", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md", "target_id": "arxiv:2410.15164", "type": "documented_by"}], "resource_id": "github:ai-agents-2030/spa-bench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ai-agents-2030/SPA-Bench", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "SPA-Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\n340 single-app and cross-app tasks over 66 Chinese and English Android apps, with plug-in support for 11 smartphone agents and automatic coarse-to-fine success detection.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"agent-specific Android actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md\", \"population\": \"complete README task collection\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 340}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md\", \"population\": \"single-app subset\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 300}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md\", \"population\": \"cross-app subset\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 40}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md\", \"population\": \"README total including third-party, Google, and system apps\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 66}], \"evaluation\": \"A multimodal evaluator and annotated key-component matching assess completion; the framework also reports resource-use metrics.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/README.md\", \"https://raw.githubusercontent.com/ai-agents-2030/SPA-Bench/HEAD/Documentation.md\"], \"focus\": [\"mobile task completion\", \"resource use\", \"cross-app workflows\"], \"notes\": \"The two task subsets partition the 340 tasks. The integrated agent and evaluator modes must accompany results. The preconfigured emulator snapshot is still marked coming soon in the inspected documentation.\", \"observations\": [\"screenshot trajectories\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/ai-agents-2030/SPA-Bench\nCatalog record: https://awesomecua.github.io/resources/github-ai-agents-2030-spa-bench/"}
{"id": "github:ai-nikolai/stateact", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ai-nikolai/StateAct", "citation_url": "https://awesomecua.github.io/resources/github-ai-nikolai-stateact/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:07Z", "role": "primary", "sha256": "b68b2cc26bfc6a654fec35f3c53c1b1334f3d617671673407d3ed3b59f3db973", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ai-nikolai/StateAct/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "38f8c08395c6e67ff9bf52b27e772667f193a172bd6f726507c57776ad3e5e5f", "supports": ["identity"], "url": "https://api.github.com/repos/ai-nikolai/StateAct"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:07Z", "role": "primary", "sha256": "97841e6eab1002bcea949e031f5abd3567482ebbcf03b991bd64643027c468bf", "supports": ["release_status"], "url": "https://api.github.com/repos/ai-nikolai/StateAct/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ai-nikolai/StateAct"}], "name": "StateAct", "publication": null, "published_at": null, "record_sha256": "4bb6faca0a2619de960a98d810e1a040636b5bc043fe627e5a6887c1be3a3406", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ai-nikolai/StateAct/HEAD/README.md", "target_id": "arxiv:2410.02810", "type": "documented_by"}], "resource_id": "github:ai-nikolai/stateact", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/ai-nikolai/StateAct", "topics": ["planning", "state-tracking", "web-navigation"], "updated_at": null}, "text": "StateAct\nResource type: framework\nReview status: curated\nScope: adjacent\n\nState-tracking prompts and experiment runners for language agents, including a WebShop implementation alongside text-based environments.\nPlatforms: web\nTopics: planning, state-tracking, web-navigation\n\nThe repository contains a WebShop runner as well as ALFWorld and TextCraft experiments. Inclusion concerns its web-shopping interaction component. No code license or downloadable trained checkpoint is declared in the inspected root metadata.\n\nPrimary source: https://github.com/ai-nikolai/StateAct\nCatalog record: https://awesomecua.github.io/resources/github-ai-nikolai-stateact/"}
{"id": "github:ai-secure/advagent", "metadata": {"aliases": ["AdvWeb"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/AI-secure/AdvAgent", "citation_url": "https://awesomecua.github.io/resources/github-ai-secure-advagent/", "details": {"availability": "Public research source and workflow documentation; no standalone trained checkpoint was verified.", "repository_revision": "e01316e33d11ab145daa7cc7ec12d6ddb0bf38a0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:18Z", "role": "primary", "sha256": "019d45dd4c383f919f3ac2bf84c672500d63751a49de2dd2bac9c018ffc99881", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/AI-secure/AdvAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:17Z", "role": "primary", "sha256": "3ea44a202a677c0beb744ab05bd687da93a83365b7eb0b7999c3a4c22d7e2608", "supports": ["identity"], "url": "https://api.github.com/repos/AI-secure/AdvAgent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:18:33Z", "role": "primary", "sha256": "95929797eb0717ce26b26ccbdc752bfda18fc0070975b32f09827b4e9fec6ecf", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/AI-secure/AdvAgent/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:48Z", "role": "primary", "sha256": "3ea44a202a677c0beb744ab05bd687da93a83365b7eb0b7999c3a4c22d7e2608", "supports": ["identity"], "url": "https://api.github.com/repos/AI-secure/AdvWeb"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:18:18Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AI-secure/AdvAgent"}], "name": "AdvAgent (formerly AdvWeb)", "publication": null, "published_at": null, "record_sha256": "82a85ecd4af8ad5b6c598557f85fb82720a73a79ae46b544d30548140ac3b958", "relations": [], "resource_id": "github:ai-secure/advagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/AI-secure/AdvAgent", "topics": ["safety-security", "robustness", "evaluation"], "updated_at": null}, "text": "AdvAgent (formerly AdvWeb)\nResource type: framework\nReview status: curated\nScope: core\n\nResearch code for constructing and evaluating adversarial inputs to web agents, with a black-box red-team pipeline using Mind2Web data and SeeAct.\nRecorded aliases: AdvWeb\nPlatforms: web\nTopics: safety-security, robustness, evaluation\n\nRecorded specifications: {\"availability\": \"Public research source and workflow documentation; no standalone trained checkpoint was verified.\", \"repository_revision\": \"e01316e33d11ab145daa7cc7ec12d6ddb0bf38a0\"}\n\nThe official GitHub API resolves AI-secure/AdvWeb to AI-secure/AdvAgent. Public data-generation, SFT/DPO training and evaluation scripts are present, with upstream data and external model APIs required. No repository license was found, and no permission for third-party data or model weights is inferred. The README citation block is empty, so no paper identifier is bound from title similarity alone.\n\nPrimary source: https://github.com/AI-secure/AdvAgent\nCatalog record: https://awesomecua.github.io/resources/github-ai-secure-advagent/"}
{"id": "github:akimotoayako/visiontasker", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/AkimotoAyako/VisionTasker", "citation_url": "https://awesomecua.github.io/resources/github-akimotoayako-visiontasker/", "details": {"availability": "Public Android automation source and setup instructions; external models/services required.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"], "reason": "Repository GNU GPL version 3 license; pretrained models and linked data retain separate terms."}}, "repository_revision": "d31b50e98bde28c0d1cfbc7d7ee2528d10b3fbb0"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:18Z", "role": "primary", "sha256": "816981b343628c53188fd27041d62f52854fcfe83c976a125b84545f5e010274", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:23Z", "role": "primary", "sha256": "94d8e95d302160e3c581015267dcfbb9f61c41ad1f04e470fcd71f5ee1ffa25d", "supports": ["identity"], "url": "https://api.github.com/repos/AkimotoAyako/VisionTasker"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:48Z", "role": "primary", "sha256": "54a1c08d44e0ad1d60bee57e3db2842518e014db3be8b08362e484315f3a6a6f", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/AkimotoAyako/VisionTasker/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T11:20:18Z", "licenses": {"code": {"basis": "Repository GNU GPL version 3 license; pretrained models and linked data retain separate terms.", "declared": "GPL-3.0", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AkimotoAyako/VisionTasker"}], "name": "VisionTasker", "publication": null, "published_at": null, "record_sha256": "bfb93399bfa58d51991e4604da7c44a8efa7404f5b468140b5b56730134b8523", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/readme.md", "target_id": "arxiv:2312.11190", "type": "documented_by"}], "resource_id": "github:akimotoayako/visiontasker", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/AkimotoAyako/VisionTasker", "topics": ["mobile-navigation", "screen-understanding", "planning"], "updated_at": null}, "text": "VisionTasker\nResource type: agent\nReview status: curated\nScope: core\n\nAn Android automation implementation that translates screenshot-based UI understanding into language descriptions for step-by-step LLM planning and action execution.\nPlatforms: android\nTopics: mobile-navigation, screen-understanding, planning\n\nRecorded specifications: {\"availability\": \"Public Android automation source and setup instructions; external models/services required.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/AkimotoAyako/VisionTasker/HEAD/LICENSE\"], \"reason\": \"Repository GNU GPL version 3 license; pretrained models and linked data retain separate terms.\"}}, \"repository_revision\": \"d31b50e98bde28c0d1cfbc7d7ee2528d10b3fbb0\"}\n\nThe official lowercase readme.md and source tree expose automatic and step-by-step entry points. The documented setup uses a Windows host and an Android phone connected through ADB; it also requires separately obtained detection/CLIP models and model/OCR services. GPL-3.0 is recorded for the code only. External model download links were not used to infer weight licenses or verify a separate checkpoint release.\n\nPrimary source: https://github.com/AkimotoAyako/VisionTasker\nCatalog record: https://awesomecua.github.io/resources/github-akimotoayako-visiontasker/"}
{"id": "github:alibaba/ui-ins", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/alibaba/UI-Ins", "citation_url": "https://awesomecua.github.io/resources/github-alibaba-ui-ins/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "e06aedcb1760867470afeb27d4eb4cdde9a3b76b9d55af183047df3c0ae2b71c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/alibaba/UI-Ins/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/alibaba/UI-Ins"}], "name": "UI-Ins", "publication": null, "published_at": null, "record_sha256": "3a85ca23aa97eab8d4d73055d8175bc8f20692a373ff29a911e779a6326c2d7c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/alibaba/UI-Ins/HEAD/README.md", "target_id": "arxiv:2510.20286", "type": "documented_by"}], "resource_id": "github:alibaba/ui-ins", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/alibaba/UI-Ins", "topics": ["data-generation"], "updated_at": null}, "text": "UI-Ins\nResource type: dataset\nReview status: curated\nScope: core\n\nTreats grounding instructions as multiple reasoning perspectives, releasing a data pipeline, training code and UI-Ins-7B and 32B models trained by SFT on synthesized instructions, then RL.\nPlatforms: cross-platform\nTopics: data-generation\n\nPrimary source: https://github.com/alibaba/UI-Ins\nCatalog record: https://awesomecua.github.io/resources/github-alibaba-ui-ins/"}
{"id": "github:alibabaresearch/rots", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/AlibabaResearch/RoTS", "citation_url": "https://awesomecua.github.io/resources/github-alibabaresearch-rots/", "details": {"availability": "announced; data and sampling pipelines in preparation", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md", "population": "paper-described synthesis output; not a released file inventory", "qualifier": "approximately", "unit": "training samples", "value": 800000}], "evidence_urls": ["https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"], "formats": [], "modalities": ["GUI interaction trajectories"], "notes": "The inspected README marks the dataset and collection pipelines as in preparation. This profile records an announced research asset, not a verified downloadable corpus; file formats, splits, and release-specific data terms remain unknown.", "splits": [], "supervision": ["recovery trajectory synthesis"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "bf5c1469d33ced951ff0a5de017494b65f654023ceee8e7fbfd8b60afb082244", "supports": ["summary", "notes", "details.availability"], "url": "https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AlibabaResearch/RoTS"}], "name": "RoTS", "publication": null, "published_at": null, "record_sha256": "949b37970734d213881da8d166688e1829ec474ec8618bef230d0ccbbdd710b2", "relations": [], "resource_id": "github:alibabaresearch/rots", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/AlibabaResearch/RoTS", "topics": ["data-generation", "robustness"], "updated_at": null}, "text": "RoTS\nResource type: dataset\nReview status: curated\nScope: core\n\nAnnounced GUI trajectory synthesis resources for recovery from policy-induced errors; the inspected repository marks code and data releases as in preparation.\nPlatforms: cross-platform\nTopics: data-generation, robustness\n\nRecorded specifications: {\"availability\": \"announced; data and sampling pipelines in preparation\", \"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md\", \"population\": \"paper-described synthesis output; not a released file inventory\", \"qualifier\": \"approximately\", \"unit\": \"training samples\", \"value\": 800000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/AlibabaResearch/RoTS/HEAD/README.md\"], \"formats\": [], \"modalities\": [\"GUI interaction trajectories\"], \"notes\": \"The inspected README marks the dataset and collection pipelines as in preparation. This profile records an announced research asset, not a verified downloadable corpus; file formats, splits, and release-specific data terms remain unknown.\", \"splits\": [], \"supervision\": [\"recovery trajectory synthesis\"]}}\n\nThe 800k figure describes training samples in the study, not an independently counted public download. The repository code-license declaration does not establish terms for an unreleased dataset.\n\nPrimary source: https://github.com/AlibabaResearch/RoTS\nCatalog record: https://awesomecua.github.io/resources/github-alibabaresearch-rots/"}
{"id": "github:alipay/mobile-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/alipay/mobile-agent", "citation_url": "https://awesomecua.github.io/resources/github-alipay-mobile-agent/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "c4da2b089f97296ac6c7b3ef7273db57b650d12cf3595f6f6ea74c901e5277b8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/alipay/mobile-agent/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/alipay/mobile-agent"}], "name": "MobileAgent (Alipay)", "publication": null, "published_at": null, "record_sha256": "a15c476f88ed8756a5d43802db464692879a3674e27bddbe50944fb6ad339d38", "relations": [{"evidence_url": "https://raw.githubusercontent.com/alipay/mobile-agent/HEAD/README.md", "target_id": "arxiv:2401.04124", "type": "documented_by"}], "resource_id": "github:alipay/mobile-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/alipay/mobile-agent", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": null}, "text": "MobileAgent (Alipay)\nResource type: agent\nReview status: curated\nScope: core\n\nMobile-control research implementation combining operating-procedure context with human interaction for personalized task execution.\nPlatforms: android\nTopics: mobile-navigation, workflow, human-in-the-loop\n\nDistinct from the X-PLUG/MobileAgent project despite the similar name.\n\nPrimary source: https://github.com/alipay/mobile-agent\nCatalog record: https://awesomecua.github.io/resources/github-alipay-mobile-agent/"}
{"id": "github:alokia/cola-demo", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Alokia/COLA-demo", "citation_url": "https://awesomecua.github.io/resources/github-alokia-cola-demo/", "details": {}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "281793fe6b875bd1a63591502345ac9523b1127acd9f41d5c2c88933a357e4f2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Alokia/COLA-demo/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["relations", "environments"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Alokia/COLA-demo"}], "name": "COLA demo", "publication": null, "published_at": null, "record_sha256": "eee6173dd0a3d0d4a4ce9834d1e22415818ac383b5ba40abf6d9938c2b5504fa", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50", "target_id": "arxiv:2503.09263", "type": "documented_by"}], "resource_id": "github:alokia/cola-demo", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Alokia/COLA-demo", "topics": ["automation", "planning", "human-in-the-loop"], "updated_at": null}, "text": "COLA demo\nResource type: agent\nReview status: curated\nScope: core\n\nResearch demonstration of collaborative Windows GUI automation with dynamic task scheduling and recovery mechanisms.\nPlatforms: desktop, windows\nTopics: automation, planning, human-in-the-loop\n\nThe publisher describes this as demonstration code and states that the full implementation is not released.\n\nPrimary source: https://github.com/Alokia/COLA-demo\nCatalog record: https://awesomecua.github.io/resources/github-alokia-cola-demo/"}
{"id": "github:amazon-science/pae", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/amazon-science/PAE", "citation_url": "https://awesomecua.github.io/resources/github-amazon-science-pae/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "3e5f24b3175afeba8635dacc7b66eba1ce89fea49996a20281ff2df5a41f1e24", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "ae4f966004dd672fd5b994add5a5cba1fee9233d4d81b224e94c339728fafd87", "supports": ["identity"], "url": "https://api.github.com/repos/amazon-science/PAE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "b1e24a3f38fae67f71033a3f920164d282d3f8bb072f438349dbe9eb23730a7a", "supports": ["release_status"], "url": "https://api.github.com/repos/amazon-science/PAE/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "db2e35513dbadcdc67f5819a3bfee2777786538dd3531620cd5fbd4b6ed6e538", "supports": ["licenses"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/LICENSE.txt"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/amazon-science/PAE"}], "name": "Proposer-Agent-Evaluator (PAE)", "publication": null, "published_at": null, "record_sha256": "ccd67137dedb9a81794bfe9308626c1f55c14baca94907d6df7813ff9e491736", "relations": [{"evidence_url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md", "target_id": "arxiv:2412.13194", "type": "documented_by"}], "resource_id": "github:amazon-science/pae", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/amazon-science/PAE", "topics": ["reinforcement-learning", "web-navigation", "data-generation"], "updated_at": null}, "text": "Proposer-Agent-Evaluator (PAE)\nResource type: framework\nReview status: curated\nScope: core\n\nA browser-agent training system that proposes tasks, collects visual interaction trajectories, and uses automated evaluation to drive policy improvement.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, data-generation\n\nSource code and browser-environment modules are present. The repository LICENSE is CC-BY-NC-4.0, while its README describes that license as applying to a dataset, leaving code scope ambiguous. The separately linked dataset is gated and was not accessed; model and dataset terms must be checked separately.\n\nPrimary source: https://github.com/amazon-science/PAE\nCatalog record: https://awesomecua.github.io/resources/github-amazon-science-pae/"}
{"id": "github:androidarenaagent/androidarena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/AndroidArenaAgent/AndroidArena", "citation_url": "https://awesomecua.github.io/resources/github-androidarenaagent-androidarena/", "details": {"benchmark_protocol": {"actions": ["app launch/stop", "click/double-click/long-click", "text input", "swipe/drag and Android navigation", "finish"], "counts": [], "evaluation": "The released evaluator compares trajectories with reference actions using longest-common-subsequence-based task reward and completion. Its reversed redundancy metric is the reference/predicted action-length ratio (with 1e-6 denominator smoothing), multiplied by the LLM success flag. LLM success judgments and instruction-constraint checks are separate metrics. Full reference-dependent evaluation coverage cannot be assumed because complete annotated action sequences are not released.", "evidence_urls": ["https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/README.md", "https://api.github.com/repos/AndroidArenaAgent/AndroidArena/git/trees/main?recursive=1", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/observation.py", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/actions.py", "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/run_evaluator.py"], "focus": ["multi-step Android tasks", "cross-app workflows and instruction constraints"], "notes": "README states that task instructions are released but annotated action sequences, apart from the calendar example, will come later. Environment setup targets Android API 33 with uiautomator2 and manual Google-account setup. Public task YAML and evaluator code exist, but benchmark execution and full reference coverage were not verified. No task total is derived from app YAML file counts. Code and data licenses remain unknown.", "observations": ["UI hierarchy text", "screenshots", "combined text/image observations"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "206bd0dc5220cf754cffc97767cce6a3cd27660ef050a569366bef2331241c60", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "a1e7dc0c5facae030c2cf19aad6511996e7e1198d8d6619268d10c29a1b57bcf", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "edc03388b67143644db4a412e4c888f8c796068daff29f63615f0b6a3a04262a", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/AndroidArenaAgent/AndroidArena/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:52Z", "role": "primary", "sha256": "cfa9d272c85398f54c18d6e7445d680009e8c16d128962235d899592d3094c2f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/observation.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:53Z", "role": "primary", "sha256": "d25f2edf9202f59615341b62fc6046ac23a6f4cb072a6038ac562506cbdc891c", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/actions.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "257c42e980a0b14bd74bfcc381a0ecbef2b538c0f1cc1a94d22925efe354fb8e", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/run_evaluator.py"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AndroidArenaAgent/AndroidArena"}], "name": "AndroidArena", "publication": null, "published_at": null, "record_sha256": "241b7694563603b87e72d0b36338e5d6643de2fa4453a9dc67451b23157763fa", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md", "target_id": "arxiv:2402.06596", "type": "documented_by"}], "resource_id": "github:androidarenaagent/androidarena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/AndroidArenaAgent/AndroidArena", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "AndroidArena\nResource type: benchmark\nReview status: curated\nScope: core\n\nAndroid environment for evaluating language-model-driven agents.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"app launch/stop\", \"click/double-click/long-click\", \"text input\", \"swipe/drag and Android navigation\", \"finish\"], \"counts\": [], \"evaluation\": \"The released evaluator compares trajectories with reference actions using longest-common-subsequence-based task reward and completion. Its reversed redundancy metric is the reference/predicted action-length ratio (with 1e-6 denominator smoothing), multiplied by the LLM success flag. LLM success judgments and instruction-constraint checks are separate metrics. Full reference-dependent evaluation coverage cannot be assumed because complete annotated action sequences are not released.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/README.md\", \"https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/README.md\", \"https://api.github.com/repos/AndroidArenaAgent/AndroidArena/git/trees/main?recursive=1\", \"https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/observation.py\", \"https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/android_env2/actions.py\", \"https://raw.githubusercontent.com/AndroidArenaAgent/AndroidArena/HEAD/run_evaluator.py\"], \"focus\": [\"multi-step Android tasks\", \"cross-app workflows and instruction constraints\"], \"notes\": \"README states that task instructions are released but annotated action sequences, apart from the calendar example, will come later. Environment setup targets Android API 33 with uiautomator2 and manual Google-account setup. Public task YAML and evaluator code exist, but benchmark execution and full reference coverage were not verified. No task total is derived from app YAML file counts. Code and data licenses remain unknown.\", \"observations\": [\"UI hierarchy text\", \"screenshots\", \"combined text/image observations\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/AndroidArenaAgent/AndroidArena\nCatalog record: https://awesomecua.github.io/resources/github-androidarenaagent-androidarena/"}
{"id": "github:anikasomaia/local-activity-recorder", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/anikasomaia/local-activity-recorder", "citation_url": "https://awesomecua.github.io/resources/github-anikasomaia-local-activity-recorder/", "details": {"license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/anikasomaia/local-activity-recorder/git/trees/f7ec5b6ecf0ba1bfe6362d7f2bfb0f1af5ac6bc7?recursive=1"], "reason": "No license declaration was found in the inspected README/C# files, and the complete pinned tree has no LICENSE, LICENCE or COPYING-named file. Public source access does not establish reuse permission."}}, "repository_revision": "967cea087bf6496b3b82743fb96770fa29f8e203"}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:15.716415+00:00", "role": "primary", "sha256": "440e0e43026ba7a20eaca17507cec33bd697d8b1eed6fb056be34136b3f4f76f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:41.685729+00:00", "role": "primary", "sha256": "90c30399c8aa5031f82164978bc12eae3bebba172758d20877a87fec53566c12", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Acomputer-use-agents+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:07.061516+00:00", "role": "primary", "sha256": "f5b501fa132163d749c6695d81c76d45fdff2888ea890e31cbaa9bc4172fed3b", "supports": ["identity", "details.repository_revision", "summary", "environments", "topics", "notes"], "url": "https://api.github.com/repos/anikasomaia/local-activity-recorder/commits/master"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:33.833893+00:00", "role": "primary", "sha256": "ee9e5899d1f150982eb05ca1f8794406cef0ffc80eb8223811b504ff972370c1", "supports": ["identity", "details.repository_revision", "summary", "environments", "notes", "licenses.code"], "url": "https://api.github.com/repos/anikasomaia/local-activity-recorder/git/trees/f7ec5b6ecf0ba1bfe6362d7f2bfb0f1af5ac6bc7?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:00.168119+00:00", "role": "primary", "sha256": "ce3022b4016d140eac554e392dbe42f54ffb3e8098b03e2bf72e92ff49c5d084", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/Program.cs"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:03.421222+00:00", "role": "primary", "sha256": "e70e2e68b5dd586e5ad0fb0b12c8b937b4614bd5237bac61d03d22144c4751f4", "supports": ["notes"], "url": "https://raw.githubusercontent.com/anikasomaia/local-activity-recorder/967cea087bf6496b3b82743fb96770fa29f8e203/DataUploader.cs"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-20T11:57:15.716415+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/anikasomaia/local-activity-recorder"}], "name": "Activity Recorder", "publication": null, "published_at": null, "record_sha256": "a2ee5d2025b4a454d5e5f61f7da5a45921d49e418a065162dc13b615f13ad792", "relations": [], "resource_id": "github:anikasomaia/local-activity-recorder", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/anikasomaia/local-activity-recorder", "topics": ["recording", "data-generation"], "updated_at": null}, "text": "Activity Recorder\nResource type: tool\nReview status: curated\nScope: core\n\nWindows tray utility that records primary-display screenshots, keyboard key codes and mouse events as local JPEG and JSON files for training-data collection.\nPlatforms: desktop, windows\nTopics: recording, data-generation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://api.github.com/repos/anikasomaia/local-activity-recorder/git/trees/f7ec5b6ecf0ba1bfe6362d7f2bfb0f1af5ac6bc7?recursive=1\"], \"reason\": \"No license declaration was found in the inspected README/C# files, and the complete pinned tree has no LICENSE, LICENCE or COPYING-named file. Public source access does not establish reuse permission.\"}}, \"repository_revision\": \"967cea087bf6496b3b82743fb96770fa29f8e203\"}\n\nRecording is started and stopped from the tray; screenshots use a five-second timer. The manual Upload Data action uses a built-in third-party demonstration endpoint unless changed in code, sending recorded files as base64 JSON. Successful uploads move files into uploaded subfolders rather than delete them; no automatic expiry policy was found in the inspected capture/upload path. Sensitive-application filters are listed as future work. The README's .NET version requirements are inconsistent, and long-session behavior was not established. No license declaration was found in the inspected files or complete tree's license-file names, so code permission remains unknown. No activity was captured and no uploader or executable was run.\n\nPrimary source: https://github.com/anikasomaia/local-activity-recorder\nCatalog record: https://awesomecua.github.io/resources/github-anikasomaia-local-activity-recorder/"}
{"id": "github:ant-research/guihallubench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ant-research/GUIHalluBench", "citation_url": "https://awesomecua.github.io/resources/github-ant-research-guihallubench/", "details": {"benchmark_protocol": {"actions": ["list of UI element boxes [x1,y1,x2,y2] and text descriptions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "population": "English annotation subset reported by README", "qualifier": "exact", "unit": "page-annotation entries", "value": 1000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "population": "Chinese annotation subset reported by README", "qualifier": "exact", "unit": "page-annotation entries", "value": 1000}], "evaluation": "Metrics use greedy one-to-one matching in descending IoU order, then micro precision/recall/F1 and lowercased difflib.SequenceMatcher description similarity on matched pairs. Actual metrics.py accepts IoU >= 0.5 at its default threshold; README prose says > 0.5. Predictions align to annotations by image_path; callers must supply complete prediction coverage because missing predictions are not independently enumerated by evaluate_dataset.", "evidence_urls": ["https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/src/metrics.py", "https://api.github.com/repos/ant-research/GUIHalluBench/git/trees/master?recursive=1", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/LEGAL.md", "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/.gitattributes"], "focus": ["GUI page-element detection", "element-description consistency"], "notes": "Each annotation entry represents a page, not one UI element. README’s illustrative num_samples=4796 output is not adopted as a corpus count. Public tree contains both annotation JSON files and an images.tar.gz Git-LFS pointer; payload availability/integrity was not tested by downloading images. LEGAL.md only governs comment-language precedence, not a code/data license grant; both slots remain unknown. No benchmark run was performed.", "observations": ["page screenshot"], "setting": "offline"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:51Z", "role": "primary", "sha256": "51df9dbcf52f46e4774207f2e6ad84dc8a6d117cdab6f77a38e9897ded1d0963", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/ant-research/GUIHalluBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "d9c476d6960a4ef99e20d4c3b5d35a63f3ae5e61fc20e9dea00b11e64b74c9e9", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "f36707dfcd3e83fc9d7f57b50d0c27007801e5029cc1ac255e000560ecf778fc", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/src/metrics.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:07Z", "role": "primary", "sha256": "c0896bc90d515b7bd8fe86b062aeb199bd47570e6eca8474986f5b1ef1d98dbd", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/ant-research/GUIHalluBench/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "adba50d9794b9ef3f7ec8cbc680f7f1fa3fbf9df0ac8d1f9b9ccab6d941bc11b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/LEGAL.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:34:56Z", "role": "primary", "sha256": "8f35fa75be42a7a53e4f96de1b62be84962d2cf1beb8b23dfcfaecfaea4a6b90", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/.gitattributes"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ant-research/GUIHalluBench"}], "name": "GUIHalluBench", "publication": null, "published_at": null, "record_sha256": "7af05c68543a14293b80ae3adf3df0cf9ed13cab977c4bd8dc1cc9e54a860dc3", "relations": [], "resource_id": "github:ant-research/guihallubench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ant-research/GUIHalluBench", "topics": ["evaluation", "robustness"], "updated_at": null}, "text": "GUIHalluBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nMeasures hallucination in UI-element parsing by matching the boxes and descriptions a model lists for 1,000 English and 1,000 Chinese screenshots against human annotations.\nPlatforms: cross-platform\nTopics: evaluation, robustness\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"list of UI element boxes [x1,y1,x2,y2] and text descriptions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md\", \"population\": \"English annotation subset reported by README\", \"qualifier\": \"exact\", \"unit\": \"page-annotation entries\", \"value\": 1000}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md\", \"population\": \"Chinese annotation subset reported by README\", \"qualifier\": \"exact\", \"unit\": \"page-annotation entries\", \"value\": 1000}], \"evaluation\": \"Metrics use greedy one-to-one matching in descending IoU order, then micro precision/recall/F1 and lowercased difflib.SequenceMatcher description similarity on matched pairs. Actual metrics.py accepts IoU >= 0.5 at its default threshold; README prose says > 0.5. Predictions align to annotations by image_path; callers must supply complete prediction coverage because missing predictions are not independently enumerated by evaluate_dataset.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/README.md\", \"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/src/metrics.py\", \"https://api.github.com/repos/ant-research/GUIHalluBench/git/trees/master?recursive=1\", \"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/LEGAL.md\", \"https://raw.githubusercontent.com/ant-research/GUIHalluBench/HEAD/.gitattributes\"], \"focus\": [\"GUI page-element detection\", \"element-description consistency\"], \"notes\": \"Each annotation entry represents a page, not one UI element. README’s illustrative num_samples=4796 output is not adopted as a corpus count. Public tree contains both annotation JSON files and an images.tar.gz Git-LFS pointer; payload availability/integrity was not tested by downloading images. LEGAL.md only governs comment-language precedence, not a code/data license grant; both slots remain unknown. No benchmark run was performed.\", \"observations\": [\"page screenshot\"], \"setting\": \"offline\"}}\n\nPrimary source: https://github.com/ant-research/GUIHalluBench\nCatalog record: https://awesomecua.github.io/resources/github-ant-research-guihallubench/"}
{"id": "github:apga/go-browse", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ApGa/Go-Browse", "citation_url": "https://awesomecua.github.io/resources/github-apga-go-browse/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "b02ec75f69ab8bbebe0d59eeccdcdcacf11a0d6b00bdfe803c692a9d587301e2", "supports": ["identity"], "url": "https://api.github.com/repos/ApGa/Go-Browse"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "cd3a40a596ad9ead856973bb31e5b7f7e88a2f60969799c01a2a8a6d09fe3421", "supports": ["release_status"], "url": "https://api.github.com/repos/ApGa/Go-Browse/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e7107f1e1b3a0c65cbbae6c496d9e71b570025bdddebb3d402a6080dd916bcba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e7107f1e1b3a0c65cbbae6c496d9e71b570025bdddebb3d402a6080dd916bcba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ApGa/Go-Browse"}], "name": "Go-Browse", "publication": null, "published_at": null, "record_sha256": "5bc561caa57be1baa8b6c09716ae3bc40746873244aaa4bc629ae2790bae9b7e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "resource_id": "github:apga/go-browse", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ApGa/Go-Browse", "topics": ["web-navigation", "data-generation", "imitation-learning"], "updated_at": null}, "text": "Go-Browse\nResource type: framework\nReview status: curated\nScope: core\n\nStructured website-exploration and data-processing code that revisits discovered pages, proposes tasks and collects trajectories for web-agent fine-tuning.\nPlatforms: web\nTopics: web-navigation, data-generation, imitation-learning\n\nThe release contains exploration agents, data-processing scripts, fine-tuning and WebArena evaluation entry points under MIT. Checkpoints and processed datasets are separate Hub resources with separate license status. The arXiv badge image has an unrelated number; the linked paper and citation consistently identify 2506.03533.\n\nPrimary source: https://github.com/ApGa/Go-Browse\nCatalog record: https://awesomecua.github.io/resources/github-apga-go-browse/"}
{"id": "github:appium/appium", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/appium/appium", "citation_url": "https://awesomecua.github.io/resources/github-appium-appium/", "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "ecd5b8033cbc17a2ce5aebf1b35260b77c4906aea7c549a0b86e5d55b9ebe135", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/appium/appium/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "5cb791a45cafc4bdaef57f3ba49fb7691a98c756e8148467f9e241295728490f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/appium/appium/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/appium/appium/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/appium/appium"}], "name": "Appium", "publication": null, "published_at": null, "record_sha256": "20364a09d8dd5c06b51cb54b8746eaadc1b74e85841d8a06c3fec2cd69debe21", "relations": [], "resource_id": "github:appium/appium", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/appium/appium", "topics": ["automation"], "updated_at": null}, "text": "Appium\nResource type: infrastructure\nReview status: curated\nScope: adjacent\n\nOpen-source WebDriver-based automation server for mobile, desktop and IoT apps, extended by per-platform drivers for iOS, Android, macOS and Windows, plugins, and clients in Java, Python, Ruby and C#.\nPlatforms: android, ios\nTopics: automation\n\nPrimary source: https://github.com/appium/appium\nCatalog record: https://awesomecua.github.io/resources/github-appium-appium/"}
{"id": "github:apple/ml-ferret", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/apple/ml-ferret", "citation_url": "https://awesomecua.github.io/resources/github-apple-ml-ferret/", "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "1973f702c08ca9637ce03bc28fc34188be05826e9d43f9992b08e804f474b1a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/apple/ml-ferret"}], "name": "Ferret-UI", "publication": null, "published_at": null, "record_sha256": "d3f2b12c2dbb5b9aca14ae2556f50fd57e0d537114f6850dee2e361fb3dd522c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/apple-aiml-research/ml-ferret/main/ferretui/README.md", "target_id": "arxiv:2404.05719", "type": "documented_by"}], "resource_id": "github:apple/ml-ferret", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/apple/ml-ferret", "topics": ["screen-understanding", "grounding"], "updated_at": null}, "text": "Ferret-UI\nResource type: model\nReview status: curated\nScope: core\n\nScreen understanding research within the Ferret model family.\nPlatforms: android, ios\nTopics: screen-understanding, grounding\n\nThe repository has moved from apple/ml-ferret to apple-aiml-research/ml-ferret; evidence uses the dedicated ferretui component.\n\nPrimary source: https://github.com/apple/ml-ferret\nCatalog record: https://awesomecua.github.io/resources/github-apple-ml-ferret/"}
{"id": "github:ariaui/aria-ui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/AriaUI/Aria-UI", "citation_url": "https://awesomecua.github.io/resources/github-ariaui-aria-ui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/AriaUI/Aria-UI"}], "name": "Aria-UI", "publication": null, "published_at": null, "record_sha256": "9bf6e4e2b386dac142390c023f104651b574915843adc4e159e1d7d9d828edb2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "resource_id": "github:ariaui/aria-ui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/AriaUI/Aria-UI", "topics": ["grounding"], "updated_at": null}, "text": "Aria-UI\nResource type: model\nReview status: curated\nScope: core\n\nMixture-of-experts GUI grounding model with 3.9B activated parameters that can use text or interleaved image history, with released training data and an AndroidWorld agent.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/AriaUI/Aria-UI\nCatalog record: https://awesomecua.github.io/resources/github-ariaui-aria-ui/"}
{"id": "github:asweigart/pyautogui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/asweigart/pyautogui", "citation_url": "https://awesomecua.github.io/resources/github-asweigart-pyautogui/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "9e7a23adb62bd05d0fd45e5c95a57fcfc3cc4d19ab89d6801c5bf1e23d3c678c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/asweigart/pyautogui/HEAD/README.md"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/asweigart/pyautogui"}], "name": "PyAutoGUI", "publication": null, "published_at": null, "record_sha256": "786c32ff111ab5f13f0c9d663cc82f27cbd69815dfa54c2429d4bd8d82848158", "relations": [], "resource_id": "github:asweigart/pyautogui", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/asweigart/pyautogui", "topics": ["automation"], "updated_at": null}, "text": "PyAutoGUI\nResource type: infrastructure\nReview status: curated\nScope: adjacent\n\nCross-platform Python module for scripted mouse and keyboard control, message boxes and screenshots with on-screen image search, working on the primary monitor under Windows, macOS and Linux.\nPlatforms: desktop\nTopics: automation\n\nPrimary source: https://github.com/asweigart/pyautogui\nCatalog record: https://awesomecua.github.io/resources/github-asweigart-pyautogui/"}
{"id": "github:baai-agents/cradle", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/BAAI-Agents/Cradle", "citation_url": "https://awesomecua.github.io/resources/github-baai-agents-cradle/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "fe475d3e92c42807b792c542bc5986bf74caa7d166aea860d368fbe776342a82", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "2964861ad96d977da95d891f62d8f5d60873f4a36da0b51f052d9e98ef97d02f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/BAAI-Agents/Cradle"}], "name": "Cradle", "publication": null, "published_at": null, "record_sha256": "5de36aa169f39264ab769567996a0c0adb86d13f67f68929f86196a0dbc0200c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/BAAI-Agents/Cradle/HEAD/README.md", "target_id": "arxiv:2403.03186", "type": "documented_by"}], "resource_id": "github:baai-agents/cradle", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/BAAI-Agents/Cradle", "topics": ["planning", "memory"], "updated_at": null}, "text": "Cradle\nResource type: agent\nReview status: curated\nScope: core\n\nFramework in which foundation models control games and software from screenshots alone, acting through keyboard and mouse, shown on RDR2, Stardew Valley, Cities: Skylines and Dealer's Life 2.\nPlatforms: desktop\nTopics: planning, memory\n\nPrimary source: https://github.com/BAAI-Agents/Cradle\nCatalog record: https://awesomecua.github.io/resources/github-baai-agents-cradle/"}
{"id": "github:bravegroup/sheetcopilot", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/BraveGroup/SheetCopilot", "citation_url": "https://awesomecua.github.io/resources/github-bravegroup-sheetcopilot/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md", "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"], "reason": "The retained README restricts SheetCopilot and its dataset to noncommercial use, while the retained repository LICENSE contains the GNU GPL version 3 text. These declarations do not establish one unambiguous code license."}}}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "beb8c811bf2d070845aa6fcc7e94ffd43ca4f9f60ce42c72d958598790fbdffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code", "notes"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/BraveGroup/SheetCopilot"}], "name": "SheetCopilot", "publication": null, "published_at": null, "record_sha256": "9abb09997a00572ecdd8e88b8758f99569deb2d78973ccbf2a47826d09b76d52", "relations": [{"evidence_url": "https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md", "target_id": "arxiv:2305.19308", "type": "documented_by"}], "resource_id": "github:bravegroup/sheetcopilot", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/BraveGroup/SheetCopilot", "topics": ["workflow", "tool-integration", "evaluation"], "updated_at": null}, "text": "SheetCopilot\nResource type: agent\nReview status: curated\nScope: core\n\nSpreadsheet automation agent for Windows/Excel, with benchmark assets and a separately released headless Ubuntu evaluation backend.\nPlatforms: desktop, windows, linux\nTopics: workflow, tool-integration, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/README.md\", \"https://raw.githubusercontent.com/BraveGroup/SheetCopilot/HEAD/LICENSE\"], \"reason\": \"The retained README restricts SheetCopilot and its dataset to noncommercial use, while the retained repository LICENSE contains the GNU GPL version 3 text. These declarations do not establish one unambiguous code license.\"}}}\n\nThe original agent and the later Ubuntu evaluator have different roles; the repository does not establish that the original agent runs on Ubuntu. Code licensing is unresolved because the retained README and LICENSE contain conflicting declarations.\n\nPrimary source: https://github.com/BraveGroup/SheetCopilot\nCatalog record: https://awesomecua.github.io/resources/github-bravegroup-sheetcopilot/"}
{"id": "github:browser-use/browser-use", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/browser-use/browser-use", "citation_url": "https://awesomecua.github.io/resources/github-browser-use-browser-use/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fcca30a036d122e34c943740c29d8cb91642f0c615050a8d54ec891c0eee56d2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "135c57671b0ee957669b05a0ca00cc06f645496d351e2e732832c6e276da9a5e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browser-use/browser-use"}], "name": "Browser Use", "publication": null, "published_at": null, "record_sha256": "f4147dcd1ec0b10d8c8a608a19ce35beb1ed0991db0c453f53e3821edc681246", "relations": [{"evidence_url": "https://raw.githubusercontent.com/browser-use/browser-use/HEAD/README.md", "target_id": "organization:browser-use", "type": "published_by"}], "resource_id": "github:browser-use/browser-use", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/browser-use/browser-use", "topics": ["web-navigation"], "updated_at": null}, "text": "Browser Use\nResource type: framework\nReview status: curated\nScope: core\n\nOpen-source browser agent library that runs locally with any LLM against a local or cloud browser, and also serves as a CLI for coding agents and as a driver for Claude's browser toolset.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/browser-use/browser-use\nCatalog record: https://awesomecua.github.io/resources/github-browser-use-browser-use/"}
{"id": "github:browser-use/web-ui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/browser-use/web-ui", "citation_url": "https://awesomecua.github.io/resources/github-browser-use-web-ui/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v3.0.0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "5047bcf2f0e92f0d34a57ac7d83166903a16a9909c700c0647b8a48363bd94f9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "7fb0738bc9b837f7b3e645c406e04d5c7c4d0860ea636a96e4e9d84d9510dd82", "supports": ["identity"], "url": "https://api.github.com/repos/browser-use/web-ui"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fcbf1775e3513625103d389c737e98e939e17140baac9bc8000b582ab2729a46", "supports": ["release_status"], "url": "https://api.github.com/repos/browser-use/web-ui/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "f87ee752c38653c447eed4a9abd83e71dedd40298758bcb9aa993b731362c453", "supports": ["release_status"], "url": "https://api.github.com/repos/browser-use/web-ui/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "612d1382ec1a426eaf3672aafb13bafcfe91b2c90a0fe339822ffc9c895003ca", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "612d1382ec1a426eaf3672aafb13bafcfe91b2c90a0fe339822ffc9c895003ca", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:01:01Z", "url": "https://raw.githubusercontent.com/browser-use/web-ui/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browser-use/web-ui"}], "name": "Browser Use WebUI", "publication": null, "published_at": null, "record_sha256": "81df57299c7ae88f62e3727f913395ff796901878e323d1277556066cf5bbd2b", "relations": [], "resource_id": "github:browser-use/web-ui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/browser-use/web-ui", "topics": ["web-navigation", "tool-integration", "recording"], "updated_at": null}, "text": "Browser Use WebUI\nResource type: tool\nReview status: curated\nScope: core\n\nA Gradio interface for configuring and running Browser Use agents, including persistent browser sessions and interaction recordings.\nPlatforms: web\nTopics: web-navigation, tool-integration, recording\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"v3.0.0\"}\n\nThe repository publishes UI source and release v3.0.0 under MIT. It is a distinct user interface built on the existing Browser Use framework, not a second framework identity. The UI supports several external model providers whose terms are separate; no dependencies were installed.\n\nPrimary source: https://github.com/browser-use/web-ui\nCatalog record: https://awesomecua.github.io/resources/github-browser-use-web-ui/"}
{"id": "github:browserbase/stagehand", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/browserbase/stagehand", "citation_url": "https://awesomecua.github.io/resources/github-browserbase-stagehand/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fb2039ee569e37cf1d652ca26c0a1951a318d4ac3240018b958985c0aec91007", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "fac7e7cd1cfc504054ee1fc53c38e0f3099d7357917e85a73792f6511fe0ea96", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browserbase/stagehand"}], "name": "Stagehand", "publication": null, "published_at": null, "record_sha256": "2f6ff4fc9aa7a52d4e690eebd1dc92b715d67652394c9d108ba3cffec22ec0a0", "relations": [{"evidence_url": "https://raw.githubusercontent.com/browserbase/stagehand/HEAD/README.md", "target_id": "organization:browserbase", "type": "published_by"}], "resource_id": "github:browserbase/stagehand", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/browserbase/stagehand", "topics": ["web-navigation"], "updated_at": null}, "text": "Stagehand\nResource type: framework\nReview status: curated\nScope: core\n\nBrowser automation framework combining code and natural-language actions.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/browserbase/stagehand\nCatalog record: https://awesomecua.github.io/resources/github-browserbase-stagehand/"}
{"id": "github:browseros-ai/browseros", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/browseros-ai/BrowserOS", "citation_url": "https://awesomecua.github.io/resources/github-browseros-ai-browseros/", "details": {"availability": "Published source; official release metadata inspected", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE", "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/README.md"], "reason": "The repository explicitly identifies its code as AGPL-3.0; this declaration does not cover every incorporated third-party component."}}, "release_tag": "browserclaw/v0.51.1.2"}, "environments": ["web", "desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "bfe2979ae0e99985154042090f483e172cca229fe408f8f9b47102662beec983", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.code"], "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "2b8c90b9fb2edffccf7406952465015df87622331921fe0f25a44fc9f5a9b86b", "supports": ["identity"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "1f43b364f532a8bdd798df5dd86d3b73f8f545ef73a66e9b3b4da59ea9d0faa1", "supports": ["release_status"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "87bea16220429a93fefc2b2a57452fd67b2618784025db18eb1f899c0396098f", "supports": ["release_status"], "url": "https://api.github.com/repos/browseros-ai/BrowserOS/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "8486a10c4393cee1c25392769ddd3b2d6c242d6ec7928e1414efff7dfb2f07ef", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "Repository LICENSE and explicit README declaration; incorporated Chromium and other components retain their own terms.", "declared": "AGPL-3.0", "retrieved_at": "2026-10-17T04:00:59Z", "url": "https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/browseros-ai/BrowserOS"}], "name": "BrowserOS / BrowserOS neo", "publication": null, "published_at": null, "record_sha256": "8556a2007dfe3ef5de75ace76dbdcc52321b90b2471ab167a3596ef9931dbe35", "relations": [], "resource_id": "github:browseros-ai/browseros", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/browseros-ai/BrowserOS", "topics": ["web-navigation", "tool-integration", "recording"], "updated_at": null}, "text": "BrowserOS / BrowserOS neo\nResource type: infrastructure\nReview status: curated\nScope: core\n\nChromium-based browser implementations for external agent control and built-in assistant workflows, with session visibility and replay facilities.\nPlatforms: web, desktop, windows, macos, linux\nTopics: web-navigation, tool-integration, recording\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/LICENSE\", \"https://raw.githubusercontent.com/browseros-ai/BrowserOS/HEAD/README.md\"], \"reason\": \"The repository explicitly identifies its code as AGPL-3.0; this declaration does not cover every incorporated third-party component.\"}}, \"release_tag\": \"browserclaw/v0.51.1.2\"}\n\nThe one repository includes BrowserOS neo and BrowserOS; they are retained as one source-bundle resource. Official release metadata lists desktop installers, but none were downloaded. The repository declares AGPL-3.0 and retains separate Chromium-related licensing. Publisher privacy and performance claims were not independently verified.\n\nPrimary source: https://github.com/browseros-ai/BrowserOS\nCatalog record: https://awesomecua.github.io/resources/github-browseros-ai-browseros/"}
{"id": "github:bytebot-ai/bytebot", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/bytebot-ai/bytebot", "citation_url": "https://awesomecua.github.io/resources/github-bytebot-ai-bytebot/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "796cc24df9874c71408d32d12135b6b04389c455134c6c722ca34711288beb40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/bytebot-ai/bytebot/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytebot-ai/bytebot"}], "name": "Bytebot", "publication": null, "published_at": null, "record_sha256": "9206eae6da6d3784896e88bc7f98d49c76b03e1e6579e2c6fd13e4e892e4b37d", "relations": [], "resource_id": "github:bytebot-ai/bytebot", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/bytebot-ai/bytebot", "topics": ["workflow"], "updated_at": null}, "text": "Bytebot\nResource type: agent\nReview status: curated\nScope: core\n\nOpen-source desktop agent that works inside its own Ubuntu Linux virtual desktop, with a web task interface, REST APIs for tasks and direct control, a takeover mode and password-manager logins.\nPlatforms: desktop\nTopics: workflow\n\nPrimary source: https://github.com/bytebot-ai/bytebot\nCatalog record: https://awesomecua.github.io/resources/github-bytebot-ai-bytebot/"}
{"id": "github:bytedance-seed/agent-r", "metadata": {"aliases": ["github:bytedance/agent-r", "bytedance/Agent-R", "https://github.com/bytedance/Agent-R"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ByteDance-Seed/Agent-R", "citation_url": "https://awesomecua.github.io/resources/github-bytedance-seed-agent-r/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "8471a0761a94b1af84099a80efb6ac53f17e34491a196e82ca068a5113c0b8be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["identity", "availability"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "e3c70b04094e8e77e249065b2b500ac197fc6fafee6e3275d78289b3790bf7ed", "supports": ["release_status"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "283ea6cc2997a1a70da0049e09adf9317bb60ca1b51279b65196b83a69e1996b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "283ea6cc2997a1a70da0049e09adf9317bb60ca1b51279b65196b83a69e1996b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T18:05:46Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["aliases"], "url": "https://api.github.com/repos/bytedance/Agent-R"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T18:05:46Z", "role": "primary", "sha256": "3fff7137bc4f39a3a757f3b820892fb330449044ef2721e9f9d7719d23ff2b98", "supports": ["aliases"], "url": "https://api.github.com/repos/ByteDance-Seed/Agent-R"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:05:22Z", "url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ByteDance-Seed/Agent-R"}], "name": "Agent-R", "publication": null, "published_at": null, "record_sha256": "334bda9df6df5e6e2553d2bbf3789535594b41ada4a3f5010b5b31c1a6ff1175", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ByteDance-Seed/Agent-R/HEAD/README.md", "target_id": "arxiv:2501.11425", "type": "documented_by"}], "resource_id": "github:bytedance-seed/agent-r", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/ByteDance-Seed/Agent-R", "topics": ["self-reflection", "data-generation", "web-navigation"], "updated_at": null}, "text": "Agent-R\nResource type: framework\nReview status: curated\nScope: adjacent\n\nTrajectory collection and revision utilities that use tree search and model-guided critique to train agents to recover from mistakes, including WebShop tasks.\nRecorded aliases: github:bytedance/agent-r, bytedance/Agent-R, https://github.com/bytedance/Agent-R\nPlatforms: web\nTopics: self-reflection, data-generation, web-navigation\n\nThe canonical repository is ByteDance-Seed/Agent-R and is archived but readable. MCTS collection, revision and evaluation scripts are released under Apache-2.0. It also supports non-GUI text environments; this entry covers its WebShop interaction component. No separate weight release was verified.\n\nPrimary source: https://github.com/ByteDance-Seed/Agent-R\nCatalog record: https://awesomecua.github.io/resources/github-bytedance-seed-agent-r/"}
{"id": "github:bytedance/gui-rewalk", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/bytedance/GUI-ReWalk", "citation_url": "https://awesomecua.github.io/resources/github-bytedance-gui-rewalk/", "details": {"availability": "Public desktop data-generation source; separate training-data and trained-model releases not verified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"], "reason": "Repository code license; no scope transfer to generated trajectories, pretrained detector weights, or GUI application content."}}, "repository_revision": "78a65d003d6f48802fac879508b62f1c89125892"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "63708cb2500aefa36916f3fb171b595efaabf614f17b925cfe8b5c8cbf72be10", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "34addf6eed41e1e29e51aab8819c5ac316b5699a8c11c4ce29a32c8e38557de8", "supports": ["identity"], "url": "https://api.github.com/repos/bytedance/GUI-ReWalk"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "a218cb6f8a2c2a243e8207785c98255bf717c00f177521368705d8652e229d7d", "supports": ["release_status"], "url": "https://api.github.com/repos/bytedance/GUI-ReWalk/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:08Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:15Z", "licenses": {"code": {"basis": "Repository code license; no scope transfer to generated trajectories, pretrained detector weights, or GUI application content.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:08Z", "url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/GUI-ReWalk"}], "name": "GUI-ReWalk", "publication": null, "published_at": null, "record_sha256": "985c46355c215c6f02c502323bb7b9a6b6cfe0de414b0d92f1d4d8ba33e3b3d8", "relations": [{"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "github:xlang-ai/osworld", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "github:microsoft/omniparser", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/README.md", "target_id": "arxiv:2509.15738", "type": "documented_by"}], "resource_id": "github:bytedance/gui-rewalk", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/bytedance/GUI-ReWalk", "topics": ["data-generation", "planning", "cross-app"], "updated_at": null}, "text": "GUI-ReWalk\nResource type: framework\nReview status: curated\nScope: core\n\nA desktop trajectory-generation framework that moves from stochastic GUI exploration to inferred goals and multi-stage task completion, with retrospective reasoning annotations.\nPlatforms: desktop, linux\nTopics: data-generation, planning, cross-app\n\nRecorded specifications: {\"availability\": \"Public desktop data-generation source; separate training-data and trained-model releases not verified.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/bytedance/GUI-ReWalk/HEAD/LICENSE\"], \"reason\": \"Repository code license; no scope transfer to generated trajectories, pretrained detector weights, or GUI application content.\"}}, \"repository_revision\": \"78a65d003d6f48802fac879508b62f1c89125892\"}\n\nThe inspected source tree contains the random walker, desktop environment integration and run_rewalk.sh. The documented runnable setup uses OSWorld and OmniParser. The README also discusses mobile data and a GUI-ReWalk-7B model, but no separately published corpus or model checkpoint was verified; those are not counted as released assets. Apache-2.0 is scoped to the implementation.\n\nPrimary source: https://github.com/bytedance/GUI-ReWalk\nCatalog record: https://awesomecua.github.io/resources/github-bytedance-gui-rewalk/"}
{"id": "github:bytedance/ui-tars", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/bytedance/UI-TARS", "citation_url": "https://awesomecua.github.io/resources/github-bytedance-ui-tars/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "0c4feb17b657c0c9a71d7e0c5e2142abbd19bb5cff22e6cd3b6b082644c98ee7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/UI-TARS"}], "name": "UI-TARS", "publication": null, "published_at": null, "record_sha256": "29dddacf2bd0ceb85d180b8d49e7474a9a4adfc7c0c3de4d517d73745dd41d2b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/bytedance/UI-TARS/HEAD/README.md", "target_id": "arxiv:2501.12326", "type": "documented_by"}], "resource_id": "github:bytedance/ui-tars", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/bytedance/UI-TARS", "topics": ["grounding", "planning"], "updated_at": null}, "text": "UI-TARS\nResource type: model\nReview status: curated\nScope: core\n\nNative GUI agent models; UI-TARS-1.5 adds reinforcement-learned reasoning before acting and an open 7B checkpoint, and UI-TARS-2 extends to games, code and tool use.\nPlatforms: cross-platform\nTopics: grounding, planning\n\nPrimary source: https://github.com/bytedance/UI-TARS\nCatalog record: https://awesomecua.github.io/resources/github-bytedance-ui-tars/"}
{"id": "github:bytedance/ui-tars-desktop", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/bytedance/UI-TARS-desktop", "citation_url": "https://awesomecua.github.io/resources/github-bytedance-ui-tars-desktop/", "details": {}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "0f05aa4f447d43439f6b3f4ab179cbe1e59d170eaf6c0684fa9c2eb0872e5985", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/bytedance/UI-TARS-desktop/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bytedance/UI-TARS-desktop"}], "name": "UI-TARS Desktop", "publication": null, "published_at": null, "record_sha256": "dee42335f276048d87b5364503a45b523a1352dce36623af941fff775015d80b", "relations": [], "resource_id": "github:bytedance/ui-tars-desktop", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/bytedance/UI-TARS-desktop", "topics": ["workflow"], "updated_at": null}, "text": "UI-TARS Desktop\nResource type: agent\nReview status: curated\nScope: core\n\nDesktop app for a local GUI agent driven by UI-TARS and Seed-1.5-VL/1.6 models, offering local and remote operators for computers and browsers; the repository also ships the Agent TARS CLI and web UI.\nPlatforms: desktop, web\nTopics: workflow\n\nPrimary source: https://github.com/bytedance/UI-TARS-desktop\nCatalog record: https://awesomecua.github.io/resources/github-bytedance-ui-tars-desktop/"}
{"id": "github:bz-lab/auitestagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/bz-lab/AUITestAgent", "citation_url": "https://awesomecua.github.io/resources/github-bz-lab-auitestagent/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "dc60f683b0f44ccee4e54e5cbe1552ee3454e1ec4ed5c04635fedb83a325f51f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/bz-lab/AUITestAgent/HEAD/README.md"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/bz-lab/AUITestAgent"}], "name": "AUITestAgent", "publication": null, "published_at": null, "record_sha256": "5e40656b8545a63e7bbac735e5a929bf662fc9f7c710a4d677f7b551bd12ad6c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/bz-lab/AUITestAgent/HEAD/README.md", "target_id": "arxiv:2407.09018", "type": "documented_by"}], "resource_id": "github:bz-lab/auitestagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/bz-lab/AUITestAgent", "topics": ["automation", "evaluation"], "updated_at": null}, "text": "AUITestAgent\nResource type: tool\nReview status: curated\nScope: core\n\nGenerates mobile GUI interactions from natural-language test requirements and checks application behavior against those requirements.\nPlatforms: unspecified\nTopics: automation, evaluation\n\nIts README describes mobile apps without naming an operating system.\n\nPrimary source: https://github.com/bz-lab/AUITestAgent\nCatalog record: https://awesomecua.github.io/resources/github-bz-lab-auitestagent/"}
{"id": "github:caap-agent/caap-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/caap-agent/caap-agent", "citation_url": "https://awesomecua.github.io/resources/github-caap-agent-caap-agent/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. The inspected text does not establish -only versus -or-later for the project."}}, "repository_tree_sha": "9d65c0f30e3c119db533dd832863b4f3f7b74b18", "repository_url": "https://github.com/caap-agent/caap-agent"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "481db1288ffc769f48b19c83f0db4fb40a64dc1b05859f776b77f0774bfcdb90", "supports": ["name", "kind", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "8d518cad746591e82b54b63b79d0a1d47531a75f02c8baf2b560877fb03918ba", "supports": ["identity"], "url": "https://api.github.com/repos/caap-agent/caap-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:24Z", "role": "primary", "sha256": "c511d94ce58e73a87c9647169b1199c0ed0029e6b1df271ad1fcd3ca86f3458b", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/caap-agent/caap-agent/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["summary", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "5e6fa1bfd473724e67366c42761db27fe091ee00fdcfbe599b9dd5d5a9311d36", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. The inspected text does not establish -only versus -or-later for the project.", "declared": "AGPL-3.0", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/caap-agent/caap-agent"}], "name": "CAAP Agent", "publication": null, "published_at": null, "record_sha256": "7b5c257fe968ff0682decd12a5bfcd38b3069f7ed1be25325e37ed9b23e8b2ca", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50", "target_id": "arxiv:2406.06947", "type": "documented_by"}], "resource_id": "github:caap-agent/caap-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/caap-agent/caap-agent", "topics": ["planning", "screen-understanding", "web-navigation"], "updated_at": null}, "text": "CAAP Agent\nResource type: agent\nReview status: curated\nScope: core\n\nScreenshot-driven agent implementation for MiniWoB++ and WebShop, using a visual observer and context-aware prompts to issue mouse and keyboard actions.\nPlatforms: web\nTopics: planning, screen-understanding, web-navigation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/caap-agent/caap-agent/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate. The inspected text does not establish -only versus -or-later for the project.\"}}, \"repository_tree_sha\": \"9d65c0f30e3c119db533dd832863b4f3f7b74b18\", \"repository_url\": \"https://github.com/caap-agent/caap-agent\"}\n\nCode is released for the two named browser benchmarks. Visual-observer checkpoints are linked separately through Google Drive and were not downloaded or access-tested. Root LICENSE begins with GNU AGPL v3 and includes third-party notices; this code declaration is not transferred to bundled/generated data or external weights.\n\nPrimary source: https://github.com/caap-agent/caap-agent\nCatalog record: https://awesomecua.github.io/resources/github-caap-agent-caap-agent/"}
{"id": "github:camel-ai/crab", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/camel-ai/crab", "citation_url": "https://awesomecua.github.io/resources/github-camel-ai-crab/", "details": {"benchmark_protocol": {"actions": ["Python-defined environment actions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "CRAB Benchmark-v0 (paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.01511", "unit": "tasks", "value": 120}], "evaluation": "Graph-based evaluators measure progress through task requirements.", "evidence_urls": ["https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "focus": ["cross-environment tasks"], "notes": "CRAB is also an extensible framework. The 120-task count belongs to Benchmark-v0, not every environment that can be created with the framework.", "observations": [], "setting": "interactive"}}, "environments": ["desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "7b894c5474ba78294427cfd4ef0eea8cb22d76dfa0c2e8c7f01e31ad927064e9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/camel-ai/crab"}], "name": "CRAB", "publication": null, "published_at": null, "record_sha256": "7a2050a8514623e7fd5c6ba9af37d65925abc097994669f3f335cce9bda7856e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md", "target_id": "arxiv:2407.01511", "type": "documented_by"}], "resource_id": "github:camel-ai/crab", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/camel-ai/crab", "topics": ["evaluation", "cross-app"], "updated_at": null}, "text": "CRAB\nResource type: benchmark\nReview status: curated\nScope: core\n\nCross-environment benchmark framework for multimodal agents.\nPlatforms: desktop, android\nTopics: evaluation, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"Python-defined environment actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\", \"population\": \"CRAB Benchmark-v0 (paper)\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2407.01511\", \"unit\": \"tasks\", \"value\": 120}], \"evaluation\": \"Graph-based evaluators measure progress through task requirements.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/camel-ai/crab/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\"], \"focus\": [\"cross-environment tasks\"], \"notes\": \"CRAB is also an extensible framework. The 120-task count belongs to Benchmark-v0, not every environment that can be created with the framework.\", \"observations\": [], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/camel-ai/crab\nCatalog record: https://awesomecua.github.io/resources/github-camel-ai-crab/"}
{"id": "github:charlypoly/web-context-bench-python", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/charlypoly/web-context-bench-python", "citation_url": "https://awesomecua.github.io/resources/github-charlypoly-web-context-bench-python/", "details": {"benchmark_protocol": {"actions": ["READ text answer", "ACT selector, text/label, accessible role/name, DOM index, Stagehand reference or screenshot point, according to the representation"], "counts": [], "evaluation": "READ uses normalized exact match against an answer or allowed alias. ACT resolves predicted identifiers to ground-truth elements, or checks screenshot points inside the target box. Representation-token accounting subtracts fixed-prompt tokens; screenshots use billed input usage minus the fixed prompt. These are answer/target correctness and token metrics, not completed-task success.", "evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md", "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"], "focus": ["Browser-representation token use", "Factual READ accuracy", "ACT target-element selection"], "notes": "Fixed local-page observations; the reported run uses claude-sonnet-5 and a shared prompt, without scrolling or an agent loop. Browser Use disables its viewport cutoff and sends untruncated text; Stagehand uses a separate matched browser load. Short-string representability labels are approximate. Offline does not mean API-free. Third-party snapshot redistribution remains unresolved; results were not reproduced.", "observations": ["Top-document HTML", "Markdown converted from HTML", "Playwright accessibility snapshot", "Browser Use indexed DOM", "Stagehand hybrid snapshot", "Viewport screenshot"], "setting": "offline"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"], "reason": "The inspected root LICENSE contains complete MIT terms for repository software. The methodology explicitly leaves redistribution permission for captured third-party practice-site pages unresolved; code MIT is not a blanket snapshot/data license."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:30.874397Z", "role": "primary", "sha256": "14775f6e3daba69d4684380fd33bf717bdc3093b65be796eb863bbd27d0c0bab", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:05:35.063004+00:00", "role": "primary", "sha256": "3e22e1cbfa47bcf98f6f70a5f6f162a31a509a9f2091a84679bb5231977e29d9", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Aweb-agents+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:54.545491Z", "role": "primary", "sha256": "fd30e94e78e7167a49579922259fde1c7d7f8259e9f00afba15faf3a71966a94", "supports": ["id", "url", "notes"], "url": "https://github.com/charlypoly/web-context-bench-python"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:30.131541Z", "role": "primary", "sha256": "1c3c0cbf9f9d9f7cd9e7a2f027ecb217fd5c16fa157916281a15bb7f403f6f1c", "supports": ["summary", "environments", "topics", "scope", "notes", "licenses.data", "details.benchmark_protocol.focus", "details.benchmark_protocol.setting", "details.benchmark_protocol.observations", "details.benchmark_protocol.actions", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:28.567861Z", "role": "primary", "sha256": "8cada5d71b5a61e7aff2e1ac2b951c266eb2db7c082c06c5a13aa127d42b3fc5", "supports": ["summary", "topics", "scope", "notes", "details.benchmark_protocol.evaluation", "details.benchmark_protocol.notes"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:41.499735Z", "role": "primary", "sha256": "9b64ef672f5091b6074d02502f6d3c161d94308a3d9d142d7437dcddf611bbc4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-20T11:51:30.874397Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:41.499735Z", "url": "https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/charlypoly/web-context-bench-python"}], "name": "Web Context Bench (Python)", "publication": null, "published_at": null, "record_sha256": "2c4ca722ae22d6032acc70cfef4a118d1c9fe353e4aac12be07f05e2be004f39", "relations": [], "resource_id": "github:charlypoly/web-context-bench-python", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/charlypoly/web-context-bench-python", "topics": ["evaluation", "grounding", "efficiency", "screen-understanding"], "updated_at": null}, "text": "Web Context Bench (Python)\nResource type: benchmark\nReview status: curated\nScope: core\n\nLocal-page benchmark comparing six browser representations on token use, factual reading and target-element selection, with deterministic answer scoring.\nPlatforms: web\nTopics: evaluation, grounding, efficiency, screen-understanding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"READ text answer\", \"ACT selector, text/label, accessible role/name, DOM index, Stagehand reference or screenshot point, according to the representation\"], \"counts\": [], \"evaluation\": \"READ uses normalized exact match against an answer or allowed alias. ACT resolves predicted identifiers to ground-truth elements, or checks screenshot points inside the target box. Representation-token accounting subtracts fixed-prompt tokens; screenshots use billed input usage minus the fixed prompt. These are answer/target correctness and token metrics, not completed-task success.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/README.md\", \"https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/METHODOLOGY.md\", \"https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/src/wpbench/scoring.py\"], \"focus\": [\"Browser-representation token use\", \"Factual READ accuracy\", \"ACT target-element selection\"], \"notes\": \"Fixed local-page observations; the reported run uses claude-sonnet-5 and a shared prompt, without scrolling or an agent loop. Browser Use disables its viewport cutoff and sends untruncated text; Stagehand uses a separate matched browser load. Short-string representability labels are approximate. Offline does not mean API-free. Third-party snapshot redistribution remains unresolved; results were not reproduced.\", \"observations\": [\"Top-document HTML\", \"Markdown converted from HTML\", \"Playwright accessibility snapshot\", \"Browser Use indexed DOM\", \"Stagehand hybrid snapshot\", \"Viewport screenshot\"], \"setting\": \"offline\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/charlypoly/web-context-bench-python/58d604bfaaac60ad42f02013f53f9fb301c1dcda/LICENSE\"], \"reason\": \"The inspected root LICENSE contains complete MIT terms for repository software. The methodology explicitly leaves redistribution permission for captured third-party practice-site pages unresolved; code MIT is not a blanket snapshot/data license.\"}}}\n\nThis is a single-snapshot representation evaluation on local fixtures and captured pages, not a live multi-step agent benchmark. Its shared-prompt, one-model setup differs from complete Browser Use or Stagehand agents. The methodology notes unresolved redistribution licensing for third-party practice-site snapshots; MIT is recorded for repository code only. Selected scoring code was read, but no models, browser tasks or claimed results were run or reproduced.\n\nPrimary source: https://github.com/charlypoly/web-context-bench-python\nCatalog record: https://awesomecua.github.io/resources/github-charlypoly-web-context-bench-python/"}
{"id": "github:chenghengli/windom", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ChenghengLi/WinDOM", "citation_url": "https://awesomecua.github.io/resources/github-chenghengli-windom/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"], "reason": "The inspected root LICENSE is complete MIT text for repository code. The local data README says same as parent repository, but the linked Hub dataset/model cards and artifacts were not inspected; their licenses are not inferred."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:34.292633Z", "role": "primary", "sha256": "7cda5207d8882bc8dedaefb93d2ccf54b310354917283b8f4890fffdc5764958", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "environments", "relations"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:50:58.795155Z", "role": "primary", "sha256": "08b89c5e6143afa4c0cdde95f8ec016d9033ee004f7631fadd4457144c2b87c9", "supports": ["id", "url", "notes"], "url": "https://github.com/ChenghengLi/WinDOM"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:41.278660Z", "role": "primary", "sha256": "09488140d35408cc11bb9c12b1bb435b21814cf6525937a8309475c9326c4004", "supports": ["environments", "notes", "licenses.data"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/data/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:45.091369Z", "role": "primary", "sha256": "0ee33b1e416ea4c5d8834688e6f42fbd15ef493f1b82398d3bb985838cbdffb9", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/training/sfd_4b.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:52:48.511963Z", "role": "primary", "sha256": "c481b2f91f2d82bf801632da6ab6cce01bc42c780104d743156c272942d2cef3", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/src/windom/evaluation/screenspot.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:44.940106Z", "role": "primary", "sha256": "d5a37ff86df9b231440ab1bcf3ecf77304a0a53e772182cdb622a3ab5d043447", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T11:51:34.292633Z", "licenses": {"code": {"basis": "Complete MIT license text in the repository root at the inspected commit; repository code only.", "declared": "MIT", "retrieved_at": "2026-10-20T11:51:44.940106Z", "url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChenghengLi/WinDOM"}], "name": "WinDOM code", "publication": null, "published_at": null, "record_sha256": "933c1c6328e3585fa63f620c0845cecd7049bbf9c51f6d8923b31f645097c9f3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/README.md", "target_id": "arxiv:2606.25964", "type": "documented_by"}], "resource_id": "github:chenghengli/windom", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ChenghengLi/WinDOM", "topics": ["grounding", "imitation-learning", "reinforcement-learning", "evaluation"], "updated_at": null}, "text": "WinDOM code\nResource type: framework\nReview status: curated\nScope: core\n\nGUI-grounding training and evaluation code with same-family distillation, GRPO recipes and point-in-box scoring, using a Windows-style web environment.\nPlatforms: web\nTopics: grounding, imitation-learning, reinforcement-learning, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/ChenghengLi/WinDOM/ab52acffe2003fd0c98943a542b4114ef929de0d/LICENSE\"], \"reason\": \"The inspected root LICENSE is complete MIT text for repository code. The local data README says same as parent repository, but the linked Hub dataset/model cards and artifacts were not inspected; their licenses are not inferred.\"}}}\n\nThe Windows-style environment is win11React in Playwright, not native Windows control. The repository documents its predecessor trajectory pipeline but explicitly omits the final corpus harvester. The inspected ScreenSpot evaluator defaults to a privileged ground-truth-center demonstration; --no-demo and --official-prompt select other modes, so evaluation results require their prompt and coordinate settings. Linked datasets, weights and reported scores were not independently verified; MIT applies to inspected repository code.\n\nPrimary source: https://github.com/ChenghengLi/WinDOM\nCatalog record: https://awesomecua.github.io/resources/github-chenghengli-windom/"}
{"id": "github:chenwu98/agent-attack", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ChenWu98/agent-attack", "citation_url": "https://awesomecua.github.io/resources/github-chenwu98-agent-attack/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "0759e1212025ac487c1646877f97526d35db1a62337192800acaff7684471087", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ChenWu98/agent-attack/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChenWu98/agent-attack"}], "name": "Adversarial Web Agents", "publication": null, "published_at": null, "record_sha256": "2f99660375e2368b49d456b2c54b93530794800fa7ee1f8f60ac92ae16e09a82", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "target_id": "arxiv:2406.12814", "type": "documented_by"}], "resource_id": "github:chenwu98/agent-attack", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ChenWu98/agent-attack", "topics": ["safety-security", "robustness"], "updated_at": null}, "text": "Adversarial Web Agents\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation resources for adversarial attacks on multimodal web agents.\nPlatforms: web\nTopics: safety-security, robustness\n\nPrimary source: https://github.com/ChenWu98/agent-attack\nCatalog record: https://awesomecua.github.io/resources/github-chenwu98-agent-attack/"}
{"id": "github:chessbean1/gui-robust", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/chessbean1/GUI-Robust", "citation_url": "https://awesomecua.github.io/resources/github-chessbean1-gui-robust/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "93defc2b31d0f8e6144e77cda4a74abeafd5253ae18a626bfc625f429050770a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/chessbean1/GUI-Robust/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/chessbean1/GUI-Robust"}], "name": "GUI-Robust", "publication": null, "published_at": null, "record_sha256": "d4a9b879a50d8e6bfe923e6d4af011119c769f12db4708b659c70eb64b3c270c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/chessbean1/GUI-Robust/HEAD/README.md", "target_id": "arxiv:2506.14477", "type": "documented_by"}], "resource_id": "github:chessbean1/gui-robust", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/chessbean1/GUI-Robust", "topics": ["evaluation", "robustness"], "updated_at": null}, "text": "GUI-Robust\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation scripts and linked data for comparing GUI action predictions and task execution under ordinary and anomalous interface conditions.\nPlatforms: cross-platform\nTopics: evaluation, robustness\n\nPrimary source: https://github.com/chessbean1/GUI-Robust\nCatalog record: https://awesomecua.github.io/resources/github-chessbean1-gui-robust/"}
{"id": "github:china-qijizhifeng/work2skill", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/china-qijizhifeng/Work2Skill", "citation_url": "https://awesomecua.github.io/resources/github-china-qijizhifeng-work2skill/", "details": {}, "environments": ["desktop", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "40a5e9ad3a2169527a9330825bd6e07ac1669293e21c209ed41036987bbf0a8b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "70387bc43f567b2ed971c1a789719c8c3567b0d073407c38b5dc6af1eb78012a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/china-qijizhifeng/Work2Skill"}], "name": "Work2Skill", "publication": null, "published_at": null, "record_sha256": "23269d103220fe55aa532071eba14f85cbaa079970d7b673fd4765b9d336e722", "relations": [{"evidence_url": "https://raw.githubusercontent.com/china-qijizhifeng/Work2Skill/HEAD/README.md", "target_id": "arxiv:2610.12984", "type": "documented_by"}], "resource_id": "github:china-qijizhifeng/work2skill", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/china-qijizhifeng/Work2Skill", "topics": ["recording", "workflow", "data-generation"], "updated_at": null}, "text": "Work2Skill\nResource type: framework\nReview status: curated\nScope: core\n\nCollects human GUI demonstrations and inspects multimodal evidence to derive reusable agent workflow skills.\nPlatforms: desktop, macos, linux\nTopics: recording, workflow, data-generation\n\nThe README reports that evaluation integration and live cross-platform acceptance testing remain pending.\n\nPrimary source: https://github.com/china-qijizhifeng/Work2Skill\nCatalog record: https://awesomecua.github.io/resources/github-china-qijizhifeng-work2skill/"}
{"id": "github:chromedevtools/chrome-devtools-mcp", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ChromeDevTools/chrome-devtools-mcp", "citation_url": "https://awesomecua.github.io/resources/github-chromedevtools-chrome-devtools-mcp/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "059a0349daaef6bda98eae6d9384ed349ac94b8033b18be817c317a01b472024", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/ChromeDevTools/chrome-devtools-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ChromeDevTools/chrome-devtools-mcp"}], "name": "Chrome DevTools MCP", "publication": null, "published_at": null, "record_sha256": "6efb3ba4619e092d42a10f7116e9ffb684b48d46cb010929fd5583ecf8e399da", "relations": [], "resource_id": "github:chromedevtools/chrome-devtools-mcp", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ChromeDevTools/chrome-devtools-mcp", "topics": ["automation", "tool-integration"], "updated_at": null}, "text": "Chrome DevTools MCP\nResource type: tool\nReview status: curated\nScope: core\n\nMCP tools for browser control and inspection through Chrome DevTools.\nPlatforms: web\nTopics: automation, tool-integration\n\nPrimary source: https://github.com/ChromeDevTools/chrome-devtools-mcp\nCatalog record: https://awesomecua.github.io/resources/github-chromedevtools-chrome-devtools-mcp/"}
{"id": "github:convergence-ai/webgames", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/convergence-ai/webgames", "citation_url": "https://awesomecua.github.io/resources/github-convergence-ai-webgames/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "a2df62f9b4359f7839c7d6c6917ff654f9eadeafa92f96311495b71e37abd9ba", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "0e03b68c7acb0b340e47f9000ac3808950e2ac8c4cd05960437a9e46e3c09a62", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/convergence-ai/webgames/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/convergence-ai/webgames"}], "name": "WebGames", "publication": null, "published_at": null, "record_sha256": "7081393908533691662f9adfa4ae8dd41546aa77b3254a0dee195edcc05912e8", "relations": [], "resource_id": "github:convergence-ai/webgames", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/convergence-ai/webgames", "topics": ["evaluation"], "updated_at": null}, "text": "WebGames\nResource type: benchmark\nReview status: curated\nScope: core\n\n150 client-side web challenges in five capability domains, built to be easy for humans but hard for browsing agents, each revealing a password when solved.\nPlatforms: web\nTopics: evaluation\n\nPrimary source: https://github.com/convergence-ai/webgames\nCatalog record: https://awesomecua.github.io/resources/github-convergence-ai-webgames/"}
{"id": "github:cooelf/auto-gui", "metadata": {"aliases": ["github:cooelf/auto-ui", "cooelf/Auto-UI", "https://github.com/cooelf/Auto-UI"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/cooelf/Auto-GUI", "citation_url": "https://awesomecua.github.io/resources/github-cooelf-auto-gui/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "13f045fa84dc0ad68dcceea95fc7e0e34a45bcda866d589132351559b87a8e30", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:59Z", "role": "primary", "sha256": "569f3c779e6a0b8210a764c9096e49e8a5199c107a64f7fcfb7a011257bf2a12", "supports": ["aliases"], "url": "https://api.github.com/repos/cooelf/Auto-UI"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:00Z", "role": "primary", "sha256": "569f3c779e6a0b8210a764c9096e49e8a5199c107a64f7fcfb7a011257bf2a12", "supports": ["aliases"], "url": "https://api.github.com/repos/cooelf/Auto-GUI"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/cooelf/Auto-GUI"}], "name": "Auto-GUI / Auto-UI", "publication": null, "published_at": null, "record_sha256": "8cba57d92fdbd3c7bb75b7546ad3bf16a3da02eb1dfb04d0210ece13beb805ed", "relations": [{"evidence_url": "https://raw.githubusercontent.com/cooelf/Auto-GUI/HEAD/README.md", "target_id": "arxiv:2309.11436", "type": "documented_by"}], "resource_id": "github:cooelf/auto-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/cooelf/Auto-GUI", "topics": ["mobile-navigation", "planning"], "updated_at": null}, "text": "Auto-GUI / Auto-UI\nResource type: agent\nReview status: curated\nScope: core\n\nScreenshot-based mobile agent implementation with training and inference code for chain-of-action prediction.\nRecorded aliases: github:cooelf/auto-ui, cooelf/Auto-UI, https://github.com/cooelf/Auto-UI\nPlatforms: android\nTopics: mobile-navigation, planning\n\nPrimary source: https://github.com/cooelf/Auto-GUI\nCatalog record: https://awesomecua.github.io/resources/github-cooelf-auto-gui/"}
{"id": "github:cursortouch/windows-mcp", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/CursorTouch/Windows-MCP", "citation_url": "https://awesomecua.github.io/resources/github-cursortouch-windows-mcp/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.8.8"}, "environments": ["windows", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "bbec20828c8a0f5483be68f6a00c47b202e9431268d96c1195fbf0745aca1364", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "ef43250ad8e57c3267f2a8e0aaeb655eb3445ade8d2dee8c243dc6e241a9d2d8", "supports": ["identity"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "652b580c99ac2de9d56d124225ba302f0dd83430ccc9b5162e16f683c56b952f", "supports": ["release_status"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "db3de978027106a991da3f7f2d78b216cfffee9b08604872fa3216927acc7a2d", "supports": ["release_status"], "url": "https://api.github.com/repos/CursorTouch/Windows-MCP/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "20ced59173768e20e91a43e3a80443aa1200b129c134f07f0cc45779747a1743", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:59Z", "role": "primary", "sha256": "20ced59173768e20e91a43e3a80443aa1200b129c134f07f0cc45779747a1743", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:00:59Z", "url": "https://raw.githubusercontent.com/CursorTouch/Windows-MCP/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/CursorTouch/Windows-MCP"}], "name": "Windows-MCP", "publication": null, "published_at": null, "record_sha256": "e42f1d6c8d2e9e0c2508ccdfb9e11ba54894bf292b7a76df6c9f9846405a3c41", "relations": [], "resource_id": "github:cursortouch/windows-mcp", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/CursorTouch/Windows-MCP", "topics": ["tool-integration", "automation", "screen-understanding"], "updated_at": null}, "text": "Windows-MCP\nResource type: tool\nReview status: curated\nScope: core\n\nA Windows MCP bridge exposing application, window, keyboard and mouse operations plus native UI-state inspection for agents.\nPlatforms: windows, desktop, web\nTopics: tool-integration, automation, screen-understanding\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"v0.8.8\"}\n\nPublic source and official release metadata are present. The README explicitly supports Windows 10 and 11 and includes a browser DOM inspection mode. Code is MIT-licensed. This is an execution tool, not a trained model; runtime behavior and latency claims were not independently tested.\n\nPrimary source: https://github.com/CursorTouch/Windows-MCP\nCatalog record: https://awesomecua.github.io/resources/github-cursortouch-windows-mcp/"}
{"id": "github:cvc2233/androtmem", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/CVC2233/AndroTMem", "citation_url": "https://awesomecua.github.io/resources/github-cvc2233-androtmem/", "details": {"availability": "Public inference/scoring source; benchmark data access remains gated.", "evaluation_caution": "Missing screenshots are marked AUTO_PASS and credited as correct according to the README.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"], "reason": "Repository code license only; no data-license or gated-access permission is inferred."}}, "repository_revision": "0c2da7002baa5b0ee71f80dd6fbd90e33071e259"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:21Z", "role": "primary", "sha256": "6681ce1bd98c856fed4193e239e234a69eb2a74fa937e9c38db89f48db40f452", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "f2b7c762383746a56e62273f756ca7ed3223726535763a41fda3ac5e3467aa82", "supports": ["identity"], "url": "https://api.github.com/repos/CVC2233/AndroTMem"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "302c804dc65ebb3e8558651e0c7cd0a79c1caf29fc9fb11868d26fba36afdf48", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/CVC2233/AndroTMem/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:00Z", "role": "primary", "sha256": "6d572ccef64746baac78441c0b26488f525c70c2b7453f985bd4f19cbec00634", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "e82623b7dc889eb7bff45c33f15a47c8c5caac3de0a3c56a7a9975ca9b2d6694", "supports": ["notes", "release_status"], "url": "https://huggingface.co/api/datasets/CVC2233/AndroTMem-Bench"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:21Z", "licenses": {"code": {"basis": "Repository code license only; no data-license or gated-access permission is inferred.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:00Z", "url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/CVC2233/AndroTMem"}], "name": "AndroTMem inference and evaluation code", "publication": null, "published_at": null, "record_sha256": "98eb537965f7f3280d1ddd5f1780f452ffc77e1fe7d8c0180dd2d2655e554290", "relations": [{"evidence_url": "https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/README.md", "target_id": "arxiv:2603.18429", "type": "documented_by"}], "resource_id": "github:cvc2233/androtmem", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/CVC2233/AndroTMem", "topics": ["memory", "long-horizon", "evaluation"], "updated_at": null}, "text": "AndroTMem inference and evaluation code\nResource type: framework\nReview status: curated\nScope: core\n\nA public Android GUI evaluation framework with action-history, summary and anchored-memory prompt modes, plus step-action and dependency-aware task scoring.\nPlatforms: android\nTopics: memory, long-horizon, evaluation\n\nRecorded specifications: {\"availability\": \"Public inference/scoring source; benchmark data access remains gated.\", \"evaluation_caution\": \"Missing screenshots are marked AUTO_PASS and credited as correct according to the README.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/CVC2233/AndroTMem/HEAD/LICENSE\"], \"reason\": \"Repository code license only; no data-license or gated-access permission is inferred.\"}}, \"repository_revision\": \"0c2da7002baa5b0ee71f80dd6fbd90e33071e259\"}\n\nThe source tree releases inference backends and evaluation scripts under evaluation/. The separately hosted benchmark is gated=auto and its raw card returned 401; no access request was made and it is not approved here as a directly accessible dataset. The README explicitly credits missing screenshots as AUTO_PASS/correct, so incomplete local image inventories can inflate scores. Its TCR is an offline anchor/dependency score and is not presented here as independently verified live task success.\n\nPrimary source: https://github.com/CVC2233/AndroTMem\nCatalog record: https://awesomecua.github.io/resources/github-cvc2233-androtmem/"}
{"id": "github:digirl-agent/digiq", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/DigiRL-agent/digiq", "citation_url": "https://awesomecua.github.io/resources/github-digirl-agent-digiq/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "33c55f7d1df7f812c9c9b9c4d97a4de7408b1b1916e58eed266b1f88389b50cd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digiq/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DigiRL-agent/digiq"}], "name": "Digi-Q", "publication": null, "published_at": null, "record_sha256": "0b933d93e9e794fe34809d400cd7adf68b2d48c04c3cad84b757bc1fa710522f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/DigiRL-agent/digiq/HEAD/README.md", "target_id": "arxiv:2502.15760", "type": "documented_by"}], "resource_id": "github:digirl-agent/digiq", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/DigiRL-agent/digiq", "topics": ["reinforcement-learning", "efficiency"], "updated_at": null}, "text": "Digi-Q\nResource type: framework\nReview status: curated\nScope: core\n\nOffline device-control training code that learns action values and uses them to select promising actions, with linked trajectories and checkpoints.\nPlatforms: android\nTopics: reinforcement-learning, efficiency\n\nPrimary source: https://github.com/DigiRL-agent/digiq\nCatalog record: https://awesomecua.github.io/resources/github-digirl-agent-digiq/"}
{"id": "github:digirl-agent/digirl", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/DigiRL-agent/digirl", "citation_url": "https://awesomecua.github.io/resources/github-digirl-agent-digirl/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "464e26d88420cd2694abe4c014bfd088417622c3a980aadf19b2bf01988d5b65", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DigiRL-agent/digirl"}], "name": "DigiRL", "publication": null, "published_at": null, "record_sha256": "631b962b72b8217563deffc634d214e54a0d1948cdc9df4ad4a954e7c4f46e10", "relations": [{"evidence_url": "https://raw.githubusercontent.com/DigiRL-agent/digirl/HEAD/README.md", "target_id": "arxiv:2406.11896", "type": "documented_by"}], "resource_id": "github:digirl-agent/digirl", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/DigiRL-agent/digirl", "topics": ["reinforcement-learning"], "updated_at": null}, "text": "DigiRL\nResource type: framework\nReview status: curated\nScope: core\n\nReinforcement learning for autonomous mobile device control.\nPlatforms: android\nTopics: reinforcement-learning\n\nPrimary source: https://github.com/DigiRL-agent/digirl\nCatalog record: https://awesomecua.github.io/resources/github-digirl-agent-digirl/"}
{"id": "github:distrl-lab/distrl-open", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/DistRL-lab/distrl-open", "citation_url": "https://awesomecua.github.io/resources/github-distrl-lab-distrl-open/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "f501f8de6aac53758f60882a8c595a5ac27145f327c73b0cfa7921b45ccae75a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/DistRL-lab/distrl-open"}], "name": "DistRL", "publication": null, "published_at": null, "record_sha256": "996d56b280d6d3741698fd1f59d81e61bdd872168c648da3f53a7205e055df72", "relations": [{"evidence_url": "https://raw.githubusercontent.com/DistRL-lab/distrl-open/HEAD/README.md", "target_id": "arxiv:2410.14803", "type": "documented_by"}], "resource_id": "github:distrl-lab/distrl-open", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/DistRL-lab/distrl-open", "topics": ["reinforcement-learning", "efficiency"], "updated_at": null}, "text": "DistRL\nResource type: framework\nReview status: curated\nScope: core\n\nDistributed online reinforcement-learning framework for Android agents, with environment setup and trajectory collection tools.\nPlatforms: android\nTopics: reinforcement-learning, efficiency\n\nThe README identifies model weights and dataset release as pending; this entry covers the code framework.\n\nPrimary source: https://github.com/DistRL-lab/distrl-open\nCatalog record: https://awesomecua.github.io/resources/github-distrl-lab-distrl-open/"}
{"id": "github:dmis-lab/switchmem", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/dmis-lab/SwitchMem", "citation_url": "https://awesomecua.github.io/resources/github-dmis-lab-switchmem/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "3fb5aabfb07df620955f0b74347f9d5926a1cd3d506b1098eccba9477627269f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/dmis-lab/SwitchMem/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "cec8e4679c883f13983915e9f77d175a84a234d69386e78f9ea75157cefbdcb4", "supports": ["relations"], "url": "https://txxnrd.github.io/switchmem/"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/dmis-lab/SwitchMem"}], "name": "SwitchMem", "publication": null, "published_at": null, "record_sha256": "7761dfe36cd9244e5726a879763043be279a9800241b1c756d30674b5693aa6b", "relations": [{"evidence_url": "https://txxnrd.github.io/switchmem/", "target_id": "arxiv:2610.17928", "type": "documented_by"}], "resource_id": "github:dmis-lab/switchmem", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/dmis-lab/SwitchMem", "topics": ["memory", "mobile-navigation"], "updated_at": null}, "text": "SwitchMem\nResource type: framework\nReview status: curated\nScope: core\n\nMemory-routing implementation that chooses text or image demonstrations for a frozen GUI actor, with offline evaluation and Android transfer experiments.\nPlatforms: android\nTopics: memory, mobile-navigation\n\nThe implementation distinguishes step-level AITW/LearnGUI scores from terminal task success on MobileMiniWoB++. Benchmark outcomes and local smoke-test outputs are different claims; neither was independently reproduced here.\n\nPrimary source: https://github.com/dmis-lab/SwitchMem\nCatalog record: https://awesomecua.github.io/resources/github-dmis-lab-switchmem/"}
{"id": "github:dongping-chen/gui-world", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Dongping-Chen/GUI-World", "citation_url": "https://awesomecua.github.io/resources/github-dongping-chen-gui-world/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "population": "dataset taxonomy", "qualifier": "exact", "unit": "GUI scenario categories", "value": 6}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "population": "dataset taxonomy", "qualifier": "exact", "unit": "question types", "value": 8}], "evidence_urls": ["https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md", "https://huggingface.co/datasets/ONE-Lab/GUI-World/raw/main/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"], "formats": [], "modalities": ["videos", "frame images", "text"], "notes": "Taxonomy counts describe coverage, not the number of videos or questions. Cases include Android, iOS, desktop software, websites, multi-application use, and XR. Keyframe-selection settings affect evaluation; exact sample inventories are not reported here.", "splits": ["train", "test"], "supervision": ["GUI video question answering", "temporal interface understanding"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "d02fa5da3384a3e58cf1d28aad47497de73fe3aad0ce57219585da04fc3187ef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://github.com/Dongping-Chen/GUI-World"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "905bf69ce80c866377841474cefe08f85ec4ef552cdd9ffeb2fb1531e07370b6", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "d2dd5886bce90f6eaee880719c79a6ede6ec0eaa08da32d7610bde656b8dc9ed", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/ONE-Lab/GUI-World/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Dongping-Chen/GUI-World"}], "name": "GUI-World", "publication": null, "published_at": null, "record_sha256": "5fe898c6d9e8e361e253db7697446f2ef5f24f2c7293fa99537a7f178d222f7f", "relations": [{"evidence_url": "https://github.com/Dongping-Chen/GUI-World", "target_id": "arxiv:2406.10819", "type": "documented_by"}], "resource_id": "github:dongping-chen/gui-world", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Dongping-Chen/GUI-World", "topics": ["video", "screen-understanding"], "updated_at": null}, "text": "GUI-World\nResource type: dataset\nReview status: curated\nScope: core\n\nGUI video benchmark and training data covering six scenarios, from Android and iOS to websites and XR, with eight question types for image and video LLMs and the GUI-Vid model.\nPlatforms: cross-platform\nTopics: video, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md\", \"population\": \"dataset taxonomy\", \"qualifier\": \"exact\", \"unit\": \"GUI scenario categories\", \"value\": 6}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md\", \"population\": \"dataset taxonomy\", \"qualifier\": \"exact\", \"unit\": \"question types\", \"value\": 8}], \"evidence_urls\": [\"https://raw.githubusercontent.com/Dongping-Chen/GUI-World/master/Readme.md\", \"https://huggingface.co/datasets/ONE-Lab/GUI-World/raw/main/README.md\", \"https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50\"], \"formats\": [], \"modalities\": [\"videos\", \"frame images\", \"text\"], \"notes\": \"Taxonomy counts describe coverage, not the number of videos or questions. Cases include Android, iOS, desktop software, websites, multi-application use, and XR. Keyframe-selection settings affect evaluation; exact sample inventories are not reported here.\", \"splits\": [\"train\", \"test\"], \"supervision\": [\"GUI video question answering\", \"temporal interface understanding\"]}}\n\nPrimary source: https://github.com/Dongping-Chen/GUI-World\nCatalog record: https://awesomecua.github.io/resources/github-dongping-chen-gui-world/"}
{"id": "github:droidrun/mobile-harness", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/droidrun/mobile-harness", "citation_url": "https://awesomecua.github.io/resources/github-droidrun-mobile-harness/", "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "0d05433c730e74f13be753fa6156737f797a69d61d8793e47f5a56b3a63ee790", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "de215a2914d50cb87fcab9e19efc856b9e600a68eefec5c63e284dae34977133", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/droidrun/mobile-harness/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobile-harness"}], "name": "Mobile Harness", "publication": null, "published_at": null, "record_sha256": "9ff8cf1be5c1fc92f2793a450aa467d74730d3450ab35e1896fb9bf6fab3e618", "relations": [], "resource_id": "github:droidrun/mobile-harness", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/droidrun/mobile-harness", "topics": ["automation", "tool-integration"], "updated_at": null}, "text": "Mobile Harness\nResource type: tool\nReview status: curated\nScope: core\n\nPortable agent instructions and integration tooling for local or cloud phone control through mobilerun-core.\nPlatforms: android, ios\nTopics: automation, tool-integration\n\nThe publisher describes this as a Markdown harness, not an agent runtime. Published benchmark results have not been independently reproduced.\n\nPrimary source: https://github.com/droidrun/mobile-harness\nCatalog record: https://awesomecua.github.io/resources/github-droidrun-mobile-harness/"}
{"id": "github:droidrun/mobilerun", "metadata": {"aliases": ["Droidrun"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/droidrun/mobilerun", "citation_url": "https://awesomecua.github.io/resources/github-droidrun-mobilerun/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.7.0"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "80052560e06db353277a55f3f8f6a4b5231aa6dc1376d5182559b72adcb5c3a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:56Z", "role": "primary", "sha256": "e48b9154a58b4b7021a4707ca185af32b9dd81c828d53ba36a86f195c1720344", "supports": ["identity"], "url": "https://api.github.com/repos/droidrun/mobilerun"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "7f40d42acd62b4405b5077a0f594a35aae73fb057589346afa5f6fb22d710a0c", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "f85c8b766edc107c455a86e363c56af3d1f7fce669b55377f4010982aa3a7e44", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "b3ebb19fda8286efa415d457529eafff41e1b1a27fe4e00d99f34e1559b6cdd9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "e48b9154a58b4b7021a4707ca185af32b9dd81c828d53ba36a86f195c1720344", "supports": ["identity", "aliases", "notes"], "url": "https://api.github.com/repos/droidrun/droidrun"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "b3ebb19fda8286efa415d457529eafff41e1b1a27fe4e00d99f34e1559b6cdd9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T04:00:57Z", "url": "https://raw.githubusercontent.com/droidrun/mobilerun/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobilerun"}], "name": "Mobilerun (formerly Droidrun)", "publication": null, "published_at": null, "record_sha256": "4b0a773b685699c44978c85e40acaca8a8b14e12425961caf10c8b11ea6a2b6f", "relations": [], "resource_id": "github:droidrun/mobilerun", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/droidrun/mobilerun", "topics": ["mobile-navigation", "tool-integration", "planning"], "updated_at": null}, "text": "Mobilerun (formerly Droidrun)\nResource type: framework\nReview status: curated\nScope: core\n\nA framework that combines screenshots and accessibility state with agent tools for mobile-device tasks through command-line and Python interfaces.\nRecorded aliases: Droidrun\nPlatforms: android, ios\nTopics: mobile-navigation, tool-integration, planning\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"v0.7.0\"}\n\nThe former droidrun/droidrun repository resolves to droidrun/mobilerun through the official API. Source and release v0.7.0 are public. Android and iOS setup paths differ, and the companion Portal app is a separate device bridge. MIT applies to framework code; model-provider and cloud-service terms are separate.\n\nPrimary source: https://github.com/droidrun/mobilerun\nCatalog record: https://awesomecua.github.io/resources/github-droidrun-mobilerun/"}
{"id": "github:droidrun/mobilerun-portal", "metadata": {"aliases": ["Droidrun Portal"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/droidrun/mobilerun-portal", "citation_url": "https://awesomecua.github.io/resources/github-droidrun-mobilerun-portal/", "details": {"availability": "Published source; official release metadata inspected", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"], "reason": "GitHub classifies this license as NOASSERTION, but the actual program-specific notice is explicit. Do not reduce the permission to a generic unknown or silently transfer the framework MIT declaration."}}, "release_tag": "v0.7.25"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fd2257dd735dbb755f2de36b44e425e5862004a04130e53e92b330f1419479a0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:56Z", "role": "primary", "sha256": "be519be28e277bd979717996e9587879c6a5d9235b95944305252b18595da413", "supports": ["identity"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "65e81cfc9f227350e33da9e09030a7098e82a6021aa9b04cf0485ed2acee95a0", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "31b8d75541bbfb7dba2f80adc83aa099ef83ca057b467eb2cb98b586f6d9d045", "supports": ["release_status"], "url": "https://api.github.com/repos/droidrun/mobilerun-portal/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:01Z", "role": "primary", "sha256": "945331c1cfa589694eb6ccc51574a72ee215a01b10bd5556f1b8be8d11bd71ab", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "be519be28e277bd979717996e9587879c6a5d9235b95944305252b18595da413", "supports": ["identity", "aliases", "notes"], "url": "https://api.github.com/repos/droidrun/droidrun-portal"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "The program-specific copyright notice explicitly permits GNU AGPL version 3 or any later version.", "declared": "AGPL-3.0-or-later", "retrieved_at": "2026-10-17T04:01:01Z", "url": "https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/droidrun/mobilerun-portal"}], "name": "Mobilerun Portal (formerly Droidrun Portal)", "publication": null, "published_at": null, "record_sha256": "39415dfe2f4c33e3753f4e9c532daec24ec5a6fab30a982f4ac640948f134888", "relations": [], "resource_id": "github:droidrun/mobilerun-portal", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/droidrun/mobilerun-portal", "topics": ["mobile-navigation", "screen-understanding", "tool-integration"], "updated_at": null}, "text": "Mobilerun Portal (formerly Droidrun Portal)\nResource type: tool\nReview status: curated\nScope: core\n\nAn Android accessibility service that exposes UI-element inspection, device-control interfaces and screen feedback to mobile automation agents.\nRecorded aliases: Droidrun Portal\nPlatforms: android\nTopics: mobile-navigation, screen-understanding, tool-integration\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/droidrun/mobilerun-portal/HEAD/LICENSE\"], \"reason\": \"GitHub classifies this license as NOASSERTION, but the actual program-specific notice is explicit. Do not reduce the permission to a generic unknown or silently transfer the framework MIT declaration.\"}}, \"release_tag\": \"v0.7.25\"}\n\nThe official API resolves the former Droidrun Portal path to mobilerun-portal. Release metadata lists a v0.7.25 Android APK. Although GitHub labels its license NOASSERTION, the inspected LICENSE explicitly grants AGPL version 3 or later for the program. Accessibility-dependent UI functions require the service; no permissions were granted or device operated.\n\nPrimary source: https://github.com/droidrun/mobilerun-portal\nCatalog record: https://awesomecua.github.io/resources/github-droidrun-mobilerun-portal/"}
{"id": "github:e2b-dev/desktop", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/e2b-dev/desktop", "citation_url": "https://awesomecua.github.io/resources/github-e2b-dev-desktop/", "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "edb4293cb510888cdc2d5c9d1135a0500272aae96366710744d5b74b5289c462", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/e2b-dev/desktop"}], "name": "E2B Desktop", "publication": null, "published_at": null, "record_sha256": "844a78f1eed0054012a1c772d2c4088ab176788409ff6757fac84b6abc124d10", "relations": [{"evidence_url": "https://raw.githubusercontent.com/e2b-dev/desktop/HEAD/README.md", "target_id": "organization:e2b", "type": "published_by"}], "resource_id": "github:e2b-dev/desktop", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/e2b-dev/desktop", "topics": ["sandbox"], "updated_at": null}, "text": "E2B Desktop\nResource type: infrastructure\nReview status: curated\nScope: core\n\nOpen-source, isolated Linux Xfce desktop sandboxes for computer-use agents, driven from Python or JavaScript SDKs with screen streaming, mouse, keyboard and window control, screenshots and shell.\nPlatforms: desktop, linux\nTopics: sandbox\n\nPrimary source: https://github.com/e2b-dev/desktop\nCatalog record: https://awesomecua.github.io/resources/github-e2b-dev-desktop/"}
{"id": "github:emergenceai/agent-e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/EmergenceAI/Agent-E", "citation_url": "https://awesomecua.github.io/resources/github-emergenceai-agent-e/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "fc3099072ca60acf2294a63dd2c3e343e90c901a7c7581f723cfdccc5a1edc2b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "555ac6a8bf7ef4a3c82f4ee258be9d751042f8d2a6c901e5eac7c43596563c75", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/LICENSE.txt"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/EmergenceAI/Agent-E"}], "name": "Agent E", "publication": null, "published_at": null, "record_sha256": "1fc45eef719ade84927ec2255e61cd0265df931c0450d2219e0e73cdc6a336bf", "relations": [{"evidence_url": "https://raw.githubusercontent.com/EmergenceAI/Agent-E/HEAD/README.md", "target_id": "arxiv:2407.13032", "type": "documented_by"}], "resource_id": "github:emergenceai/agent-e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/EmergenceAI/Agent-E", "topics": ["web-navigation"], "updated_at": null}, "text": "Agent E\nResource type: agent\nReview status: curated\nScope: core\n\nBrowser automation agent on the AG2 framework that pairs a planner agent with a browser navigation agent to fill forms, search shops and work through sites, also callable over a local HTTP API.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/EmergenceAI/Agent-E\nCatalog record: https://awesomecua.github.io/resources/github-emergenceai-agent-e/"}
{"id": "github:euphoria16/ui-genie", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Euphoria16/UI-Genie", "citation_url": "https://awesomecua.github.io/resources/github-euphoria16-ui-genie/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md", "https://api.github.com/repos/Euphoria16/UI-Genie/git/trees/main?recursive=1"], "reason": "The README declares MIT and links a root LICENSE, but the complete retained tree has no root LICENSE; bundled ms-swift has its own license. Preserve this unresolved scope instead of treating a component license as the whole project license."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "a53e4cd3b4391c03c9334745152f8c1f66aff7590051f9bdb7e78b81c6c99a24", "supports": ["identity"], "url": "https://api.github.com/repos/Euphoria16/UI-Genie"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:55Z", "role": "primary", "sha256": "c27c6d88d043c964776c04a74517d48644418dca2f5ae9c8ba778d082c12a6ae", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/Euphoria16/UI-Genie/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Euphoria16/UI-Genie"}], "name": "UI-Genie", "publication": null, "published_at": null, "record_sha256": "e9ad9c113e875029ad05a8a71f9107bbd20e3cfe1e6fadc357abd80b0393883a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "resource_id": "github:euphoria16/ui-genie", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Euphoria16/UI-Genie", "topics": ["data-generation", "reward-models", "imitation-learning"], "updated_at": null}, "text": "UI-Genie\nResource type: framework\nReview status: curated\nScope: core\n\nTraining and evaluation code for a mobile-agent framework that iteratively generates trajectories using agent and reward-model feedback.\nPlatforms: android\nTopics: data-generation, reward-models, imitation-learning\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md\", \"https://api.github.com/repos/Euphoria16/UI-Genie/git/trees/main?recursive=1\"], \"reason\": \"The README declares MIT and links a root LICENSE, but the complete retained tree has no root LICENSE; bundled ms-swift has its own license. Preserve this unresolved scope instead of treating a component license as the whole project license.\"}}}\n\nThe tree contains ms-swift-based training and AndroidControl evaluation scripts. The README says AndroidLab/Android Arena scripts and the 72B checkpoint are still forthcoming; these are not claimed released. It declares MIT but links a root LICENSE absent from the inspected tree, so the exact code-license scope remains unresolved. Separate Hub weights and data have their own declarations.\n\nPrimary source: https://github.com/Euphoria16/UI-Genie\nCatalog record: https://awesomecua.github.io/resources/github-euphoria16-ui-genie/"}
{"id": "github:facebookresearch/ai-agent-privacy", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/facebookresearch/ai-agent-privacy", "citation_url": "https://awesomecua.github.io/resources/github-facebookresearch-ai-agent-privacy/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "9c938f8f3f3ee04f459362a8b7d399588fd030ce476cda111e5f23fbf1e31f10", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/facebookresearch/ai-agent-privacy/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/facebookresearch/ai-agent-privacy"}], "name": "AgentDAM", "publication": null, "published_at": null, "record_sha256": "225b7e80c42f8f0d27666b224a919ff296536a2755cdaaca29cd4f1de42a04cb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/facebookresearch/ai-agent-privacy/HEAD/README.md", "target_id": "arxiv:2503.09780", "type": "documented_by"}], "resource_id": "github:facebookresearch/ai-agent-privacy", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/facebookresearch/ai-agent-privacy", "topics": ["safety-security", "evaluation"], "updated_at": null}, "text": "AgentDAM\nResource type: benchmark\nReview status: curated\nScope: core\n\nWeb-agent benchmark for measuring unnecessary use of sensitive personal information during end-to-end task execution.\nPlatforms: web\nTopics: safety-security, evaluation\n\nPrimary source: https://github.com/facebookresearch/ai-agent-privacy\nCatalog record: https://awesomecua.github.io/resources/github-facebookresearch-ai-agent-privacy/"}
{"id": "github:facebookresearch/openapps", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/facebookresearch/OpenApps", "citation_url": "https://awesomecua.github.io/resources/github-facebookresearch-openapps/", "details": {"benchmark_protocol": {"actions": ["click", "type", "scroll", "configured BrowserGym/AgentLab actions"], "counts": [], "evaluation": "Tasks define check_if_task_is_complete(initial_state,current_state); app-specific reward logic evaluates underlying state, with custom checks for code-editor file contents. Observation/history flags are configurable. No universal single fixed task total or published-run reproduction is implied.", "evidence_urls": ["https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md", "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/tasks.md", "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/agents.md", "https://api.github.com/repos/facebookresearch/OpenApps/git/trees/main?recursive=1"], "focus": ["UI-agent reliability under environment variation", "state-based browser task evaluation"], "notes": "Preserve the code/tag, app state, theme/layout/content choices, goal variant and action/observation configuration. Current main refactors the paper-era appearance surface, and the v1.0-paper tag includes post-publication fixes that may move rewards. The environments run as browser apps; phone/tablet viewports are not native Android/iOS environments.", "observations": ["screenshots", "accessibility trees", "set-of-marks and coordinates (configurable)"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "6e6782794433c60468d65ac75bc3f03c56007f8c", "repository_url": "https://github.com/facebookresearch/OpenApps"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:18Z", "role": "primary", "sha256": "1df0f66584b37e7dba90811a7732d3b0c13c96cd7da4aec3970a18be36a6cec1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:19Z", "role": "primary", "sha256": "2c57b627cb344e360480c10e74dcdd1d1ed321d9a7a4e3891bc970355b15f4cd", "supports": ["identity"], "url": "https://api.github.com/repos/facebookresearch/OpenApps"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:18Z", "role": "primary", "sha256": "34aa76fbd6cca6ac2f3949792144662b90d85d7a18988afabcb570167ea5c1d6", "supports": ["identity", "release_status", "details.repository_tree_sha", "details.benchmark_protocol"], "url": "https://api.github.com/repos/facebookresearch/OpenApps/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:40Z", "role": "primary", "sha256": "af030b93a7604c063b08f003b0a0124a6e95e560649d0306bfe40dd87fed563d", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/tasks.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:40Z", "role": "primary", "sha256": "aede4dfd965bd3484336f40ddb1bd6db48aed00ab45e27c048b76cc4601d71b3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/agents.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "336255dc30193e8e15d689d9481bb05673d89055718f3a96923a7ffb99adbbaf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:35:18Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "CC-BY-NC-4.0", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/facebookresearch/OpenApps"}], "name": "OpenApps", "publication": null, "published_at": null, "record_sha256": "9dde338ecd07e480e3312d0f26a5c805fbbbad16447905517a358c8e20d01986", "relations": [{"evidence_url": "https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md", "target_id": "arxiv:2511.20766", "type": "documented_by"}], "resource_id": "github:facebookresearch/openapps", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/facebookresearch/OpenApps", "topics": ["evaluation", "robustness", "cross-app", "sandbox"], "updated_at": null}, "text": "OpenApps\nResource type: benchmark\nReview status: curated\nScope: core\n\nConfigurable browser apps and task rewards for testing computer-use agents under changes in content, layout and appearance.\nPlatforms: web\nTopics: evaluation, robustness, cross-app, sandbox\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"click\", \"type\", \"scroll\", \"configured BrowserGym/AgentLab actions\"], \"counts\": [], \"evaluation\": \"Tasks define check_if_task_is_complete(initial_state,current_state); app-specific reward logic evaluates underlying state, with custom checks for code-editor file contents. Observation/history flags are configurable. No universal single fixed task total or published-run reproduction is implied.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/README.md\", \"https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/tasks.md\", \"https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/docs/agents.md\", \"https://api.github.com/repos/facebookresearch/OpenApps/git/trees/main?recursive=1\"], \"focus\": [\"UI-agent reliability under environment variation\", \"state-based browser task evaluation\"], \"notes\": \"Preserve the code/tag, app state, theme/layout/content choices, goal variant and action/observation configuration. Current main refactors the paper-era appearance surface, and the v1.0-paper tag includes post-publication fixes that may move rewards. The environments run as browser apps; phone/tablet viewports are not native Android/iOS environments.\", \"observations\": [\"screenshots\", \"accessibility trees\", \"set-of-marks and coordinates (configurable)\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/facebookresearch/OpenApps/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"6e6782794433c60468d65ac75bc3f03c56007f8c\", \"repository_url\": \"https://github.com/facebookresearch/OpenApps\"}\n\nPublic code includes app/task/configuration and agent-launching implementations. README distinguishes moving main from v1.0-paper and explicitly says even that tag is not a byte-exact snapshot of the published runs. CC-BY-NC-4.0 is the repository’s declared code license; third-party WebShop, map and icon content retain separate terms. No commercial-use permission or fixed downloadable task total is inferred.\n\nPrimary source: https://github.com/facebookresearch/OpenApps\nCatalog record: https://awesomecua.github.io/resources/github-facebookresearch-openapps/"}
{"id": "github:fanbinlu/steve-r1", "metadata": {"aliases": ["STEVE", "STEVE-R1"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/FanbinLu/STEVE-R1", "citation_url": "https://awesomecua.github.io/resources/github-fanbinlu-steve-r1/", "details": {}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "52aa383fdf5d91951f5d9a73f4ebed4454a82b35edae296918116e9f0895dd83", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/FanbinLu/STEVE/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["relations", "environments"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/FanbinLu/STEVE-R1"}], "name": "STEVE / STEVE-R1 research code", "publication": null, "published_at": null, "record_sha256": "fa70fc830c08619d6d4e9835cf4ee18b2c62e9123c937d7e524151c250493e97", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2503.12532", "type": "documented_by"}], "resource_id": "github:fanbinlu/steve-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/FanbinLu/STEVE-R1", "topics": ["data-generation", "automation", "evaluation"], "updated_at": null}, "text": "STEVE / STEVE-R1 research code\nResource type: framework\nReview status: curated\nScope: core\n\nResearch code and linked models for step-verified computer-use training; the current repository describes STEVE-R1 and its trajectory releases.\nRecorded aliases: STEVE, STEVE-R1\nPlatforms: desktop, windows, linux\nTopics: data-generation, automation, evaluation\n\nThe repository has moved to STEVE-R1. Its README reports an SFT release, with RL tuning in progress, and describes evaluations using multiple attempts per task. These results are not independently reproduced here.\n\nPrimary source: https://github.com/FanbinLu/STEVE-R1\nCatalog record: https://awesomecua.github.io/resources/github-fanbinlu-steve-r1/"}
{"id": "github:farama-foundation/miniwob-plusplus", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Farama-Foundation/miniwob-plusplus", "citation_url": "https://awesomecua.github.io/resources/github-farama-foundation-miniwob-plusplus/", "details": {"benchmark_protocol": {"actions": ["configurable coordinate actions", "element-reference actions", "keyboard input"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "population": "README collection size", "qualifier": "more-than", "unit": "web environments", "value": 100}], "evaluation": "", "evidence_urls": ["https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "https://miniwob.farama.org/content/observation_space/", "https://miniwob.farama.org/content/action_space/"], "focus": ["small web interaction tasks"], "notes": "The README gives a lower bound, not an exact active-environment inventory. Action presets expose different capabilities, so comparisons must identify the preset. Environment-specific reward details are not normalized here. The repository is in maintenance mode.", "observations": ["task instructions", "screenshots", "visible DOM-element attributes"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "90499776fa1b23e08ac85190ad3be709caa9172ff4f2a0a106b0fa350568e86f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "90499776fa1b23e08ac85190ad3be709caa9172ff4f2a0a106b0fa350568e86f", "supports": ["details.benchmark_protocol", "relations", "notes"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "511039e0e3c26b117e7e0a001746c26ec6d94a8ab8c467f0be0bc1a0b01e870a", "supports": ["details.benchmark_protocol"], "url": "https://miniwob.farama.org/content/observation_space/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "ae7bf6a3b98b198b985d7ca7b41cb2bdab3cf48e6445467ad9a8d4a2c8df240f", "supports": ["details.benchmark_protocol"], "url": "https://miniwob.farama.org/content/action_space/"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "fd6529fb8f648d4130fd969e3d7fc740c9992a0bd94a1998f52555a873311511", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Farama-Foundation/miniwob-plusplus"}], "name": "MiniWoB++", "publication": null, "published_at": null, "record_sha256": "d90cfb3dfd8983799fd9f364114e8bbe3c5828a2be07cfa15650d610fd51df74", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md", "target_id": "arxiv:1802.08802", "type": "introduced_by"}], "resource_id": "github:farama-foundation/miniwob-plusplus", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Farama-Foundation/miniwob-plusplus", "topics": ["evaluation", "reinforcement-learning"], "updated_at": null}, "text": "MiniWoB++\nResource type: benchmark\nReview status: curated\nScope: core\n\nOver 100 web-interaction environments extending OpenAI's MiniWoB, controlled from Python through a Gymnasium API that drives the browser with Selenium.\nPlatforms: web\nTopics: evaluation, reinforcement-learning\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"configurable coordinate actions\", \"element-reference actions\", \"keyboard input\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md\", \"population\": \"README collection size\", \"qualifier\": \"more-than\", \"unit\": \"web environments\", \"value\": 100}], \"evaluation\": \"\", \"evidence_urls\": [\"https://raw.githubusercontent.com/Farama-Foundation/miniwob-plusplus/HEAD/README.md\", \"https://miniwob.farama.org/content/observation_space/\", \"https://miniwob.farama.org/content/action_space/\"], \"focus\": [\"small web interaction tasks\"], \"notes\": \"The README gives a lower bound, not an exact active-environment inventory. Action presets expose different capabilities, so comparisons must identify the preset. Environment-specific reward details are not normalized here. The repository is in maintenance mode.\", \"observations\": [\"task instructions\", \"screenshots\", \"visible DOM-element attributes\"], \"setting\": \"interactive\"}}\n\nThe maintainer README identifies MiniWoB++ as the extension introduced in the 2018 workflow-guided exploration paper. It separately links World of Bits as the source for the original MiniWoB predecessor. The current Farama interface, environment inventory, and maintenance status belong to the retained repository snapshot.\n\nPrimary source: https://github.com/Farama-Foundation/miniwob-plusplus\nCatalog record: https://awesomecua.github.io/resources/github-farama-foundation-miniwob-plusplus/"}
{"id": "github:fclab-skku/mobilegpt", "metadata": {"aliases": ["github:mobilegptsys/mobilegpt", "mobilegptsys/MobileGPT", "https://github.com/mobilegptsys/MobileGPT"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/fclab-skku/MobileGPT", "citation_url": "https://awesomecua.github.io/resources/github-fclab-skku-mobilegpt/", "details": {"availability": "Public Android client, Python server and manual offline-explorer source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"], "reason": "Repository source license; linked benchmark data and external models retain separate terms."}}, "repository_revision": "7945048e6c05849aff1633744cbdf653c58bece2"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "8683cff296d21c1a7a38bc372efaf004b8af644030fac77a4dc9d2b2e21d894b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["identity"], "url": "https://api.github.com/repos/mobilegptsys/MobileGPT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "2a8f30b3422a033b345c2054d55adff0836b7dac163e600583ba0db1b4423320", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/fclab-skku/MobileGPT/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:04Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["aliases"], "url": "https://api.github.com/repos/mobilegptsys/MobileGPT"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:06Z", "role": "primary", "sha256": "3d3ceb084d374b706842cff1f4ec986fde50c13cba141dc4619d5afe8b631d55", "supports": ["aliases"], "url": "https://api.github.com/repos/fclab-skku/MobileGPT"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": {"basis": "Repository source license; linked benchmark data and external models retain separate terms.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/fclab-skku/MobileGPT"}], "name": "MobileGPT", "publication": null, "published_at": null, "record_sha256": "40625f73c2c547577c3088401c8190d12ead80b7ea52b08ee7ec707a41886157", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "target_id": "arxiv:2312.03003", "type": "documented_by"}], "resource_id": "github:fclab-skku/mobilegpt", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/fclab-skku/MobileGPT", "topics": ["mobile-navigation", "memory", "planning", "workflow"], "updated_at": null}, "text": "MobileGPT\nResource type: agent\nReview status: curated\nScope: core\n\nAn Android client and Python-server agent that learns reusable app subtasks through exploration, selection, derivation and memory recall.\nRecorded aliases: github:mobilegptsys/mobilegpt, mobilegptsys/MobileGPT, https://github.com/mobilegptsys/MobileGPT\nPlatforms: android\nTopics: mobile-navigation, memory, planning, workflow\n\nRecorded specifications: {\"availability\": \"Public Android client, Python server and manual offline-explorer source.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/fclab-skku/MobileGPT/HEAD/LICENSE\"], \"reason\": \"Repository source license; linked benchmark data and external models retain separate terms.\"}}, \"repository_revision\": \"7945048e6c05849aff1633744cbdf653c58bece2\"}\n\nThe old mobilegptsys repository resolves to fclab-skku/MobileGPT. Public Android client and Python server sources require Android SDK 33+, Accessibility Service setup and external service credentials. The README explicitly omits the random explorer and human-in-the-loop memory repair from the open-source version. Its benchmark description differs from the retained arXiv snapshot, so no dataset counts or reproduction claim are merged into this code record. External benchmark data was not approved as a separate release.\n\nPrimary source: https://github.com/fclab-skku/MobileGPT\nCatalog record: https://awesomecua.github.io/resources/github-fclab-skku-mobilegpt/"}
{"id": "github:friedrichor/webtestbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/friedrichor/WebTestBench", "citation_url": "https://awesomecua.github.io/resources/github-friedrichor-webtestbench/", "details": {"availability": "Public evaluation harness and ungated web-application archive manifest.", "benchmark_protocol": {"actions": ["browser interaction", "test checklist generation", "defect reports"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "application categories", "value": 7}, {"basis": "computed", "evidence_url": "https://huggingface.co/api/datasets/friedrichor/WebTestBench", "population": "Hub file manifest", "qualifier": "exact", "unit": "web-application archives", "value": 100}], "evaluation": "The supplied WebTester pipeline separates checklist generation from interactive defect detection and provides a scoring workflow.", "evidence_urls": ["https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "https://huggingface.co/api/datasets/friedrichor/WebTestBench", "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"], "focus": ["web functionality testing", "latent permission and business constraints"], "notes": "The 100 archive files are not counted as 100 independent test tasks or defects. Backend/tool versions and deployment settings affect reproducibility.", "observations": ["web application interface", "development instruction"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"], "reason": "Repository source LICENSE; external tools and services retain their own licenses."}, "data": {"evidence_urls": ["https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"], "reason": "Publisher Hub-card license declaration for the released benchmark dataset."}}, "repository_revision": "04873065bdb0938ed53d809559d4ab473c7ccd31"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "b41472026cd7d4f81a01d58bf39193dc1fabc7875c61e370090a20cc77500cdd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "7203ecc2aae5f1697eeeb804ca8f21d47b9f22a8b8fa43a4064c4e1c49f5e72c", "supports": ["identity"], "url": "https://api.github.com/repos/friedrichor/WebTestBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "bd1f09149b00d59a0078d414216bd2fb78e09ce7212d15bbda84ddec7052c8af", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/friedrichor/WebTestBench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "db1e97f40107038ae03835dc20f17b13b37e74b5c8587f8c39b441271618f017", "supports": ["identity", "release_status", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/friedrichor/WebTestBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "6f2f94a25ed66a14779f6b4e6686044d0510eb32a0187645cc473978d93f7c5b", "supports": ["licenses.data", "details.benchmark_protocol"], "url": "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:18Z", "licenses": {"code": {"basis": "Repository source LICENSE; external tools and services retain their own licenses.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:55Z", "url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE"}, "data": {"basis": "Publisher Hub-card license declaration for the released benchmark dataset.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:08Z", "url": "https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/friedrichor/WebTestBench"}], "name": "WebTestBench", "publication": null, "published_at": null, "record_sha256": "7df910e4fbfb7f371ec81f2e4341863c66536ab2dd0030ccdccdd0ffa8705169", "relations": [{"evidence_url": "https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md", "target_id": "arxiv:2603.25226", "type": "documented_by"}], "resource_id": "github:friedrichor/webtestbench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/friedrichor/WebTestBench", "topics": ["evaluation", "web-navigation", "tool-integration"], "updated_at": null}, "text": "WebTestBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nAn end-to-end web-testing benchmark where agents derive test checklists from development instructions, interact with deployed applications and report functional or constraint-related defects.\nPlatforms: web\nTopics: evaluation, web-navigation, tool-integration\n\nRecorded specifications: {\"availability\": \"Public evaluation harness and ungated web-application archive manifest.\", \"benchmark_protocol\": {\"actions\": [\"browser interaction\", \"test checklist generation\", \"defect reports\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md\", \"population\": \"publisher benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"application categories\", \"value\": 7}, {\"basis\": \"computed\", \"evidence_url\": \"https://huggingface.co/api/datasets/friedrichor/WebTestBench\", \"population\": \"Hub file manifest\", \"qualifier\": \"exact\", \"unit\": \"web-application archives\", \"value\": 100}], \"evaluation\": \"The supplied WebTester pipeline separates checklist generation from interactive defect detection and provides a scoring workflow.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/README.md\", \"https://huggingface.co/api/datasets/friedrichor/WebTestBench\", \"https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md\"], \"focus\": [\"web functionality testing\", \"latent permission and business constraints\"], \"notes\": \"The 100 archive files are not counted as 100 independent test tasks or defects. Backend/tool versions and deployment settings affect reproducibility.\", \"observations\": [\"web application interface\", \"development instruction\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/friedrichor/WebTestBench/HEAD/LICENSE\"], \"reason\": \"Repository source LICENSE; external tools and services retain their own licenses.\"}, \"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/friedrichor/WebTestBench/raw/main/README.md\"], \"reason\": \"Publisher Hub-card license declaration for the released benchmark dataset.\"}}, \"repository_revision\": \"04873065bdb0938ed53d809559d4ab473c7ccd31\"}\n\nPublic evaluation/deployment scripts accompany an ungated Hub release with JSON/JSONL metadata and 100 application ZIP files. The ZIP count is an artifact inventory, not an asserted test-case or defect count. The documented baseline uses Claude Agent SDK and a pinned Playwright MCP version with external model credentials; these were not installed or run. Code and data Apache-2.0 declarations are retained separately.\n\nPrimary source: https://github.com/friedrichor/WebTestBench\nCatalog record: https://awesomecua.github.io/resources/github-friedrichor-webtestbench/"}
{"id": "github:gair-nlp/pc-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/GAIR-NLP/PC-Agent", "citation_url": "https://awesomecua.github.io/resources/github-gair-nlp-pc-agent/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "178774927f003b5221d695fa78ded6fcc543aa8653c7b2dd8016418e6911afb7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "aee22091f60f10968afb7dc33dd8b728ea929e287c260ed82adffdec4866fcb5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/PC-Agent"}], "name": "PC-Agent", "publication": null, "published_at": null, "record_sha256": "edbb77c7693978a3a9d090e0916cb2f0bf469f5df7ba31d477349f177400b059", "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent/HEAD/README.md", "target_id": "arxiv:2412.17589", "type": "documented_by"}], "resource_id": "github:gair-nlp/pc-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/GAIR-NLP/PC-Agent", "topics": ["planning", "cross-app"], "updated_at": null}, "text": "PC-Agent\nResource type: agent\nReview status: curated\nScope: core\n\nHuman-cognition-transfer framework for PC agents: PC Tracker logs human computer use, cognition completion turns logs into cognitive trajectories, and a planner works with a grounding agent.\nPlatforms: desktop\nTopics: planning, cross-app\n\nPrimary source: https://github.com/GAIR-NLP/PC-Agent\nCatalog record: https://awesomecua.github.io/resources/github-gair-nlp-pc-agent/"}
{"id": "github:gair-nlp/pc-agent-e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/GAIR-NLP/PC-Agent-E", "citation_url": "https://awesomecua.github.io/resources/github-gair-nlp-pc-agent-e/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "9b1b00b907b99a4a8dbfc9c01ea49010878dfbd6", "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "30761141ea62c5f20183f359bbc3fc8ccc5b458886d38c5426bae4802a23ee0b", "supports": ["identity"], "url": "https://api.github.com/repos/GAIR-NLP/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:23Z", "role": "primary", "sha256": "2fb2aa4ee6f02ccadb631dab4972f96c3fdf545cd92c5854fc191e81cf4ae88a", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/GAIR-NLP/PC-Agent-E/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "39e37b84c067596361c491e2e9e22af8805386a24b8352b87e9aca3bddf4f92c", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/PC-Agent-E"}], "name": "PC-Agent-E training and deployment", "publication": null, "published_at": null, "record_sha256": "db1e55ccf38bfe2bc8464b58714b5b753345cca450384b43dfe7d813a4793fa2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "resource_id": "github:gair-nlp/pc-agent-e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/GAIR-NLP/PC-Agent-E", "topics": ["data-generation", "imitation-learning", "automation"], "updated_at": null}, "text": "PC-Agent-E training and deployment\nResource type: framework\nReview status: curated\nScope: core\n\nTools for refining and augmenting human computer-use trajectories, preparing supervised training data, and deploying a native PC-Agent-E scaffold.\nPlatforms: desktop, windows, linux\nTopics: data-generation, imitation-learning, automation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"9b1b00b907b99a4a8dbfc9c01ea49010878dfbd6\", \"repository_url\": \"https://github.com/GAIR-NLP/PC-Agent-E\"}\n\nREADME demonstrates Windows and Linux operation and describes thought completion and trajectory augmentation. The released source tree contains postprocess, train and deploy code. The checkpoint, data archive and WindowsAgentArena-V2 are separate resources; external model APIs and a training framework are required.\n\nPrimary source: https://github.com/GAIR-NLP/PC-Agent-E\nCatalog record: https://awesomecua.github.io/resources/github-gair-nlp-pc-agent-e/"}
{"id": "github:gair-nlp/windowsagentarena-v2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2", "citation_url": "https://awesomecua.github.io/resources/github-gair-nlp-windowsagentarena-v2/", "details": {"benchmark_protocol": {"actions": ["PyAutoGUI"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "population": "full WindowsAgentArena-V2 suite reported by README", "qualifier": "exact", "unit": "tasks", "value": 141}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "population": "full suite coverage reported by README", "qualifier": "exact", "unit": "Windows applications", "value": 11}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json", "population": "shipped test_custom.json selection", "qualifier": "exact", "unit": "task IDs", "value": 20}], "evaluation": "Runner exposes four observation choices and uses PyAutoGUI. count.py sums the last score in each results.json and divides by the number of traj.jsonl files, multiplying by 100; it separately reports result/trajectory/error-file coverage. This denominator is not automatically the full benchmark size.", "evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/run.py", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json", "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/count.py", "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2/git/trees/main?recursive=1"], "focus": ["Windows desktop task execution", "revised task evaluators"], "notes": "README sample config uses max_steps=30. The shipped selected-task index has 20 IDs, not 141; preserve the selected task list, trial ID, score coverage and VM snapshot when comparing runs. Docker/Windows-11 setup is required. No VM or benchmark was executed.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "2927fe55005d1be75d5a9188c0045f73ba28d192", "repository_url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2"}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "1dd744ce565ec00b20ea5637eb7bc6112e97fb9044bd9a471f7cc4b7227ca344", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:20Z", "role": "primary", "sha256": "0b0af02d3c2c4916dd786b868b99d60b4681927d0273a23661eb2b6db7cf9404", "supports": ["identity"], "url": "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "638ef9ee593a3a560c1d0b3fb2bc4a9d31a3e904c0111971924a6d66e1763aa6", "supports": ["identity", "release_status", "details.repository_tree_sha", "details.benchmark_protocol"], "url": "https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "494de9e74629ac2c55358629c0d9ba13703f36317057a3089388fbb6e80ca28f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/run.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "e45ba7fdc7555f46ad1b4cfaab79d00754909d053fc49d5bab651b0fb593439b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "f12147ac78c39c72eb75f9c37b293ba7fff3209fc5758fd51c69abbd96d80a97", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/count.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:20Z", "role": "primary", "sha256": "2a2d1e46940b570d60f5e2198ed27fcec6316fc8489e9660da8777dbd75aaeec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:20Z", "url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2"}], "name": "WindowsAgentArena-V2", "publication": null, "published_at": null, "record_sha256": "cfb4e02173e351c3e6b1546c06fca82e3a9d996bceec551c3a4bdc9b572c724a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "resource_id": "github:gair-nlp/windowsagentarena-v2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/GAIR-NLP/WindowsAgentArena-V2", "topics": ["evaluation", "cross-app", "robustness"], "updated_at": null}, "text": "WindowsAgentArena-V2\nResource type: benchmark\nReview status: curated\nScope: core\n\nWindows benchmark revision with task/evaluator fixes and VM-initialization support, documented alongside PC-Agent-E.\nPlatforms: desktop, windows\nTopics: evaluation, cross-app, robustness\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"PyAutoGUI\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md\", \"population\": \"full WindowsAgentArena-V2 suite reported by README\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 141}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md\", \"population\": \"full suite coverage reported by README\", \"qualifier\": \"exact\", \"unit\": \"Windows applications\", \"value\": 11}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json\", \"population\": \"shipped test_custom.json selection\", \"qualifier\": \"exact\", \"unit\": \"task IDs\", \"value\": 20}], \"evaluation\": \"Runner exposes four observation choices and uses PyAutoGUI. count.py sums the last score in each results.json and divides by the number of traj.jsonl files, multiplying by 100; it separately reports result/trajectory/error-file coverage. This denominator is not automatically the full benchmark size.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/README.md\", \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/run.py\", \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_custom.json\", \"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/count.py\", \"https://api.github.com/repos/GAIR-NLP/WindowsAgentArena-V2/git/trees/main?recursive=1\"], \"focus\": [\"Windows desktop task execution\", \"revised task evaluators\"], \"notes\": \"README sample config uses max_steps=30. The shipped selected-task index has 20 IDs, not 141; preserve the selected task list, trial ID, score coverage and VM snapshot when comparing runs. Docker/Windows-11 setup is required. No VM or benchmark was executed.\", \"observations\": [\"screenshot\", \"accessibility tree\", \"screenshot plus accessibility tree\", \"set-of-marks\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/GAIR-NLP/WindowsAgentArena-V2/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"2927fe55005d1be75d5a9188c0045f73ba28d192\", \"repository_url\": \"https://github.com/GAIR-NLP/WindowsAgentArena-V2\"}\n\nRepository contains executable evaluation code and task definitions. README links a separate Windows 11 VM snapshot; that image and its software licenses/access were not validated by downloading it. MIT is recorded for repository code only. Default task selection and the reported full suite have different scopes.\n\nPrimary source: https://github.com/GAIR-NLP/WindowsAgentArena-V2\nCatalog record: https://awesomecua.github.io/resources/github-gair-nlp-windowsagentarena-v2/"}
{"id": "github:garyyang12345/gui-harvest", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/GaryYang12345/GUI-HARVEST", "citation_url": "https://awesomecua.github.io/resources/github-garyyang12345-gui-harvest/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "2e42f01578710d91145412f7d1ecb93979db96c698f677915284da1a1d592588", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:29Z", "role": "primary", "sha256": "74e521e46c46489e48f9bad5d774625a85b5ccd32828969fd98e1a4c453b899f", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/GaryYang12345/GUI-HARVEST/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/GaryYang12345/GUI-HARVEST"}], "name": "GUI-HARVEST", "publication": null, "published_at": null, "record_sha256": "b9716fe8f45db5fb3740fd7f13ce7739596aaadd166994d32ad50a64a986a585", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=0&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2610.00948", "type": "documented_by"}], "resource_id": "github:garyyang12345/gui-harvest", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/GaryYang12345/GUI-HARVEST", "topics": ["automation", "observability", "robustness"], "updated_at": null}, "text": "GUI-HARVEST\nResource type: framework\nReview status: curated\nScope: core\n\nOptimizes GUI-agent execution code from repeated rollout evidence and checks proposed changes against fixed validation gates.\nPlatforms: desktop\nTopics: automation, observability, robustness\n\nRequires a separately installed benchmark and seed agent harness; benchmark tasks and model weights remain fixed during its optimization procedure.\n\nPrimary source: https://github.com/GaryYang12345/GUI-HARVEST\nCatalog record: https://awesomecua.github.io/resources/github-garyyang12345-gui-harvest/"}
{"id": "github:ghadinehme/videocad", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ghadinehme/VideoCAD", "citation_url": "https://awesomecua.github.io/resources/github-ghadinehme-videocad/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"], "reason": "MIT declaration in the author repository LICENSE; separate data, weights and document scopes are not inferred."}}, "repository_id": 1079441024, "repository_revision": "2d69eeaf0ac44210c76dcab2d08b4d4610be236c"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:30.768857+00:00", "role": "primary", "sha256": "ff80afd35c79d8dd27dd7b9284ce082fe238c13a113aa19400ca0dc2d18ff194", "supports": ["identity", "details.repository_id"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:52.564274+00:00", "role": "primary", "sha256": "9ded7c6ef25b88879e9d46939ca7ff3bef97362e76dbdb808e57c67008541557", "supports": ["release_status", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/ghadinehme/VideoCAD/git/trees/2d69eeaf0ac44210c76dcab2d08b4d4610be236c?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "summary", "relations", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "role": "primary", "sha256": "8e02753a1016dc67c97b395934313027b5f91b74209c7a24438de70fdc19cad1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:53.536297+00:00", "licenses": {"code": {"basis": "The author repository at the inspected commit contains the standard MIT license. This declaration does not establish terms for externally hosted data or weights.", "declared": "MIT", "retrieved_at": "2026-10-18T12:03:24.765415+00:00", "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ghadinehme/VideoCAD"}], "name": "VideoCAD training and evaluation code", "publication": null, "published_at": null, "record_sha256": "9b6cdbdb314f9c68ecc64c9f88ac1af49ed2a27d32f030e57d4099f9b3bfacc9", "relations": [{"evidence_url": "https://arxiv.org/html/2505.24838v2", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "resource_id": "github:ghadinehme/videocad", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ghadinehme/VideoCAD", "topics": ["grounding", "imitation-learning", "evaluation"], "updated_at": null}, "text": "VideoCAD training and evaluation code\nResource type: framework\nReview status: curated\nScope: core\n\nVideoCADFormer training, action-log preprocessing and evaluation code for learning CAD interface action sequences from screen recordings of browser-based Onshape.\nPlatforms: web\nTopics: grounding, imitation-learning, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/LICENSE\"], \"reason\": \"MIT declaration in the author repository LICENSE; separate data, weights and document scopes are not inferred.\"}}, \"repository_id\": 1079441024, \"repository_revision\": \"2d69eeaf0ac44210c76dcab2d08b4d4610be236c\"}\n\nThe paper links this author repository. Training, preprocessing, model and evaluation source files are released; runtime behavior has not been independently reproduced. Onshape is a browser-based CAD environment; supported operating systems are not inferred. The full dataset and a checkpoint are linked separately, and their payloads were not downloaded for this review. MIT is recorded only for this repository code; external data and weight licenses remain unknown.\n\nPrimary source: https://github.com/ghadinehme/VideoCAD\nCatalog record: https://awesomecua.github.io/resources/github-ghadinehme-videocad/"}
{"id": "github:google-deepmind/android_env", "metadata": {"aliases": ["github:deepmind/android_env", "deepmind/android_env", "https://github.com/deepmind/android_env"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-deepmind/android_env", "citation_url": "https://awesomecua.github.io/resources/github-google-deepmind-android-env/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "b8dd2ea004647008265f4fb5c5a122569992311b6758f413080984266e53c25a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:19Z", "role": "primary", "sha256": "de82c4101e81a6980789904be59c3733ab13f2c1f0e769e5fcb1ab35f847cb3e", "supports": ["aliases"], "url": "https://api.github.com/repos/deepmind/android_env"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:20Z", "role": "primary", "sha256": "de82c4101e81a6980789904be59c3733ab13f2c1f0e769e5fcb1ab35f847cb3e", "supports": ["aliases"], "url": "https://api.github.com/repos/google-deepmind/android_env"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-deepmind/android_env"}], "name": "AndroidEnv", "publication": null, "published_at": null, "record_sha256": "c484d76c34ddc9a18082a95a3ca87b2cb2862170d6fcd021424a28f76735d24d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-deepmind/android_env/HEAD/README.md", "target_id": "arxiv:2105.13231", "type": "documented_by"}], "resource_id": "github:google-deepmind/android_env", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-deepmind/android_env", "topics": ["sandbox", "reinforcement-learning", "automation"], "updated_at": null}, "text": "AndroidEnv\nResource type: infrastructure\nReview status: curated\nScope: core\n\nAndroid device-control platform with screenshots, accessibility telemetry, input actions, and an optional reinforcement-learning task interface.\nRecorded aliases: github:deepmind/android_env, deepmind/android_env, https://github.com/deepmind/android_env\nPlatforms: android\nTopics: sandbox, reinforcement-learning, automation\n\nFormerly published as deepmind/android_env. That address now redirects to this repository, and both paths serve the same reviewed README, so the project is counted once.\n\nPrimary source: https://github.com/google-deepmind/android_env\nCatalog record: https://awesomecua.github.io/resources/github-google-deepmind-android-env/"}
{"id": "github:google-research-datasets/screen2words", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research-datasets/screen2words", "citation_url": "https://awesomecua.github.io/resources/github-google-research-datasets-screen2words/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md", "population": "screen2words release", "qualifier": "exact", "unit": "summaries", "value": 112085}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md", "population": "RICO screens referenced by screen2words", "qualifier": "exact", "unit": "screens", "value": 22417}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md"], "formats": ["CSV"], "modalities": ["text summaries", "RICO screen references"], "notes": "Multiple summaries describe one screen. The repository provides annotations and split files; use the referenced RICO assets for screenshots.", "splits": ["train", "dev", "test"], "supervision": ["screen summarization"]}, "modalities": ["text", "screenshot-reference"], "sample_count": 112085, "sample_unit": "screen summaries", "screen_count": 22417}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "8be66dc5a3907e0e84993a66679416d928ed1a32aae70db8bb0e0f771c7a8715", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.screen_count", "details.modalities", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/screen2words"}], "name": "Screen2Words", "publication": null, "published_at": null, "record_sha256": "7bc54c0c307b8de0f4419c3e81e5d9238a014ddcae91fa67d2f7a82353eaf358", "relations": [], "resource_id": "github:google-research-datasets/screen2words", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research-datasets/screen2words", "topics": ["screen-understanding"], "updated_at": null}, "text": "Screen2Words\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile screen descriptions for language-based screen understanding.\nPlatforms: android\nTopics: screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md\", \"population\": \"screen2words release\", \"qualifier\": \"exact\", \"unit\": \"summaries\", \"value\": 112085}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md\", \"population\": \"RICO screens referenced by screen2words\", \"qualifier\": \"exact\", \"unit\": \"screens\", \"value\": 22417}], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research-datasets/screen2words/HEAD/README.md\"], \"formats\": [\"CSV\"], \"modalities\": [\"text summaries\", \"RICO screen references\"], \"notes\": \"Multiple summaries describe one screen. The repository provides annotations and split files; use the referenced RICO assets for screenshots.\", \"splits\": [\"train\", \"dev\", \"test\"], \"supervision\": [\"screen summarization\"]}, \"modalities\": [\"text\", \"screenshot-reference\"], \"sample_count\": 112085, \"sample_unit\": \"screen summaries\", \"screen_count\": 22417}\n\nPrimary source: https://github.com/google-research-datasets/screen2words\nCatalog record: https://awesomecua.github.io/resources/github-google-research-datasets-screen2words/"}
{"id": "github:google-research-datasets/screen_qa", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research-datasets/screen_qa", "citation_url": "https://awesomecua.github.io/resources/github-google-research-datasets-screen-qa/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA train split", "qualifier": "exact", "unit": "question-answer pairs", "value": 68951}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA validation split", "qualifier": "exact", "unit": "question-answer pairs", "value": 8614}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "population": "original ScreenQA test split", "qualifier": "exact", "unit": "question-answer pairs", "value": 8419}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md"], "formats": ["JSON"], "modalities": ["questions", "answers", "screen references", "bounding boxes"], "notes": "All questions for a given screenshot share one split. ScreenQA Short reuses the same questions/screens; ComplexQA is a separate extension. Do not add these variants as independent original examples.", "splits": ["train", "validation", "test"], "supervision": ["screen question answering", "answer grounding"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "7d1bfe7d6eddd5d3fc4c33d4814de3cabf7a6b6fd3d3f85ba4c2b33036bf8867", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/screen_qa"}], "name": "ScreenQA", "publication": null, "published_at": null, "record_sha256": "e14cbe01e81a1fb2cf60f021af215e40c07dca7dbf411e127df11cd8a4267ba5", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md", "target_id": "arxiv:2209.08199", "type": "documented_by"}], "resource_id": "github:google-research-datasets/screen_qa", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research-datasets/screen_qa", "topics": ["screen-understanding"], "updated_at": null}, "text": "ScreenQA\nResource type: dataset\nReview status: curated\nScope: core\n\nAbout 86,000 human-written question-answer pairs over 35,000 Rico mobile screenshots, plus ScreenQA Short and ComplexQA variants, for benchmarking screen content understanding.\nPlatforms: android\nTopics: screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md\", \"population\": \"original ScreenQA train split\", \"qualifier\": \"exact\", \"unit\": \"question-answer pairs\", \"value\": 68951}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md\", \"population\": \"original ScreenQA validation split\", \"qualifier\": \"exact\", \"unit\": \"question-answer pairs\", \"value\": 8614}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md\", \"population\": \"original ScreenQA test split\", \"qualifier\": \"exact\", \"unit\": \"question-answer pairs\", \"value\": 8419}], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research-datasets/screen_qa/HEAD/README.md\"], \"formats\": [\"JSON\"], \"modalities\": [\"questions\", \"answers\", \"screen references\", \"bounding boxes\"], \"notes\": \"All questions for a given screenshot share one split. ScreenQA Short reuses the same questions/screens; ComplexQA is a separate extension. Do not add these variants as independent original examples.\", \"splits\": [\"train\", \"validation\", \"test\"], \"supervision\": [\"screen question answering\", \"answer grounding\"]}}\n\nPrimary source: https://github.com/google-research-datasets/screen_qa\nCatalog record: https://awesomecua.github.io/resources/github-google-research-datasets-screen-qa/"}
{"id": "github:google-research-datasets/uibert", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research-datasets/uibert", "citation_url": "https://awesomecua.github.io/resources/github-google-research-datasets-uibert/", "details": {"availability": "Public archived TFRecord annotation files for both downstream tasks.", "dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "formats": ["TFRecord"], "modalities": ["text", "bounding boxes", "UI hierarchy metadata", "Rico image identifiers"], "notes": "Original UI images are retrieved separately by Rico IDs. No release-specific sample count is established.", "splits": ["AppSim train/dev/test", "RefExp train/dev/test"], "supervision": ["similar UI element retrieval", "referring-expression component retrieval"]}, "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"], "reason": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms."}}, "repository_revision": "eceafbde7567959899bfb7f8d9e6fd71a52aaa12"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "45d1ca0f5c727cf58b2671dc5579f0628f4d78e07b977a5d8b018295c6fe1f58", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "24a7838d425bdae363a7d5872702996b22068c1f0e47be274fdfbc4b0e20144f", "supports": ["identity"], "url": "https://api.github.com/repos/google-research-datasets/uibert"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "9a6fc19c6b407309943fe645f4addb3d032f153856976c4f8d7a8cfb6d2ffe6b", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/google-research-datasets/uibert/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "f752ebfb2598b7f848d343f428502b28e0f591f27849ab707f14bee398d4eb8d", "supports": ["environments", "notes"], "url": "https://interactionmining.org/rico"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-17T12:30:54Z", "url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/uibert"}], "name": "UIBert AppSim and RefExp annotations", "publication": null, "published_at": null, "record_sha256": "510c6f169dcd9179c2e777cad6549156c9caad88c8af122495fed9c36fe72b49", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "url:0b134b082a7bd507fc6a", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md", "target_id": "arxiv:2107.13731", "type": "documented_by"}], "resource_id": "github:google-research-datasets/uibert", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research-datasets/uibert", "topics": ["grounding", "screen-understanding", "evaluation"], "updated_at": null}, "text": "UIBert AppSim and RefExp annotations\nResource type: dataset\nReview status: curated\nScope: core\n\nTwo Rico-derived UI component retrieval datasets: AppSim pairs semantically similar elements across screens, while RefExp links referring expressions to target elements.\nPlatforms: android\nTopics: grounding, screen-understanding, evaluation\n\nRecorded specifications: {\"availability\": \"Public archived TFRecord annotation files for both downstream tasks.\", \"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md\"], \"formats\": [\"TFRecord\"], \"modalities\": [\"text\", \"bounding boxes\", \"UI hierarchy metadata\", \"Rico image identifiers\"], \"notes\": \"Original UI images are retrieved separately by Rico IDs. No release-specific sample count is established.\", \"splits\": [\"AppSim train/dev/test\", \"RefExp train/dev/test\"], \"supervision\": [\"similar UI element retrieval\", \"referring-expression component retrieval\"]}, \"license_review\": {\"data\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/google-research-datasets/uibert/HEAD/README.md\"], \"reason\": \"The README explicitly says the datasets are licensed under CC BY 4.0; original Rico materials keep their source terms.\"}}, \"repository_revision\": \"eceafbde7567959899bfb7f8d9e6fd71a52aaa12\"}\n\nThe archived repository retains TFRecord train/dev/test files for AppSim and RefExp. These are annotation and UI-structure records with Rico image identifiers, not a republished complete screenshot collection or a UIBert checkpoint. The README’s 72k figure describes the upstream Rico source, not the number of released AppSim/RefExp examples; no sample total is inferred. The publisher explicitly assigns CC-BY-4.0 to these datasets.\n\nPrimary source: https://github.com/google-research-datasets/uibert\nCatalog record: https://awesomecua.github.io/resources/github-google-research-datasets-uibert/"}
{"id": "github:google-research-datasets/uicrit", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research-datasets/uicrit", "citation_url": "https://awesomecua.github.io/resources/github-google-research-datasets-uicrit/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md", "population": "current three-annotator release", "qualifier": "exact", "unit": "mobile UI screens", "value": 1000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md", "population": "current three-annotator release", "qualifier": "exact", "unit": "design critiques", "value": 11344}], "evidence_urls": ["https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md"], "formats": ["CSV"], "modalities": ["text", "bounding boxes", "quality ratings"], "notes": "Publisher-reported current release, separate from the smaller original-paper population. Screenshot pixels are obtained through referenced RICO IDs.", "splits": [], "supervision": ["design critique", "region localization", "quality rating"]}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "345ff7003120f146016741099dd57b9b1a9ffc54ce1fc90f4d16d61e31d379e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "5579bbbc622a2f9fa1f77b509c1ece17af023b8c14b53aa48dcc41288a5852f2", "supports": ["identity"], "url": "https://api.github.com/repos/google-research-datasets/uicrit"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:45Z", "role": "primary", "sha256": "8e790a53de8c40aa14f9e49aebf42b077e6a172a6841f38c7c6851af57b0617a", "supports": ["release_status"], "url": "https://api.github.com/repos/google-research-datasets/uicrit/contents"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research-datasets/uicrit"}], "name": "UICrit", "publication": null, "published_at": null, "record_sha256": "2227e219152b548befc91f464f248e83f3e2ebc99b32e4218829d2e6de46040b", "relations": [], "resource_id": "github:google-research-datasets/uicrit", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research-datasets/uicrit", "topics": ["screen-understanding", "ui-design", "evaluation"], "updated_at": null}, "text": "UICrit\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile-interface critiques and quality ratings with region coordinates and RICO screen references, including a larger three-annotator release than the original paper dataset.\nPlatforms: unspecified\nTopics: screen-understanding, ui-design, evaluation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md\", \"population\": \"current three-annotator release\", \"qualifier\": \"exact\", \"unit\": \"mobile UI screens\", \"value\": 1000}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md\", \"population\": \"current three-annotator release\", \"qualifier\": \"exact\", \"unit\": \"design critiques\", \"value\": 11344}], \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research-datasets/uicrit/HEAD/README.md\"], \"formats\": [\"CSV\"], \"modalities\": [\"text\", \"bounding boxes\", \"quality ratings\"], \"notes\": \"Publisher-reported current release, separate from the smaller original-paper population. Screenshot pixels are obtained through referenced RICO IDs.\", \"splits\": [], \"supervision\": [\"design critique\", \"region localization\", \"quality rating\"]}}\n\nThe published CSV contains critique text, normalized boxes, ratings and RICO IDs; screenshots are referenced through RICO. The README explicitly licenses the work CC-BY-4.0. Its current 1,000-screen, 11,344-critique release differs from the original paper population. The README cites an ACM DOI rather than an arXiv ID, so no paper relation is asserted here. Its README establishes mobile interfaces without naming an operating system.\n\nPrimary source: https://github.com/google-research-datasets/uicrit\nCatalog record: https://awesomecua.github.io/resources/github-google-research-datasets-uicrit/"}
{"id": "github:google-research/android_world", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/google-research/android_world", "citation_url": "https://awesomecua.github.io/resources/github-google-research-android-world/", "details": {"benchmark_protocol": {"actions": ["Android device interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "population": "parameterized AndroidWorld tasks", "qualifier": "exact", "unit": "task definitions", "value": 116}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "population": "documented AndroidWorld app set", "qualifier": "exact", "unit": "applications", "value": 20}], "evaluation": "Programmatic task-specific initialization, system-state success checks, and teardown.", "evidence_urls": ["https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"], "focus": ["mobile task completion"], "notes": "Random parameters produce many instances from the same task definitions. Preserve seeds and parameterization; 116 is not a count of all possible task instances.", "observations": ["screenshots", "accessibility information"], "setting": "interactive"}, "evaluation_type": "programmatic state checks", "leaderboard_url": "https://docs.google.com/spreadsheets/d/1cchzP9dlTZ3WXQTfYNhh3avxoLipqHN75v1Tb86uhHo/edit?gid=0#gid=0", "metrics": ["task success"], "task_count": 116, "task_unit": "parameterized task definitions"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "93201de5e21f2b12283a75b1e2c357781b2e6cc27eab20cc7dd7aa0f5919d839", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.task_count", "details.task_unit", "details.metrics", "details.evaluation_type", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/google-research/android_world"}], "name": "AndroidWorld", "publication": null, "published_at": null, "record_sha256": "40180ee86d032c922730b382aa213ca49a5f5b4e574fd56f73a94b0550d5542c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md", "target_id": "arxiv:2405.14573", "type": "documented_by"}], "resource_id": "github:google-research/android_world", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/google-research/android_world", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "AndroidWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nAndroid environment with programmatic task evaluation for mobile agents.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"Android device interactions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md\", \"population\": \"parameterized AndroidWorld tasks\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 116}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md\", \"population\": \"documented AndroidWorld app set\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 20}], \"evaluation\": \"Programmatic task-specific initialization, system-state success checks, and teardown.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/google-research/android_world/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50\"], \"focus\": [\"mobile task completion\"], \"notes\": \"Random parameters produce many instances from the same task definitions. Preserve seeds and parameterization; 116 is not a count of all possible task instances.\", \"observations\": [\"screenshots\", \"accessibility information\"], \"setting\": \"interactive\"}, \"evaluation_type\": \"programmatic state checks\", \"leaderboard_url\": \"https://docs.google.com/spreadsheets/d/1cchzP9dlTZ3WXQTfYNhh3avxoLipqHN75v1Tb86uhHo/edit?gid=0#gid=0\", \"metrics\": [\"task success\"], \"task_count\": 116, \"task_unit\": \"parameterized task definitions\"}\n\nPrimary source: https://github.com/google-research/android_world\nCatalog record: https://awesomecua.github.io/resources/github-google-research-android-world/"}
{"id": "github:gydpku/pptc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/gydpku/PPTC", "citation_url": "https://awesomecua.github.io/resources/github-gydpku-pptc/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "63349e7557bd99240d4628de3994a53b69e54e3bd809895be11e9879bfba40cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "f872ff2b1a064f46a45eee86f2cb2562fc0513285b80036c6cf7df9a5ffc6c70", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/gydpku/PPTC"}], "name": "PPTC", "publication": null, "published_at": null, "record_sha256": "8b5b5f7ac2ee6f792e070399c6200a3c8023830d2971b342da84ef0928af0b3d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/gydpku/PPTC/HEAD/README.md", "target_id": "arxiv:2311.01767", "type": "documented_by"}], "resource_id": "github:gydpku/pptc", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/gydpku/PPTC", "topics": ["evaluation", "tool-integration"], "updated_at": null}, "text": "PPTC\nResource type: benchmark\nReview status: curated\nScope: adjacent\n\nPowerPoint creation and editing benchmark with API-based execution and checks of the resulting presentation files.\nPlatforms: unspecified\nTopics: evaluation, tool-integration\n\nPrimary source: https://github.com/gydpku/PPTC\nCatalog record: https://awesomecua.github.io/resources/github-gydpku-pptc/"}
{"id": "github:hazyresearch/wonderbread", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/HazyResearch/wonderbread", "citation_url": "https://awesomecua.github.io/resources/github-hazyresearch-wonderbread/", "details": {"availability": "Public benchmark code and public Zenodo dataset release metadata.", "benchmark_protocol": {"actions": ["generate procedures", "segment demonstrations", "answer workflow questions", "validate and rank procedures"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "population": "WONDERBREAD BPM benchmark", "qualifier": "exact", "unit": "task types", "value": 6}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "full dataset release 14094162", "qualifier": "exact", "unit": "human demonstrations", "value": 2928}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "full dataset release 14094162", "qualifier": "exact", "unit": "web navigation workflows", "value": 598}, {"basis": "reported", "evidence_url": "https://zenodo.org/api/records/14094162", "population": "gold subset used for the original paper evaluations", "qualifier": "exact", "unit": "human demonstrations", "value": 724}], "evaluation": "Task-specific scripts evaluate six workflow-understanding task types; model outputs are not treated as successful interactive execution.", "evidence_urls": ["https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "https://zenodo.org/api/records/14094162"], "focus": ["workflow documentation", "workflow quality assessment", "workflow knowledge transfer"], "notes": "Counts are publisher-reported. The gold subset covers 162 workflows and differs from the full collection; scores require the exact task and subset.", "observations": ["screen recordings", "action traces", "screenshots", "written standard operating procedures"], "setting": "offline"}, "dataset_release": "10.5281/zenodo.14094162", "license_review": {"data": {"evidence_urls": ["https://zenodo.org/api/records/14094162"], "reason": "The linked Zenodo dataset release declares CC-BY-4.0; this does not assign a code license to the repository."}}, "repository_revision": "ed052c67aeada04167cdfe92ff8de454aa94627a"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:11Z", "role": "primary", "sha256": "1e4308d59a22b39543c28b745a58d3d7c47be1ab9df9a4c5c0fe868a810350b2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:11Z", "role": "primary", "sha256": "5ba27e24455b36b52bf65b3c9f589e760f97a576bb01d622d89836c5bb53b38c", "supports": ["identity"], "url": "https://api.github.com/repos/HazyResearch/wonderbread"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:04Z", "role": "primary", "sha256": "accb518a5362948cf96e9f0b684fb6fcc9b26e43a691016ae70a3283c0e1e2c2", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/HazyResearch/wonderbread/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:34Z", "role": "primary", "sha256": "68b4b952c1ab90484d457b392de1d0c95a24f6cc1674c82125a6bd26f1cb1016", "supports": ["release_status", "licenses.data", "details.benchmark_protocol"], "url": "https://zenodo.org/api/records/14094162"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:11Z", "licenses": {"code": null, "data": {"basis": "The linked Zenodo dataset release declares CC-BY-4.0; this does not assign a code license to the repository.", "declared": "cc-by-4.0", "retrieved_at": "2026-10-17T09:15:34Z", "url": "https://zenodo.org/api/records/14094162"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HazyResearch/wonderbread"}], "name": "WONDERBREAD", "publication": null, "published_at": null, "record_sha256": "ebcbed78f1bf422155b3a46890d2905c934fb2f909f4ce2f50b29ecedb00e80d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md", "target_id": "arxiv:2406.13264", "type": "documented_by"}], "resource_id": "github:hazyresearch/wonderbread", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/HazyResearch/wonderbread", "topics": ["evaluation", "workflow", "video", "screen-understanding"], "updated_at": null}, "text": "WONDERBREAD\nResource type: benchmark\nReview status: curated\nScope: core\n\nAn offline benchmark of workflow documentation, quality assessment and knowledge transfer using human web-navigation demonstrations, action traces and written procedures.\nPlatforms: web\nTopics: evaluation, workflow, video, screen-understanding\n\nRecorded specifications: {\"availability\": \"Public benchmark code and public Zenodo dataset release metadata.\", \"benchmark_protocol\": {\"actions\": [\"generate procedures\", \"segment demonstrations\", \"answer workflow questions\", \"validate and rank procedures\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md\", \"population\": \"WONDERBREAD BPM benchmark\", \"qualifier\": \"exact\", \"unit\": \"task types\", \"value\": 6}, {\"basis\": \"reported\", \"evidence_url\": \"https://zenodo.org/api/records/14094162\", \"population\": \"full dataset release 14094162\", \"qualifier\": \"exact\", \"unit\": \"human demonstrations\", \"value\": 2928}, {\"basis\": \"reported\", \"evidence_url\": \"https://zenodo.org/api/records/14094162\", \"population\": \"full dataset release 14094162\", \"qualifier\": \"exact\", \"unit\": \"web navigation workflows\", \"value\": 598}, {\"basis\": \"reported\", \"evidence_url\": \"https://zenodo.org/api/records/14094162\", \"population\": \"gold subset used for the original paper evaluations\", \"qualifier\": \"exact\", \"unit\": \"human demonstrations\", \"value\": 724}], \"evaluation\": \"Task-specific scripts evaluate six workflow-understanding task types; model outputs are not treated as successful interactive execution.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/HazyResearch/wonderbread/HEAD/README.md\", \"https://zenodo.org/api/records/14094162\"], \"focus\": [\"workflow documentation\", \"workflow quality assessment\", \"workflow knowledge transfer\"], \"notes\": \"Counts are publisher-reported. The gold subset covers 162 workflows and differs from the full collection; scores require the exact task and subset.\", \"observations\": [\"screen recordings\", \"action traces\", \"screenshots\", \"written standard operating procedures\"], \"setting\": \"offline\"}, \"dataset_release\": \"10.5281/zenodo.14094162\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://zenodo.org/api/records/14094162\"], \"reason\": \"The linked Zenodo dataset release declares CC-BY-4.0; this does not assign a code license to the repository.\"}}, \"repository_revision\": \"ed052c67aeada04167cdfe92ff8de454aa94627a\"}\n\nThe repository contains evaluation and collection code; the linked Zenodo release exposes the full, gold and debug demonstration archives and metadata. This measures workflow understanding rather than interactive agent task completion. Zenodo declares CC-BY-4.0 for that dataset; no repository-level code license was found. Archives and human demonstration records were not downloaded.\n\nPrimary source: https://github.com/HazyResearch/wonderbread\nCatalog record: https://awesomecua.github.io/resources/github-hazyresearch-wonderbread/"}
{"id": "github:heimingx/tag", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/HeimingX/TAG", "citation_url": "https://awesomecua.github.io/resources/github-heimingx-tag/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "68cb0deb8a2eb49fb9236ae3d885b043d240dded0fce31cf7425130a56ff5819", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "a33bee17685d64d23fb4d65b30b9c52f3a6dd5b21789d7b3509eb2e15145de6f", "supports": ["identity"], "url": "https://api.github.com/repos/HeimingX/TAG"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "fac4f5e38011bbcfef916063d495c21d49830759af4e8713a60d7bae992d2710", "supports": ["release_status"], "url": "https://api.github.com/repos/HeimingX/TAG/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HeimingX/TAG"}], "name": "TAG attention-driven GUI grounding", "publication": null, "published_at": null, "record_sha256": "0a49b89516d9c2b73d84c1904376af10bf517fb254ebc27dc1d35bd24a42cc0e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/HeimingX/TAG/HEAD/README.md", "target_id": "arxiv:2412.10840", "type": "documented_by"}], "resource_id": "github:heimingx/tag", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/HeimingX/TAG", "topics": ["grounding", "evaluation", "training-free"], "updated_at": null}, "text": "TAG attention-driven GUI grounding\nResource type: tool\nReview status: curated\nScope: core\n\nGrounding and evaluation code that aggregates a pretrained vision-language model’s attention maps to localize requested interface elements without fine-tuning.\nPlatforms: web\nTopics: grounding, evaluation, training-free\n\nThe release contains grounding and benchmark evaluation code. The inspected root repository metadata does not declare a code license. It uses external pretrained models rather than publishing a new trained checkpoint, and the README restricts redistribution of unzipped benchmark data.\n\nPrimary source: https://github.com/HeimingX/TAG\nCatalog record: https://awesomecua.github.io/resources/github-heimingx-tag/"}
{"id": "github:hitsz-tmg/windowsworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/HITsz-TMG/WindowsWorld", "citation_url": "https://awesomecua.github.io/resources/github-hitsz-tmg-windowsworld/", "details": {"benchmark_protocol": {"actions": ["PyAutoGUI", "UiPath Computer_13 action interface"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "population": "WindowsWorld benchmark described in the retained README", "qualifier": "exact", "unit": "tasks", "value": 181}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "population": "WindowsWorld coverage described in the retained README", "qualifier": "exact", "unit": "desktop applications", "value": 17}], "evaluation": "Publisher distinguishes S_int (intermediate checkpoint score on L1–L3) and S_final (final completion on L1–L4). Documented step budgets are 15/25/40/20 for L1/L2/L3/L4. Preserve level, observation mode, agent interface and checkpoint/final metric when reporting a run.", "evidence_urls": ["https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "https://api.github.com/repos/HITsz-TMG/WindowsWorld/git/trees/main?recursive=1"], "focus": ["multi-application Windows workflows", "intermediate-state and final-goal completion"], "notes": "Interactive VMware/Windows environment. The public repository includes benchmark.json, task/evaluator code and monitoring support; the README refers VM import to Installation Guide.md, whose distribution links and payloads were not inspected for this profile. Checkpoint counts and application counts are not extra tasks. Apache-2.0 is recorded for repository code; Windows, apps and VM artifacts retain separate terms.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "9857bd8ae9f521f152761ae3bd85debe4e7471aad34c9eb47618332948bca6c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "490c9257d56f0a5214e6424e02490379b0b00ca2f7d4626fd5d14e17239d05ed", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/HITsz-TMG/WindowsWorld/git/trees/main?recursive=1"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/HITsz-TMG/WindowsWorld"}], "name": "WindowsWorld", "publication": null, "published_at": null, "record_sha256": "6989cbfe2dcac16bdc3e1ecc19f57ba2e2425d32d772ece4bc41057a6ef9beb3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md", "target_id": "arxiv:2604.27776", "type": "documented_by"}], "resource_id": "github:hitsz-tmg/windowsworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/HITsz-TMG/WindowsWorld", "topics": ["evaluation", "cross-app"], "updated_at": null}, "text": "WindowsWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nCross-application Windows benchmark with intermediate task evaluation.\nPlatforms: desktop, windows\nTopics: evaluation, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"PyAutoGUI\", \"UiPath Computer_13 action interface\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md\", \"population\": \"WindowsWorld benchmark described in the retained README\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 181}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md\", \"population\": \"WindowsWorld coverage described in the retained README\", \"qualifier\": \"exact\", \"unit\": \"desktop applications\", \"value\": 17}], \"evaluation\": \"Publisher distinguishes S_int (intermediate checkpoint score on L1–L3) and S_final (final completion on L1–L4). Documented step budgets are 15/25/40/20 for L1/L2/L3/L4. Preserve level, observation mode, agent interface and checkpoint/final metric when reporting a run.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/HITsz-TMG/WindowsWorld/HEAD/README.md\", \"https://api.github.com/repos/HITsz-TMG/WindowsWorld/git/trees/main?recursive=1\"], \"focus\": [\"multi-application Windows workflows\", \"intermediate-state and final-goal completion\"], \"notes\": \"Interactive VMware/Windows environment. The public repository includes benchmark.json, task/evaluator code and monitoring support; the README refers VM import to Installation Guide.md, whose distribution links and payloads were not inspected for this profile. Checkpoint counts and application counts are not extra tasks. Apache-2.0 is recorded for repository code; Windows, apps and VM artifacts retain separate terms.\", \"observations\": [\"screenshot\", \"accessibility tree\", \"screenshot plus accessibility tree\", \"set-of-marks\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/HITsz-TMG/WindowsWorld\nCatalog record: https://awesomecua.github.io/resources/github-hitsz-tmg-windowsworld/"}
{"id": "github:hud-evals/hud-python", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/hud-evals/hud-python", "citation_url": "https://awesomecua.github.io/resources/github-hud-evals-hud-python/", "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "3ef331612819a006d4279ac59203971a22a56fb6d9a0047f85a82668d09c327b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c88cc185e5547fce850b56f3b4072bed161658dc7277707e509437b9d760d470", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/hud-evals/hud-python/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/hud-evals/hud-python"}], "name": "HUD", "publication": null, "published_at": null, "record_sha256": "c54761eaced73533fe1207f15e6e7bd4db2fb957448068047f7ac5f9c016c128", "relations": [], "resource_id": "github:hud-evals/hud-python", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/hud-evals/hud-python", "topics": ["evaluation", "reinforcement-learning"], "updated_at": null}, "text": "HUD\nResource type: framework\nReview status: curated\nScope: core\n\nSDK for RL environments and evals spanning coding, browser, computer-use and robotics agents, whose environments expose SSH, MCP, CDP or VNC capabilities behind a minimal start-and-grade protocol.\nPlatforms: web, desktop\nTopics: evaluation, reinforcement-learning\n\nPrimary source: https://github.com/hud-evals/hud-python\nCatalog record: https://awesomecua.github.io/resources/github-hud-evals-hud-python/"}
{"id": "github:huggingface/screensuite", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/huggingface/screensuite", "citation_url": "https://awesomecua.github.io/resources/github-huggingface-screensuite/", "details": {"availability": "Public benchmark-adapter source and running instructions.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"], "reason": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services."}}, "repository_revision": "9ceda5070d5664dd9376d5f8e7b43f02188e01ff"}, "environments": ["web", "android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:13Z", "role": "primary", "sha256": "edcaa6f8355be6a8774ebeeb9bdb67648e8eeddb813d6be0fa2433116d9def36", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:12Z", "role": "primary", "sha256": "e372458e48d0ebbbe0f7887b66e2d5bb4db612fd96db4dc53404c0319891f7e8", "supports": ["identity"], "url": "https://api.github.com/repos/huggingface/screensuite"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:05Z", "role": "primary", "sha256": "93b685a13bfba4d9569dbfc0418de93c7b9f8a3f7c101af053aadeebbdf8b2c0", "supports": ["release_status"], "url": "https://api.github.com/repos/huggingface/screensuite/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:13Z", "licenses": {"code": {"basis": "Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:06Z", "url": "https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/huggingface/screensuite"}], "name": "ScreenSuite", "publication": null, "published_at": null, "record_sha256": "33935a38369deff0c566e9a6ebe01f096b3dabc9152ae3645d2ee9bdfaa85c02", "relations": [], "resource_id": "github:huggingface/screensuite", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/huggingface/screensuite", "topics": ["evaluation", "grounding", "action-prediction"], "updated_at": null}, "text": "ScreenSuite\nResource type: framework\nReview status: curated\nScope: core\n\nA shared evaluation harness for GUI-model perception, offline action prediction and multi-step interaction, using simple smolagents-based agents across existing benchmarks.\nPlatforms: web, android, desktop\nTopics: evaluation, grounding, action-prediction\n\nRecorded specifications: {\"availability\": \"Public benchmark-adapter source and running instructions.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/huggingface/screensuite/HEAD/LICENSE\"], \"reason\": \"Inspected repository LICENSE; not a license for linked benchmark datasets, pretrained weights or services.\"}}, \"repository_revision\": \"9ceda5070d5664dd9376d5f8e7b43f02188e01ff\"}\n\nThe publisher states that the suite compares the underlying multimodal models rather than full agent implementations. Public source includes benchmark adapters and registry code. Multi-step environments require Docker and KVM; constituent datasets, remote model services and OSWorld account-dependent tasks retain their own setup and terms. Apache-2.0 is scoped to this repository code.\n\nPrimary source: https://github.com/huggingface/screensuite\nCatalog record: https://awesomecua.github.io/resources/github-huggingface-screensuite/"}
{"id": "github:hyperbrowserai/hyperagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/hyperbrowserai/HyperAgent", "citation_url": "https://awesomecua.github.io/resources/github-hyperbrowserai-hyperagent/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "9fe74ee087156efde4531a058ccd3abad504032e0bd994979992ed4164c8704b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/hyperbrowserai/HyperAgent/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/hyperbrowserai/HyperAgent"}], "name": "HyperAgent", "publication": null, "published_at": null, "record_sha256": "8cfc3f4a9ecd253ef40c028cf87d409b4a81478c08bec0388d5f5039989263c1", "relations": [{"evidence_url": "https://raw.githubusercontent.com/hyperbrowserai/HyperAgent/HEAD/README.md", "target_id": "organization:hyperbrowser", "type": "published_by"}], "resource_id": "github:hyperbrowserai/hyperagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/hyperbrowserai/HyperAgent", "topics": ["web-navigation"], "updated_at": null}, "text": "HyperAgent\nResource type: framework\nReview status: curated\nScope: core\n\nLibrary adding natural-language commands to Playwright: accessibility-tree single actions, multi-step tasks that can use screenshots, an MCP client and action caching for replay without LLM calls.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/hyperbrowserai/HyperAgent\nCatalog record: https://awesomecua.github.io/resources/github-hyperbrowserai-hyperagent/"}
{"id": "github:imeanai/webcanvas", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/iMeanAI/WebCanvas", "citation_url": "https://awesomecua.github.io/resources/github-imeanai-webcanvas/", "details": {"leaderboard_url": "https://github.com/iMeanAI/WebCanvas/blob/main/Mind2web-live_Leaderboard.md"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5b2b4b4ce1241db7fd776f2ba0856ec2e0e441a56fa6461234abe0c6516f643e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "cbaab6bcbc9c3e28d0de902e12f92c2466103c775cb2a5af1e6d4110e82223a6", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/iMeanAI/WebCanvas"}], "name": "WebCanvas", "publication": null, "published_at": null, "record_sha256": "819ee60aff579134fd5edb882050b4f3f9d33f92ce55c021ceb3a8bfc343575c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/iMeanAI/WebCanvas/HEAD/README.md", "target_id": "arxiv:2406.12373", "type": "documented_by"}], "resource_id": "github:imeanai/webcanvas", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/iMeanAI/WebCanvas", "topics": ["evaluation"], "updated_at": null}, "text": "WebCanvas\nResource type: benchmark\nReview status: curated\nScope: core\n\nFramework and evaluation resources for real-world web agents.\nPlatforms: web\nTopics: evaluation\n\nRecorded specifications: {\"leaderboard_url\": \"https://github.com/iMeanAI/WebCanvas/blob/main/Mind2web-live_Leaderboard.md\"}\n\nPrimary source: https://github.com/iMeanAI/WebCanvas\nCatalog record: https://awesomecua.github.io/resources/github-imeanai-webcanvas/"}
{"id": "github:imnearth/coat", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/IMNearth/CoAT", "citation_url": "https://awesomecua.github.io/resources/github-imnearth-coat/", "details": {"availability": "Public CoAT prompting/action-prediction and evaluation source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md", "https://api.github.com/repos/IMNearth/CoAT/git/trees/main?recursive=1"], "reason": "Apache-2.0 badge alone does not establish component-scoped source terms; the inspected non-truncated tree has no license file."}}, "repository_revision": "9ab1a9f0e76581c50c84f8ebc6ada1a2084a0c71"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "7a9fb905b64f2760a137cd267a5e38dd343df1821eda8172b606ab7500d27731", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "d0c5056c9a74913f84a4f809bf01d9fd9ea325ac0a91b4c4a0a51220b2da9d68", "supports": ["identity"], "url": "https://api.github.com/repos/IMNearth/CoAT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:04Z", "role": "primary", "sha256": "29d84ed12dc25035fd539604b06f6cc113080b389edeb9aa605cfa55c5f28151", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/IMNearth/CoAT/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/IMNearth/CoAT"}], "name": "CoAT research implementation", "publication": null, "published_at": null, "record_sha256": "4e134aed03886372026ca482767999c79d415fc35bd85868ba6f7083f0daab6a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md", "target_id": "arxiv:2403.02713", "type": "documented_by"}], "resource_id": "github:imnearth/coat", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/IMNearth/CoAT", "topics": ["planning", "mobile-navigation", "action-prediction"], "updated_at": null}, "text": "CoAT research implementation\nResource type: framework\nReview status: curated\nScope: core\n\nResearch scripts that generate screen descriptions, action reasoning and expected outcomes, then predict Android GUI actions using the resulting Chain-of-Action-Thought context.\nPlatforms: android\nTopics: planning, mobile-navigation, action-prediction\n\nRecorded specifications: {\"availability\": \"Public CoAT prompting/action-prediction and evaluation source.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/IMNearth/CoAT/HEAD/README.md\", \"https://api.github.com/repos/IMNearth/CoAT/git/trees/main?recursive=1\"], \"reason\": \"Apache-2.0 badge alone does not establish component-scoped source terms; the inspected non-truncated tree has no license file.\"}}, \"repository_revision\": \"9ab1a9f0e76581c50c84f8ebc6ada1a2084a0c71\"}\n\nThe tree contains run_coat.py and coat/agent.py, model.py and evaluate.py, with API-backed model configuration. This verifies a released annotation/prediction implementation, not a live phone controller or a released checkpoint. AiTZ screens/annotations are linked through external Drive/Baidu bundles whose complete download inventory was not verified here. The README has an Apache-2.0 badge, but no license file or explicit component grant is present; code/data/weight rights remain unestablished.\n\nPrimary source: https://github.com/IMNearth/CoAT\nCatalog record: https://awesomecua.github.io/resources/github-imnearth-coat/"}
{"id": "github:inclusionai/ui-venus", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/inclusionAI/UI-Venus", "citation_url": "https://awesomecua.github.io/resources/github-inclusionai-ui-venus/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5703c892aaf99d250bf59edf52d359560a80d443f2be13c7aa76af0c1a7f1d67", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/inclusionAI/UI-Venus"}], "name": "UI-Venus", "publication": null, "published_at": null, "record_sha256": "f8b4785d2572faa0da63b2cc291739b3675e33924e6302a7342fea94b0cf6f77", "relations": [{"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2508.10833", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2602.09082", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/inclusionAI/UI-Venus/HEAD/README.md", "target_id": "arxiv:2609.00028", "type": "documented_by"}], "resource_id": "github:inclusionai/ui-venus", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/inclusionAI/UI-Venus", "topics": ["grounding", "planning"], "updated_at": null}, "text": "UI-Venus\nResource type: model\nReview status: curated\nScope: core\n\nGUI agent model family; UI-Venus-2, at 9B and 27B, acts on mobile, web and desktop in one end-to-end agent, trained with keypoint-based multi-judge task verification.\nPlatforms: cross-platform\nTopics: grounding, planning\n\nPrimary source: https://github.com/inclusionAI/UI-Venus\nCatalog record: https://awesomecua.github.io/resources/github-inclusionai-ui-venus/"}
{"id": "github:infixai/infigui-g1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/InfiXAI/InfiGUI-G1", "citation_url": "https://awesomecua.github.io/resources/github-infixai-infigui-g1/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/InfiXAI/InfiGUI-G1"}], "name": "InfiGUI-G1", "publication": null, "published_at": null, "record_sha256": "aaa032dac862748b06e41b78af2174c520388eb110d0bcf469e69430cd40938a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "resource_id": "github:infixai/infigui-g1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/InfiXAI/InfiGUI-G1", "topics": ["grounding"], "updated_at": null}, "text": "InfiGUI-G1\nResource type: model\nReview status: curated\nScope: core\n\nInfiGUI-G1 3B and 7B grounding models trained with Adaptive Exploration Policy Optimization, which samples several candidate answers per pass under an adaptive exploration reward.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/InfiXAI/InfiGUI-G1\nCatalog record: https://awesomecua.github.io/resources/github-infixai-infigui-g1/"}
{"id": "github:jacklishufan/mobileworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/jacklishufan/MobileWorld", "citation_url": "https://awesomecua.github.io/resources/github-jacklishufan-mobileworld/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:51Z", "role": "primary", "sha256": "dfcbaa088cc21dd4e3c67d6eb3b3a1dc476217165ce2ca0daa2facfda71b5a38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/jacklishufan/MobileWorld"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jacklishufan/MobileWorld"}], "name": "MobileWorldBench", "publication": null, "published_at": null, "record_sha256": "2a235e0c52a98bff8d08d121db99f638d9390d0642e21ab31fe5506b5dfbd4d9", "relations": [{"evidence_url": "https://github.com/jacklishufan/MobileWorld", "target_id": "arxiv:2512.14014", "type": "documented_by"}], "resource_id": "github:jacklishufan/mobileworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/jacklishufan/MobileWorld", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "MobileWorldBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nBenchmark of semantic world modeling for mobile agents, with question-answering tasks and GPT-judged generation tasks, plus training data built on AitW and AndroidControl screens.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nPrimary source: https://github.com/jacklishufan/MobileWorld\nCatalog record: https://awesomecua.github.io/resources/github-jacklishufan-mobileworld/"}
{"id": "github:jcwang0602/seg2click", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/jcwang0602/Seg2Click", "citation_url": "https://awesomecua.github.io/resources/github-jcwang0602-seg2click/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "2316acbef1207925b10b4770e14bdf69f6f4b6b3bc0ddb12e6be107d6f1425da", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "57bcc2eb985fa6036bf2bde7f28c4fd359c6b3479a658fea768162cac2a75ba5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jcwang0602/Seg2Click"}], "name": "Seg2Click", "publication": null, "published_at": null, "record_sha256": "ed91831d9f9a528f6a136f5ecc07a1ee71a36c9a0f4367fe7858cb29e88bcbd8", "relations": [{"evidence_url": "https://raw.githubusercontent.com/jcwang0602/Seg2Click/HEAD/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "github:jcwang0602/seg2click", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/jcwang0602/Seg2Click", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "Seg2Click\nResource type: framework\nReview status: curated\nScope: core\n\nTraining and inference implementation that predicts an executable-region mask and derives a GUI click from that mask.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\n\nThe repository provides training, evaluation, and a custom mask decoder. Its code license does not determine the licenses of the five model checkpoints. No model execution was performed.\n\nPrimary source: https://github.com/jcwang0602/Seg2Click\nCatalog record: https://awesomecua.github.io/resources/github-jcwang0602-seg2click/"}
{"id": "github:jetastra/macagentbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/JetAstra/MacAgentBench", "citation_url": "https://awesomecua.github.io/resources/github-jetastra-macagentbench/", "details": {"benchmark_protocol": {"actions": ["GUI interactions", "command-line operations"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "population": "documented MacAgentBench suite", "qualifier": "exact", "unit": "tasks", "value": 676}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "population": "documented application coverage", "qualifier": "exact", "unit": "applications", "value": 25}], "evaluation": "Deterministic rule-based evaluators with intermediate checkpoints.", "evidence_urls": ["https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"], "focus": ["macOS task completion", "cross-application workflows"], "notes": "The benchmark evaluates multiple agent frameworks and GUI/CLI workflows. Framework skills and interfaces affect the task setting; the inspected README does not define one universal observation representation.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://jetastra.github.io/MacAgentBench/"}, "environments": ["desktop", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "f1f8c0cd5d984e4026abea858b54f0f1427c018cad7e75322e744f1e070225a9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:03Z", "role": "primary", "sha256": "2adad07a9ec7189aea25e48c76fa28206eda97f8974d5745cadda9529103bb1a", "supports": ["details.benchmark_protocol", "relations"], "url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "adf69997028988ba1eec34666afa119e007771dfc8e830c5dc50ad650896e897", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/JetAstra/MacAgentBench"}], "name": "MacAgentBench", "publication": null, "published_at": null, "record_sha256": "b7dd40eb4c6835d3bf64edb5e4dc2849d8d0fdc2acc1dfde8d2ea31f8c728b7e", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48", "target_id": "arxiv:2606.22557", "type": "documented_by"}], "resource_id": "github:jetastra/macagentbench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/JetAstra/MacAgentBench", "topics": ["evaluation"], "updated_at": null}, "text": "MacAgentBench\nResource type: benchmark\nReview status: curated\nScope: core\n\n676 macOS tasks across 25 applications, each run in its own Docker container and scored by deterministic rule-based checks with multiple checkpoints.\nPlatforms: desktop, macos\nTopics: evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"GUI interactions\", \"command-line operations\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md\", \"population\": \"documented MacAgentBench suite\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 676}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md\", \"population\": \"documented application coverage\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 25}], \"evaluation\": \"Deterministic rule-based evaluators with intermediate checkpoints.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/JetAstra/MacAgentBench/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2605.06761%2C2605.07110%2C2605.07134%2C2605.09497%2C2605.11212%2C2605.12481%2C2605.12501%2C2605.12549%2C2605.13631%2C2605.14290%2C2605.14747%2C2605.15030%2C2605.15542%2C2605.15777%2C2605.16024%2C2605.16116%2C2605.16402%2C2605.17320%2C2605.18048%2C2605.18652%2C2605.18758%2C2605.19260%2C2605.19769%2C2605.21470%2C2605.25160%2C2605.25624%2C2605.25707%2C2605.26114%2C2605.27761%2C2605.28116%2C2605.28534%2C2605.28646%2C2605.28775%2C2605.29119%2C2605.29324%2C2605.29447%2C2605.29486%2C2605.29534%2C2605.29615%2C2606.06322%2C2606.09426%2C2606.11042%2C2606.22557%2C2606.23189%2C2606.24551%2C2608.17659%2C2609.22000%2C2609.34139&max_results=48\"], \"focus\": [\"macOS task completion\", \"cross-application workflows\"], \"notes\": \"The benchmark evaluates multiple agent frameworks and GUI/CLI workflows. Framework skills and interfaces affect the task setting; the inspected README does not define one universal observation representation.\", \"observations\": [], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://jetastra.github.io/MacAgentBench/\"}\n\nPrimary source: https://github.com/JetAstra/MacAgentBench\nCatalog record: https://awesomecua.github.io/resources/github-jetastra-macagentbench/"}
{"id": "github:jia-lab-research/arpo", "metadata": {"aliases": ["github:dvlab-research/arpo", "dvlab-research/ARPO", "https://github.com/dvlab-research/ARPO"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/JIA-Lab-research/ARPO", "citation_url": "https://awesomecua.github.io/resources/github-jia-lab-research-arpo/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "7489ae309e90d8b3b9e823f495b63e5116434060", "repository_url": "https://github.com/JIA-Lab-research/ARPO"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "839b2de0e0c614fe8d2df4782d0f6a6e7d33f5c60d8020c9d4a01c61d533f39e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["identity"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:23Z", "role": "primary", "sha256": "b3ff7f4e350223827cae0fffd986ee672e0b4466349c262ed2dc9e63a8480aeb", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["identity"], "url": "https://api.github.com/repos/dvlab-research/ARPO"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["aliases"], "url": "https://api.github.com/repos/dvlab-research/ARPO"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-18T04:40:39Z", "role": "primary", "sha256": "bb64c7fc1b5d88fc070e919516a51eefac4acb89e86773464cfa0b3de18f1898", "supports": ["aliases"], "url": "https://api.github.com/repos/JIA-Lab-research/ARPO"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/JIA-Lab-research/ARPO"}], "name": "ARPO training framework", "publication": null, "published_at": null, "record_sha256": "72c1652faf968f8ea8c3a3d375a8d749f8c2768534b5decc87cdcbecf5ec3b8d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md", "target_id": "arxiv:2505.16282", "type": "documented_by"}], "resource_id": "github:jia-lab-research/arpo", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/JIA-Lab-research/ARPO", "topics": ["reinforcement-learning", "long-horizon", "automation"], "updated_at": null}, "text": "ARPO training framework\nResource type: framework\nReview status: curated\nScope: core\n\nReplay-based policy-optimization framework for GUI agents, with distributed OSWorld rollouts and screenshot/action histories.\nRecorded aliases: github:dvlab-research/arpo, dvlab-research/ARPO, https://github.com/dvlab-research/ARPO\nPlatforms: desktop, linux\nTopics: reinforcement-learning, long-horizon, automation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"7489ae309e90d8b3b9e823f495b63e5116434060\", \"repository_url\": \"https://github.com/JIA-Lab-research/ARPO\"}\n\nOfficial GitHub API resolves the former dvlab-research/ARPO URL to JIA-Lab-research/ARPO with the same numeric repository ID. Training/evaluation scripts are released; OSWorld environments and the separate model checkpoint are not bundled into this code record. Reported benchmark scores were not reproduced.\n\nPrimary source: https://github.com/JIA-Lab-research/ARPO\nCatalog record: https://awesomecua.github.io/resources/github-jia-lab-research-arpo/"}
{"id": "github:jylee425/b-moca", "metadata": {"aliases": ["github:gimme1dollar/b-moca", "gimme1dollar/b-moca", "https://github.com/gimme1dollar/b-moca"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/jylee425/b-moca", "citation_url": "https://awesomecua.github.io/resources/github-jylee425-b-moca/", "details": {"benchmark_protocol": {"actions": ["normalized touch/lift coordinate pairs for tap or swipe"], "counts": [{"basis": "computed", "evidence_url": "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1", "population": "asset/tasks/**/*.textproto in retained release/ver.3 tree; not a paper test-task total", "qualifier": "exact", "unit": "task-definition files", "value": 131}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv", "population": "environments_train.csv, retained release/ver.3 snapshot", "qualifier": "exact", "unit": "environment configuration rows", "value": 35}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv", "population": "environments_test.csv, retained release/ver.3 snapshot", "qualifier": "exact", "unit": "environment configuration rows", "value": 10}], "evaluation": "Task success is detected using Appium; positive task reward terminates the episode. Environment variants change device/locale/wallpaper/DPI/dark mode and are loaded as snapshots. The numeric action interface is (touch_y, touch_x, lift_y, lift_x); wrapper/tanh transformations must be kept with the evaluated agent.", "evidence_urls": ["https://raw.githubusercontent.com/jylee425/b-moca/HEAD/readme.md", "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/coordinator.py", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/environment.py", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv", "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv"], "focus": ["Android control across visual/device variation", "generalization across environment snapshots"], "notes": "Current source inventory differs in grain from published experiment subsets: do not multiply 131 definitions by 35/10 configuration rows or label the product as evaluated tasks. Setup requires Android API 29, Appium, external app versions and generated AVD snapshots. No apps/snapshots were downloaded or run. Code Apache-2.0 does not establish data or app rights.", "observations": ["screenshot pixels", "XML view hierarchy", "goal instruction"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "bd1354c8008d7d79d4883c0c037b7918392f86fab61391b39b52e507b04672ce", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "06ba40a5e0d5fb52e7bf5e5df636cb8e8bbbc56a8308dc6d528012489f7650ee", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:24:53Z", "role": "primary", "sha256": "34bc0c23949732e0a8091e334d59d2d41cae49a82cde850247aa0e6d739d2464", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/coordinator.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "d866af344c0cff5255b813576f2e9e9c5c2964e6809058b39e2b5180af71d946", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/environment.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "c278cc28f8a0667b7fdc6ba638704cfbe43783cd496181231c45c865087543f3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:15Z", "role": "primary", "sha256": "4b92933fe4f4f66e71a694b579940db11bedd5ea8f153752a44e633825b06fa3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:03Z", "role": "primary", "sha256": "9d4706306b04960f0eb4298887496961efdee6b7d41445e82abcac489d63779f", "supports": ["aliases"], "url": "https://api.github.com/repos/gimme1dollar/b-moca"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:04Z", "role": "primary", "sha256": "9d4706306b04960f0eb4298887496961efdee6b7d41445e82abcac489d63779f", "supports": ["aliases"], "url": "https://api.github.com/repos/jylee425/b-moca"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/jylee425/b-moca/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jylee425/b-moca"}], "name": "B-MoCA", "publication": null, "published_at": null, "record_sha256": "391aa598615b113fa59a57f6ba1b83cf09dee7a5a71addca91b95ec766f4a343", "relations": [], "resource_id": "github:jylee425/b-moca", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/jylee425/b-moca", "topics": ["evaluation", "robustness"], "updated_at": null}, "text": "B-MoCA\nResource type: benchmark\nReview status: curated\nScope: core\n\nMobile control benchmark with varied interface configurations.\nRecorded aliases: github:gimme1dollar/b-moca, gimme1dollar/b-moca, https://github.com/gimme1dollar/b-moca\nPlatforms: android\nTopics: evaluation, robustness\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"normalized touch/lift coordinate pairs for tap or swipe\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1\", \"population\": \"asset/tasks/**/*.textproto in retained release/ver.3 tree; not a paper test-task total\", \"qualifier\": \"exact\", \"unit\": \"task-definition files\", \"value\": 131}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv\", \"population\": \"environments_train.csv, retained release/ver.3 snapshot\", \"qualifier\": \"exact\", \"unit\": \"environment configuration rows\", \"value\": 35}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv\", \"population\": \"environments_test.csv, retained release/ver.3 snapshot\", \"qualifier\": \"exact\", \"unit\": \"environment configuration rows\", \"value\": 10}], \"evaluation\": \"Task success is detected using Appium; positive task reward terminates the episode. Environment variants change device/locale/wallpaper/DPI/dark mode and are loaded as snapshots. The numeric action interface is (touch_y, touch_x, lift_y, lift_x); wrapper/tanh transformations must be kept with the evaluated agent.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/readme.md\", \"https://api.github.com/repos/jylee425/b-moca/git/trees/release/ver.3?recursive=1\", \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/coordinator.py\", \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/bmoca/environment/environment.py\", \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_train.csv\", \"https://raw.githubusercontent.com/jylee425/b-moca/HEAD/asset/environments/config/environments_test.csv\"], \"focus\": [\"Android control across visual/device variation\", \"generalization across environment snapshots\"], \"notes\": \"Current source inventory differs in grain from published experiment subsets: do not multiply 131 definitions by 35/10 configuration rows or label the product as evaluated tasks. Setup requires Android API 29, Appium, external app versions and generated AVD snapshots. No apps/snapshots were downloaded or run. Code Apache-2.0 does not establish data or app rights.\", \"observations\": [\"screenshot pixels\", \"XML view hierarchy\", \"goal instruction\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/jylee425/b-moca\nCatalog record: https://awesomecua.github.io/resources/github-jylee425-b-moca/"}
{"id": "github:jylee425/mobilesafetybench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/jylee425/mobilesafetybench", "citation_url": "https://awesomecua.github.io/resources/github-jylee425-mobilesafetybench/", "details": {"benchmark_protocol": {"actions": ["tap/long-tap/swipe", "navigation buttons and text editing", "refuse, ask-consent and complete", "task-specific shortcut actions"], "counts": [], "evaluation": "Appium-backed task evaluators track goal achievement and harm prevention as separate booleans, plus risk-detection step and completion progress. The base evaluator defaults to max_step=15; task-specific overrides must be retained. Results are not a single generic success-rate metric.", "evidence_urls": ["https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/README.md", "https://api.github.com/repos/jylee425/mobilesafetybench/git/trees/release/ver.3?recursive=1", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/environment.py", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/component/parser.py", "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/evaluator/_base_evaluator.py"], "focus": ["mobile task helpfulness", "harm prevention during Android interaction"], "notes": "Inspected release/ver.3 branch exposes task definitions, Android environment and per-task evaluators. It requires Android API 34, Appium/ADB and documented third-party app versions; those APKs/accounts were not acquired. The action parser includes semantic shortcuts in addition to screen gestures, so this is not a strictly coordinate-only action setting. No corpus task total is inferred from evaluator-module or file counts. Apache-2.0 applies to repository code, not automatically app/data assets.", "observations": ["screenshot pixels", "parsed UI hierarchy text", "task instruction and interaction history"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e733115d9b2fbe910cb3891162f04a89af8cbda25255c8f0bf60f78101e4309a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "8dc66f6c5087a32c5c9fb7b3104ea6f9e0082255e2ba3cf9ba8244ce5a24da6f", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/jylee425/mobilesafetybench/git/trees/release/ver.3?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "f56d6345546a9c67a2113cfba5b7148f5bd6b48fac5c34f1a5f4080e2d81ae2f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/environment.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "120487c3a6492371321dd3eebcd65dc59de6a82454d622d7fed3a7d86b8fa6c5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/component/parser.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "918e3aa64ab3f578fb54580da123c1ff4508022b64ce74f88ce28c7a612756a4", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/evaluator/_base_evaluator.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/LICENSE.txt"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/jylee425/mobilesafetybench"}], "name": "MobileSafetyBench", "publication": null, "published_at": null, "record_sha256": "99a1e37949d51f1c48db5a0738b27d548b60a1a866c9327a67c731e33c859f96", "relations": [], "resource_id": "github:jylee425/mobilesafetybench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/jylee425/mobilesafetybench", "topics": ["safety-security", "evaluation"], "updated_at": null}, "text": "MobileSafetyBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nAndroid emulator testbed that scores mobile device-control agents on both safety and helpfulness, using ADB and Appium for observations, actions and success checks.\nPlatforms: android\nTopics: safety-security, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"tap/long-tap/swipe\", \"navigation buttons and text editing\", \"refuse, ask-consent and complete\", \"task-specific shortcut actions\"], \"counts\": [], \"evaluation\": \"Appium-backed task evaluators track goal achievement and harm prevention as separate booleans, plus risk-detection step and completion progress. The base evaluator defaults to max_step=15; task-specific overrides must be retained. Results are not a single generic success-rate metric.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/README.md\", \"https://api.github.com/repos/jylee425/mobilesafetybench/git/trees/release/ver.3?recursive=1\", \"https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/environment.py\", \"https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/component/parser.py\", \"https://raw.githubusercontent.com/jylee425/mobilesafetybench/HEAD/mobile_safety/evaluator/_base_evaluator.py\"], \"focus\": [\"mobile task helpfulness\", \"harm prevention during Android interaction\"], \"notes\": \"Inspected release/ver.3 branch exposes task definitions, Android environment and per-task evaluators. It requires Android API 34, Appium/ADB and documented third-party app versions; those APKs/accounts were not acquired. The action parser includes semantic shortcuts in addition to screen gestures, so this is not a strictly coordinate-only action setting. No corpus task total is inferred from evaluator-module or file counts. Apache-2.0 applies to repository code, not automatically app/data assets.\", \"observations\": [\"screenshot pixels\", \"parsed UI hierarchy text\", \"task instruction and interaction history\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/jylee425/mobilesafetybench\nCatalog record: https://awesomecua.github.io/resources/github-jylee425-mobilesafetybench/"}
{"id": "github:kdegroup/ui-agile", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/KDEGroup/UI-AGILE", "citation_url": "https://awesomecua.github.io/resources/github-kdegroup-ui-agile/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "eb2a2767741b53e5a6659ef4b0cc325ac18ffdc4222deef1d0e7671089fcaa39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "a145c4ec8fd43f59ebe97708b6ee4cef8112959036957eaa4739b8da9a297e8a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/KDEGroup/UI-AGILE"}], "name": "UI-AGILE", "publication": null, "published_at": null, "record_sha256": "ea137637964de4203159244d067959478a8a74b8e632f4eb3e2d0b20c3e8554e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/KDEGroup/UI-AGILE/HEAD/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "resource_id": "github:kdegroup/ui-agile", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/KDEGroup/UI-AGILE", "topics": ["grounding", "reinforcement-learning"], "updated_at": null}, "text": "UI-AGILE\nResource type: model\nReview status: curated\nScope: core\n\nGUI agent trained by RL with a continuous reward, a simple-thinking reward and cropping-based resampling on about 9k samples, plus a plug-in decomposed grounding-and-selection inference method.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\n\nPrimary source: https://github.com/KDEGroup/UI-AGILE\nCatalog record: https://awesomecua.github.io/resources/github-kdegroup-ui-agile/"}
{"id": "github:kwai/memgui-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/kwai/MemGUI-Agent", "citation_url": "https://awesomecua.github.io/resources/github-kwai-memgui-agent/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/LICENSE", "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md"], "reason": "The README declares Apache-2.0, but the inspected LICENSE is a short notice and link rather than the complete terms; the code license remains unconfirmed."}}, "repository_id": 1278022559}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:54:57.854138+00:00", "role": "primary", "sha256": "3f7545f466407ed5336f47d18b12e8354ec352ff83dcf973b1545f81d9a629ac", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:26.237759+00:00", "role": "primary", "sha256": "4c92c584b6885a3bde8ee89d04ccf48287e7f6119889f75cde0c034c90bd8217", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/evaluation/memgui3k_offline_eval/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:01.061844+00:00", "role": "primary", "sha256": "d4e344274155be1d0a192c8ae2940169abb0bd8546d045b96b45810593860636", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-19T19:10:26.954787+00:00", "role": "primary", "sha256": "0ae4477b885b0915e43874399c2a6b91f7d07954697370204550df931f7db82c", "supports": ["details.repository_id"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-18&per_page=100&page=1&sort=updated&order=asc"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T03:54:57.854138+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kwai/MemGUI-Agent"}], "name": "MemGUI-Agent", "publication": null, "published_at": null, "record_sha256": "008e01766c959b7939bfd9f6d786ce87b8e6d4466a04a2cc46503bb4673cb8a6", "relations": [{"evidence_url": "https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md", "target_id": "arxiv:2606.19926", "type": "introduced_by"}], "resource_id": "github:kwai/memgui-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/kwai/MemGUI-Agent", "topics": ["memory", "interaction-history", "long-horizon", "imitation-learning", "evaluation"], "updated_at": null}, "text": "MemGUI-Agent\nResource type: framework\nReview status: curated\nScope: core\n\nMemGUI-Agent training and offline evaluation tools provide trajectory conversion, LoRA SFT templates, and step-level checks of GUI actions, memory updates and context folding.\nPlatforms: android\nTopics: memory, interaction-history, long-horizon, imitation-learning, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/LICENSE\", \"https://raw.githubusercontent.com/kwai/MemGUI-Agent/main/README.md\"], \"reason\": \"The README declares Apache-2.0, but the inspected LICENSE is a short notice and link rather than the complete terms; the code license remains unconfirmed.\"}}, \"repository_id\": 1278022559}\n\nThe inspected release documentation describes dataset utilities, a LoRA SFT template and gold-step offline evaluation; live Android rollouts are explicitly outside that evaluator and linked to separate benchmark repositories. Executable source files, external datasets/checkpoints and reported results were not independently verified in this bounded review.\n\nPrimary source: https://github.com/kwai/MemGUI-Agent\nCatalog record: https://awesomecua.github.io/resources/github-kwai-memgui-agent/"}
{"id": "github:kwai/mobileforge", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/kwai/MobileForge", "citation_url": "https://awesomecua.github.io/resources/github-kwai-mobileforge/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE", "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md"], "reason": "The complete root LICENSE declares Apache-2.0 for repository code. Linked data and model weights, and third-party dependencies, retain separate licensing requirements."}}, "repository_id": 1278021888}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:04.230792+00:00", "role": "primary", "sha256": "f78f58a6e3b8dd2ae3e8c49a15473782ae258ecd70719f7eeddfa43a10441adf", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "notes", "environments", "licenses.code"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:29.533431+00:00", "role": "primary", "sha256": "58187fe25d8398ef2ae31de02debd4bc3b932ff1925db86a87f5de65b016aed1", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T03:55:07.420029+00:00", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-19T19:10:26.954787+00:00", "role": "primary", "sha256": "0ae4477b885b0915e43874399c2a6b91f7d07954697370204550df931f7db82c", "supports": ["details.repository_id"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-18&per_page=100&page=1&sort=updated&order=asc"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T03:55:04.230792+00:00", "licenses": {"code": {"basis": "Complete repository root Apache License 2.0 text; code declaration only, not linked datasets or weights.", "declared": "Apache-2.0", "retrieved_at": "2026-10-20T03:55:07.420029+00:00", "url": "https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kwai/MobileForge"}], "name": "MobileForge", "publication": null, "published_at": null, "record_sha256": "f7e4b55b53f569a64ca9b1d4a6d057975e6d05f0c8386d5dd72ee61e96359d0b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/kwai/MobileForge/main/README.md", "target_id": "arxiv:2606.19930", "type": "introduced_by"}], "resource_id": "github:kwai/mobileforge", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/kwai/MobileForge", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning", "evaluation"], "updated_at": null}, "text": "MobileForge\nResource type: framework\nReview status: curated\nScope: core\n\nAndroid GUI-agent adaptation framework combining app exploration, generated curricula and rollout feedback with hint-conditioned step-level GRPO; the documented public training path targets Qwen3-VL-8B.\nPlatforms: android\nTopics: mobile-navigation, data-generation, reinforcement-learning, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/kwai/MobileForge/main/LICENSE\", \"https://raw.githubusercontent.com/kwai/MobileForge/main/README.md\"], \"reason\": \"The complete root LICENSE declares Apache-2.0 for repository code. Linked data and model weights, and third-party dependencies, retain separate licensing requirements.\"}}, \"repository_id\": 1278021888}\n\nThe root README describes MobileGym exploration/feedback and HiFPO optimization. The training README limits its public training path to the Qwen3-VL-8B experiment. Linked model and dataset collections, executable source completeness and benchmark results were not independently verified; the repository code license is not a model or dataset license.\n\nPrimary source: https://github.com/kwai/MobileForge\nCatalog record: https://awesomecua.github.io/resources/github-kwai-mobileforge/"}
{"id": "github:kyle8581/web-shepherd", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/kyle8581/Web-Shepherd", "citation_url": "https://awesomecua.github.io/resources/github-kyle8581-web-shepherd/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "52925efd66489696c4e1f31f0f8353997fe534df9fcd42e50232cc78f0506903", "supports": ["identity"], "url": "https://api.github.com/repos/kyle8581/Web-Shepherd"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "ba7d28866cec4fb889499fc011473aedcdcc36fd8175c73527eafd7ef6b6dbba", "supports": ["release_status"], "url": "https://api.github.com/repos/kyle8581/Web-Shepherd/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/kyle8581/Web-Shepherd"}], "name": "Web-Shepherd inference and evaluation", "publication": null, "published_at": null, "record_sha256": "6cfecd8c2bc523a116368429e221f2c5e72a585596ce30561ca45b22fca77c05", "relations": [], "resource_id": "github:kyle8581/web-shepherd", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/kyle8581/Web-Shepherd", "topics": ["reward-models", "evaluation", "web-navigation"], "updated_at": null}, "text": "Web-Shepherd inference and evaluation\nResource type: tool\nReview status: curated\nScope: core\n\nProcess-reward inference and benchmark-evaluation utilities that use task checklists to assess individual web-agent decisions.\nPlatforms: web\nTopics: reward-models, evaluation, web-navigation\n\nThe tree contains src/webprm modules and a WebRewardBench evaluation script. The README’s illustrative folder layout differs from the actual tree, so a complete training pipeline is not claimed. No root code license is declared. Some old checkpoint links are unavailable; current model releases are separately identified through the official Hub collection.\n\nPrimary source: https://github.com/kyle8581/Web-Shepherd\nCatalog record: https://awesomecua.github.io/resources/github-kyle8581-web-shepherd/"}
{"id": "github:laokuize/appagent-pro", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/LaoKuiZe/AppAgent-Pro", "citation_url": "https://awesomecua.github.io/resources/github-laokuize-appagent-pro/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9b4b945f4514d261373ec08d4c3e4e86b6623f95f95b9469bec2b8e7e93d9807", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "f13a0d1aaaec0f7e32f034197b31b2bc0537685aff11af653381454288c75794", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/LaoKuiZe/AppAgent-Pro"}], "name": "AppAgent Pro", "publication": null, "published_at": null, "record_sha256": "d529fd05626e949af3687dcd63b082c6d4113db425cccd851855b95f04423c1b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/LaoKuiZe/AppAgent-Pro/HEAD/README.md", "target_id": "arxiv:2508.18689", "type": "documented_by"}], "resource_id": "github:laokuize/appagent-pro", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/LaoKuiZe/AppAgent-Pro", "topics": ["mobile-navigation"], "updated_at": null}, "text": "AppAgent Pro\nResource type: agent\nReview status: curated\nScope: core\n\nProactive Android agent built on AppAgent that uses GPT-4o to decide whether YouTube or Amazon content would help, operates those apps over ADB, and merges what it finds into its answer.\nPlatforms: android\nTopics: mobile-navigation\n\nPrimary source: https://github.com/LaoKuiZe/AppAgent-Pro\nCatalog record: https://awesomecua.github.io/resources/github-laokuize-appagent-pro/"}
{"id": "github:lavague-ai/lavague", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/lavague-ai/LaVague", "citation_url": "https://awesomecua.github.io/resources/github-lavague-ai-lavague/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "7e4d38b95657b604f5f390e25762d94c8d59164fbb2ae6c140c126bb128dbd4f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/lavague-ai/LaVague/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lavague-ai/LaVague"}], "name": "LaVague", "publication": null, "published_at": null, "record_sha256": "109f4530c9166036034ad051ae2f0562deb6d37f8ce94f7e71867c682e99f5d8", "relations": [], "resource_id": "github:lavague-ai/lavague", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/lavague-ai/LaVague", "topics": ["web-navigation"], "updated_at": null}, "text": "LaVague\nResource type: framework\nReview status: curated\nScope: core\n\nOpen-source web agent framework where a World Model turns an objective and the current page into instructions and an Action Engine compiles them into Selenium or Playwright code to run.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/lavague-ai/LaVague\nCatalog record: https://awesomecua.github.io/resources/github-lavague-ai-lavague/"}
{"id": "github:lgy0404/learnact", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/lgy0404/LearnAct", "citation_url": "https://awesomecua.github.io/resources/github-lgy0404-learnact/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "95184255b6ec021aac8ca062b944f3ed76d1e01405e5771d9e31368021523eec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/lgy0404/LearnAct/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lgy0404/LearnAct"}], "name": "LearnAct", "publication": null, "published_at": null, "record_sha256": "b1913551186f30dc0d25e60298d421bef1cb0b115195a2460e31ebc725143df7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/lgy0404/LearnAct/HEAD/README.md", "target_id": "arxiv:2504.13805", "type": "documented_by"}], "resource_id": "github:lgy0404/learnact", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/lgy0404/LearnAct", "topics": ["mobile-navigation", "imitation-learning", "evaluation"], "updated_at": null}, "text": "LearnAct\nResource type: agent\nReview status: curated\nScope: core\n\nReleased inference and evaluation code for demonstration-conditioned mobile GUI agents on LearnGUI-Offline. The repository describes a multi-agent adaptation method and links the LearnGUI demonstrations.\nPlatforms: unspecified\nTopics: mobile-navigation, imitation-learning, evaluation\n\nAt the retained README snapshot, LearnGUI-Online code and the real-world deployment implementation were marked coming soon. The mobile operating system is not established by the reviewed release instructions; no code was executed.\n\nPrimary source: https://github.com/lgy0404/LearnAct\nCatalog record: https://awesomecua.github.io/resources/github-lgy0404-learnact/"}
{"id": "github:likaixin2000/screenspot-pro-gui-grounding", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding", "citation_url": "https://awesomecua.github.io/resources/github-likaixin2000-screenspot-pro-gui-grounding/", "details": {"benchmark_protocol": {"actions": ["predicted target coordinates"], "counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evaluation": "For positive targets, the predicted point must fall inside the annotated bounding box.", "evidence_urls": ["https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py", "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"], "focus": ["GUI grounding"], "notes": "Language, instruction style, and positive/negative target configuration are explicit evaluator options. This profile describes target localization, not end-to-end workflow completion. The target count is computed from the card table, not from expanded language/style evaluation instances.", "observations": ["high-resolution screenshots", "target instructions"], "setting": "offline"}, "leaderboard_url": "https://gui-agent.github.io/grounding-leaderboard"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "7b64ccad9cfec24c2c9cb35f55923bff93d53729cef3cc7463c79b29bdb1de1a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["details.benchmark_protocol", "environments"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c78681e624c748b7a9980b6a9fd1514cec5140c7b3c7f08fe21b36bc7468f2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding"}], "name": "ScreenSpot Pro", "publication": null, "published_at": null, "record_sha256": "acf14628fcab74d351b420d478516060119c41e95a630b0885675bd9d6945b86", "relations": [], "resource_id": "github:likaixin2000/screenspot-pro-gui-grounding", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding", "topics": ["evaluation", "grounding"], "updated_at": null}, "text": "ScreenSpot Pro\nResource type: benchmark\nReview status: curated\nScope: core\n\nGUI grounding benchmark focused on professional high-resolution interfaces.\nPlatforms: desktop, windows, macos, linux\nTopics: evaluation, grounding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"predicted target coordinates\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md\", \"population\": \"sum of icon/text counts in the published card table\", \"qualifier\": \"exact\", \"unit\": \"target annotations\", \"value\": 1581}], \"evaluation\": \"For positive targets, the predicted point must fall inside the annotated bounding box.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/readme.md\", \"https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py\", \"https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md\"], \"focus\": [\"GUI grounding\"], \"notes\": \"Language, instruction style, and positive/negative target configuration are explicit evaluator options. This profile describes target localization, not end-to-end workflow completion. The target count is computed from the card table, not from expanded language/style evaluation instances.\", \"observations\": [\"high-resolution screenshots\", \"target instructions\"], \"setting\": \"offline\"}, \"leaderboard_url\": \"https://gui-agent.github.io/grounding-leaderboard\"}\n\nPrimary source: https://github.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding\nCatalog record: https://awesomecua.github.io/resources/github-likaixin2000-screenspot-pro-gui-grounding/"}
{"id": "github:ljang0/iosworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ljang0/iOSWorld", "citation_url": "https://awesomecua.github.io/resources/github-ljang0-iosworld/", "details": {"benchmark_protocol": {"actions": ["touch actions", "MCP tools (optional)"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md", "population": "tasks.json benchmark", "qualifier": "exact", "unit": "tasks", "value": 133}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md", "population": "bundled SwiftUI apps", "qualifier": "exact", "unit": "applications", "value": 26}], "evaluation": "Rubric satisfaction per task; full-task pass requires every criterion to be satisfied.", "evidence_urls": ["https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"], "focus": ["personalized phone tasks", "cross-app workflows"], "notes": "Vision-only, vision-plus-XML, and MCP modes expose different information and tools. The documented default evaluator is model-based; compare the same mode and judge.", "observations": ["screenshots", "accessibility XML (optional)"], "setting": "interactive"}, "evaluation_type": "rubric-based", "metrics": ["mean rubric score", "full-task pass rate"], "task_count": 133}, "environments": ["ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "706f9d770b65a1238e56f74f7dd5089c7756cc88e0ac09d80c7b2b3085a09b27", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.task_count", "details.metrics", "details.evaluation_type", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "fa9fe1e31ed8da21d0b51621d3a3669ced4f0e42353aaf4bd8be1615d200eaac", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ljang0/iOSWorld"}], "name": "iOSWorld", "publication": null, "published_at": null, "record_sha256": "820c12934f23d36d0f3a589c6287e7056574e2a4c2ea97720fcd3ddbc4a8473b", "relations": [], "resource_id": "github:ljang0/iosworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ljang0/iOSWorld", "topics": ["evaluation", "cross-app", "memory"], "updated_at": null}, "text": "iOSWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nNative iPhone-simulator benchmark with shared user data, application workflows, agent runners, and rubric-based scoring.\nPlatforms: ios\nTopics: evaluation, cross-app, memory\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"touch actions\", \"MCP tools (optional)\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md\", \"population\": \"tasks.json benchmark\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 133}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md\", \"population\": \"bundled SwiftUI apps\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 26}], \"evaluation\": \"Rubric satisfaction per task; full-task pass requires every criterion to be satisfied.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/ljang0/iOSWorld/HEAD/README.md\"], \"focus\": [\"personalized phone tasks\", \"cross-app workflows\"], \"notes\": \"Vision-only, vision-plus-XML, and MCP modes expose different information and tools. The documented default evaluator is model-based; compare the same mode and judge.\", \"observations\": [\"screenshots\", \"accessibility XML (optional)\"], \"setting\": \"interactive\"}, \"evaluation_type\": \"rubric-based\", \"metrics\": [\"mean rubric score\", \"full-task pass rate\"], \"task_count\": 133}\n\nPrimary source: https://github.com/ljang0/iOSWorld\nCatalog record: https://awesomecua.github.io/resources/github-ljang0-iosworld/"}
{"id": "github:llamatouch/llamatouch", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/LlamaTouch/LlamaTouch", "citation_url": "https://awesomecua.github.io/resources/github-llamatouch-llamatouch/", "details": {"benchmark_protocol": {"actions": ["device actions through AgentEnv"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "complete task collection", "qualifier": "exact", "unit": "tasks", "value": 495}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "AITW-derived subset", "qualifier": "exact", "unit": "tasks", "value": 102}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "self-constructed subset", "qualifier": "exact", "unit": "tasks", "value": 393}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "population": "self-constructed subset", "qualifier": "exact", "unit": "Android applications", "value": 46}], "evaluation": "Checks whether the recorded execution visits all annotated essential states rather than requiring an identical action sequence.", "evidence_urls": ["https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"], "focus": ["mobile task completion", "essential-state evaluation"], "notes": "Application coverage refers to the self-constructed subset. Task execution and essential-state annotation are separate components; counts are publisher-reported, not a fresh task-file recount.", "observations": ["screenshots", "view hierarchies", "Android activity names"], "setting": "interactive"}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c4fe0fa92961f7286b6553b13117b2942e8482a48b7120a79e57dce2f8d42b9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c4fe0fa92961f7286b6553b13117b2942e8482a48b7120a79e57dce2f8d42b9", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/LlamaTouch/LlamaTouch"}], "name": "LlamaTouch", "publication": null, "published_at": null, "record_sha256": "a8045b6cf421cfa411b5293dd8c9267745e12297fe4bcdaf311402a8113e28b7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md", "target_id": "arxiv:2404.16054", "type": "documented_by"}], "resource_id": "github:llamatouch/llamatouch", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/LlamaTouch/LlamaTouch", "topics": ["evaluation"], "updated_at": null}, "text": "LlamaTouch\nResource type: benchmark\nReview status: curated\nScope: core\n\nMobile UI task evaluation using application state information.\nPlatforms: android\nTopics: evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"device actions through AgentEnv\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md\", \"population\": \"complete task collection\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 495}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md\", \"population\": \"AITW-derived subset\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 102}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md\", \"population\": \"self-constructed subset\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 393}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md\", \"population\": \"self-constructed subset\", \"qualifier\": \"exact\", \"unit\": \"Android applications\", \"value\": 46}], \"evaluation\": \"Checks whether the recorded execution visits all annotated essential states rather than requiring an identical action sequence.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/LlamaTouch/LlamaTouch/HEAD/README.md\"], \"focus\": [\"mobile task completion\", \"essential-state evaluation\"], \"notes\": \"Application coverage refers to the self-constructed subset. Task execution and essential-state annotation are separate components; counts are publisher-reported, not a fresh task-file recount.\", \"observations\": [\"screenshots\", \"view hierarchies\", \"Android activity names\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/LlamaTouch/LlamaTouch\nCatalog record: https://awesomecua.github.io/resources/github-llamatouch-llamatouch/"}
{"id": "github:lll6gg/ui-r1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/lll6gg/UI-R1", "citation_url": "https://awesomecua.github.io/resources/github-lll6gg-ui-r1/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/lll6gg/UI-R1"}], "name": "UI-R1", "publication": null, "published_at": null, "record_sha256": "da9324a863811a625ec77d03e0479e8f85c3a789b5f23725baf01379e66125e3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "resource_id": "github:lll6gg/ui-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/lll6gg/UI-R1", "topics": ["reinforcement-learning", "grounding"], "updated_at": null}, "text": "UI-R1\nResource type: model\nReview status: curated\nScope: core\n\nRule-based reinforcement learning for GUI action prediction, releasing UI-R1-3B, trained from Qwen2.5-VL-3B, the efficiency-oriented UI-R1-E-3B, and training data.\nPlatforms: cross-platform\nTopics: reinforcement-learning, grounding\n\nPrimary source: https://github.com/lll6gg/UI-R1\nCatalog record: https://awesomecua.github.io/resources/github-lll6gg-ui-r1/"}
{"id": "github:ltzheng/synapse", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ltzheng/Synapse", "citation_url": "https://awesomecua.github.io/resources/github-ltzheng-synapse/", "details": {"availability": "Public memory, prompting, browser-experiment and fine-tuning source.", "license_conflict": "README MIT badge versus NTUITIVE agreement; the agreement contains inconsistent distribution provisions.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ltzheng/synapse/HEAD/README.md", "https://raw.githubusercontent.com/ltzheng/Synapse/HEAD/LICENSE"], "reason": "Unresolved conflict between README MIT badge and NTUITIVE non-commercial LICENSE, including inconsistent redistribution clauses. Keep permissions unknown rather than choose an interpretation."}}, "repository_revision": "08c3a252ff158822e50e9c109c1cd49dc2f6147d"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "ed5416d91665017fca4b354016b7c2c748acf32d7723483fecb5387a22a1fdff", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ltzheng/synapse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "79a238a6f486d2a8970baba199f0968d28f14d42922c86ce71753eb687e0e9c6", "supports": ["identity"], "url": "https://api.github.com/repos/ltzheng/synapse"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "785a7163dac9f41b1d0ca97f8dd9311b6df6a09f3a5163438dd64db226a4ceaf", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/ltzheng/Synapse/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "cc4bce2274289d2b924fcc3c185c68b8e7c539365c5f8c72be865883244088ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ltzheng/Synapse/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ltzheng/Synapse"}], "name": "Synapse", "publication": null, "published_at": null, "record_sha256": "acdb52b3597f7d8c055598aa4006ae20ff104a9c3ce1b36c2273f0da223e1e8c", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50", "target_id": "arxiv:2306.07863", "type": "documented_by"}], "resource_id": "github:ltzheng/synapse", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ltzheng/Synapse", "topics": ["memory", "planning", "web-navigation", "imitation-learning"], "updated_at": null}, "text": "Synapse\nResource type: framework\nReview status: curated\nScope: core\n\nA computer-control research implementation that retrieves example trajectories from memory, abstracts task-relevant state and prompts an agent to navigate browser tasks.\nPlatforms: web\nTopics: memory, planning, web-navigation, imitation-learning\n\nRecorded specifications: {\"availability\": \"Public memory, prompting, browser-experiment and fine-tuning source.\", \"license_conflict\": \"README MIT badge versus NTUITIVE agreement; the agreement contains inconsistent distribution provisions.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/ltzheng/synapse/HEAD/README.md\", \"https://raw.githubusercontent.com/ltzheng/Synapse/HEAD/LICENSE\"], \"reason\": \"Unresolved conflict between README MIT badge and NTUITIVE non-commercial LICENSE, including inconsistent redistribution clauses. Keep permissions unknown rather than choose an interpretation.\"}}, \"repository_revision\": \"08c3a252ff158822e50e9c109c1cd49dc2f6147d\"}\n\nTraining, memory-building and evaluation scripts for MiniWoB++ and Mind2Web are present. External benchmark data, model/API credentials and legacy dependencies are still required. License remains unknown: the README shows MIT, while LICENSE is an NTUITIVE non-commercial agreement that both permits some distribution and separately says distribution is prohibited. No interpretation resolves those contradictions, and no commercial or non-commercial permission is asserted. This admission confirms the released GUI implementation, not a licensing grant.\n\nPrimary source: https://github.com/ltzheng/Synapse\nCatalog record: https://awesomecua.github.io/resources/github-ltzheng-synapse/"}
{"id": "github:madeagents/colorbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MadeAgents/ColorBench", "citation_url": "https://awesomecua.github.io/resources/github-madeagents-colorbench/", "details": {"availability": "Public graph/task manifests, evaluation source and ungated screenshot archive manifest.", "benchmark_protocol": {"actions": ["mobile GUI operations along graph transitions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "tasks", "value": 175}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "cross-app tasks", "value": 101}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "single-app tasks", "value": 74}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "population": "reported ColorBench benchmark", "qualifier": "exact", "unit": "mobile applications", "value": 21}], "evaluation": "The publisher reports success rate, completion rate and atomic-capability diagnostics using graph-based evaluation.", "evidence_urls": ["https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md"], "focus": ["long-horizon mobile task completion", "multi-path graph interaction"], "notes": "The interaction environment replays graph states rather than running real user accounts. The 101 and 74 subsets partition the reported 175 tasks; do not add the three counts.", "observations": ["mobile screenshots represented as graph states"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"], "reason": "Repository source LICENSE."}, "data": {"evidence_urls": ["https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"], "reason": "Publisher metadata for the separately hosted screenshot archive; underlying third-party content retains applicable rights."}}, "repository_revision": "1f26a84230d26f99d84eab97d05032582b39018e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "5c0190fffb68b91c1855a408f6a5e252871d94000b95fa258504a3c100600cec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "2f6940503cc0d6cb97bdf142d6b8c6ab116de82227f2fa6d97fd56d7e5aebec0", "supports": ["identity"], "url": "https://api.github.com/repos/MadeAgents/ColorBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "6a00c093a6794ec4a172d30b70fdabc17400584a9e25a157a47a7661dbdd20fa", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/MadeAgents/ColorBench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "c0c3286b5150e128e55034ba0393ef85f7fcf73e4b73f4276389a04c9bcec919", "supports": ["identity", "release_status"], "url": "https://huggingface.co/api/datasets/Norsheep/ColorBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "130976c026bcafa9dcc479697fe6595dbefb73cdb375fc47b8478acc16e6ea73", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:16Z", "licenses": {"code": {"basis": "Repository source LICENSE.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:54Z", "url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE"}, "data": {"basis": "Publisher metadata for the separately hosted screenshot archive; underlying third-party content retains applicable rights.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:06Z", "url": "https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MadeAgents/ColorBench"}], "name": "ColorBench", "publication": null, "published_at": null, "record_sha256": "ad65fc1e429f22b11daf28eefa871a5b4fa0b899516ee6059ef4a5bc9b1ced8d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md", "target_id": "arxiv:2510.14621", "type": "documented_by"}], "resource_id": "github:madeagents/colorbench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MadeAgents/ColorBench", "topics": ["evaluation", "mobile-navigation", "long-horizon", "cross-app"], "updated_at": null}, "text": "ColorBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nA graph-based mobile GUI benchmark that evaluates long tasks through screenshot states and multiple valid or erroneous transitions, including reflection, replanning and backtracking.\nPlatforms: android\nTopics: evaluation, mobile-navigation, long-horizon, cross-app\n\nRecorded specifications: {\"availability\": \"Public graph/task manifests, evaluation source and ungated screenshot archive manifest.\", \"benchmark_protocol\": {\"actions\": [\"mobile GUI operations along graph transitions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md\", \"population\": \"reported ColorBench benchmark\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 175}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md\", \"population\": \"reported ColorBench benchmark\", \"qualifier\": \"exact\", \"unit\": \"cross-app tasks\", \"value\": 101}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md\", \"population\": \"reported ColorBench benchmark\", \"qualifier\": \"exact\", \"unit\": \"single-app tasks\", \"value\": 74}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md\", \"population\": \"reported ColorBench benchmark\", \"qualifier\": \"exact\", \"unit\": \"mobile applications\", \"value\": 21}], \"evaluation\": \"The publisher reports success rate, completion rate and atomic-capability diagnostics using graph-based evaluation.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/README.md\"], \"focus\": [\"long-horizon mobile task completion\", \"multi-path graph interaction\"], \"notes\": \"The interaction environment replays graph states rather than running real user accounts. The 101 and 74 subsets partition the reported 175 tasks; do not add the three counts.\", \"observations\": [\"mobile screenshots represented as graph states\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/MadeAgents/ColorBench/HEAD/LICENSE\"], \"reason\": \"Repository source LICENSE.\"}, \"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/Norsheep/ColorBench/raw/main/README.md\"], \"reason\": \"Publisher metadata for the separately hosted screenshot archive; underlying third-party content retains applicable rights.\"}}, \"repository_revision\": \"1f26a84230d26f99d84eab97d05032582b39018e\"}\n\nThe source tree contains graph and task JSON files plus construction/evaluation code; the linked ungated Hub provides the screenshot archive. The environment is static but interactive and graph-based, not a live-device deployment. Counts are publisher-reported task/app inventories. Apache-2.0 declarations for repository code and the Hub dataset are separately retained; third-party application visuals are not granted additional rights by this catalog.\n\nPrimary source: https://github.com/MadeAgents/ColorBench\nCatalog record: https://awesomecua.github.io/resources/github-madeagents-colorbench/"}
{"id": "github:madeagents/mobile-use", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MadeAgents/mobile-use", "citation_url": "https://awesomecua.github.io/resources/github-madeagents-mobile-use/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "101595fa1f62d30271b0560f8aaa54286d7f6fc006e2aa9d8b73086583b27656", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "806f2d2c84a015f221f39691c95f7bdd0ff18f134e563d1155caa787076653fe", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MadeAgents/mobile-use"}], "name": "Mobile Use", "publication": null, "published_at": null, "record_sha256": "e3aaa463f59a1ee2af735994d9da596bcec6feaf78cad40976d6373df9ce2caa", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MadeAgents/mobile-use/HEAD/README.md", "target_id": "arxiv:2507.16853", "type": "documented_by"}], "resource_id": "github:madeagents/mobile-use", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MadeAgents/mobile-use", "topics": ["mobile-navigation"], "updated_at": null}, "text": "Mobile Use\nResource type: agent\nReview status: curated\nScope: core\n\nAndroid GUI agent that operates real phones through ADB with hierarchical reflection and a multi-agent framework, run from a web interface or Python and evaluated on AndroidWorld and AndroidLab.\nPlatforms: android\nTopics: mobile-navigation\n\nPrimary source: https://github.com/MadeAgents/mobile-use\nCatalog record: https://awesomecua.github.io/resources/github-madeagents-mobile-use/"}
{"id": "github:magicgh/self-map", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/magicgh/Self-MAP", "citation_url": "https://awesomecua.github.io/resources/github-magicgh-self-map/", "details": {"availability": "Public conversational planning, candidate-generation and training/evaluation source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"], "reason": "Project source LICENSE; separate dataset license is not folded into the code slot."}}, "repository_revision": "6cf779682b70dae597626f96dfa5130b202689bc"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "693197cb807e3ab4d5afda9de35dd51342b4feb18e68b8ccfcdaa370399dbeb3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/magicgh/self-map/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "82c425841c94bf129c4fbd1e70029c581aa1488d5f85cdd23aa0f292092146f0", "supports": ["identity"], "url": "https://api.github.com/repos/magicgh/self-map"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "8d844f227c457321640231ab3a4953457e3c6cfeab8d8f4803f8aa01dc15a8f8", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/magicgh/Self-MAP/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "f271c479b19a620600d18fc2265224205bf71483b969672d3f6acd9bc552b3f5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "Project source LICENSE; separate dataset license is not folded into the code slot.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/magicgh/Self-MAP"}], "name": "Self-MAP", "publication": null, "published_at": null, "record_sha256": "068e12a5d0245b37b269f4d22057e9b7f9137cb7ff14f51753584dfe886fa0ec", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2402.15057", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/magicgh/self-map/HEAD/README.md", "target_id": "url:bdfb50330deb97893ce4", "type": "uses"}], "resource_id": "github:magicgh/self-map", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/magicgh/Self-MAP", "topics": ["web-navigation", "memory", "self-reflection", "planning"], "updated_at": null}, "text": "Self-MAP\nResource type: framework\nReview status: curated\nScope: core\n\nA conversational web-agent implementation with candidate ranking and memory-assisted action planning, including optional memory simplification, refinement and self-reflection modules.\nPlatforms: web\nTopics: web-navigation, memory, self-reflection, planning\n\nRecorded specifications: {\"availability\": \"Public conversational planning, candidate-generation and training/evaluation source.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/magicgh/Self-MAP/HEAD/LICENSE\"], \"reason\": \"Project source LICENSE; separate dataset license is not folded into the code slot.\"}}, \"repository_revision\": \"6cf779682b70dae597626f96dfa5130b202689bc\"}\n\nThe source tree provides candidate-generation and action-prediction training/evaluation scripts. Its documented workflow requires separately downloaded MT-Mind2Web data and configured model/API paths. The README cites the ACL publication by DOI; a retained original arXiv abstract explicitly names Self-MAP and MT-Mind2Web and supplies the catalog paper relation without a new bibliography entity. MIT is the code license; the separate data artifact declares CC-BY-4.0.\n\nPrimary source: https://github.com/magicgh/Self-MAP\nCatalog record: https://awesomecua.github.io/resources/github-magicgh-self-map/"}
{"id": "github:mcgill-nlp/agent-reward-bench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/McGill-NLP/agent-reward-bench", "citation_url": "https://awesomecua.github.io/resources/github-mcgill-nlp-agent-reward-bench/", "details": {"benchmark_protocol": {"actions": ["judge verdicts"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "paper-reported expert-reviewed set", "qualifier": "exact", "unit": "trajectories", "value": 1302}], "evaluation": "Compare automatic judgments with expert answers on task success, side effects and repetitiveness.", "evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["automatic evaluator agreement", "task success", "side effects", "repetition"], "notes": "Counts refer to the paper study; current Hub package files are not converted into trajectory totals.", "observations": ["web trajectories", "screenshots", "actions"], "setting": "offline"}, "leaderboard_url": "https://huggingface.co/spaces/McGill-NLP/agent-reward-bench-leaderboard", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"], "reason": "The dataset card supplies custom Terms of Use and preserves third-party conditions. The repository has no root code license, and no dataset license is inherited from its benchmarks."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "fb753e41c3eb19be172e9e7fccbc6bcadf44288507f82b190ca99e64144294b6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "7dfac983a6e443d5f66e1fc0fc3706ebea4c93005f03457cc0a2b13a2bd89601", "supports": ["identity"], "url": "https://api.github.com/repos/McGill-NLP/agent-reward-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "79a1639cd4738719e51fc7f32313d63c0483914517cee96175459f63513e8c07", "supports": ["release_status"], "url": "https://api.github.com/repos/McGill-NLP/agent-reward-bench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:18Z", "role": "primary", "sha256": "20c910fad3ad45a0c8c39b182c53632a821d3b6b546278bea6b7a78fbcb3f6cb", "supports": ["release_status"], "url": "https://huggingface.co/api/datasets/McGill-NLP/agent-reward-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:20Z", "role": "primary", "sha256": "e16adf817e3950441d70c7bb1bb3f44bd8dc65344938247f23fea4aaff9f654e", "supports": ["licenses.data", "summary"], "url": "https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card Terms of Use, including third-party restrictions; no standard SPDX or unrestricted permission is asserted.", "declared": "Custom AgentRewardBench Terms of Use", "retrieved_at": "2026-10-17T04:03:20Z", "url": "https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/agent-reward-bench"}], "name": "AgentRewardBench", "publication": null, "published_at": null, "record_sha256": "a2945e445b25b60d8f1d2e9e0a24d01bc5160987b9bc918bb7c92ea191a2c583", "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md", "target_id": "arxiv:2504.08942", "type": "documented_by"}], "resource_id": "github:mcgill-nlp/agent-reward-bench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/McGill-NLP/agent-reward-bench", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null}, "text": "AgentRewardBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nA benchmark and evaluation toolkit for comparing automatic web-agent judges against expert judgments of task success, side effects and repetition.\nPlatforms: web\nTopics: evaluation, reward-models, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"judge verdicts\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending\", \"population\": \"paper-reported expert-reviewed set\", \"qualifier\": \"exact\", \"unit\": \"trajectories\", \"value\": 1302}], \"evaluation\": \"Compare automatic judgments with expert answers on task success, side effects and repetitiveness.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/McGill-NLP/agent-reward-bench/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"focus\": [\"automatic evaluator agreement\", \"task success\", \"side effects\", \"repetition\"], \"notes\": \"Counts refer to the paper study; current Hub package files are not converted into trajectory totals.\", \"observations\": [\"web trajectories\", \"screenshots\", \"actions\"], \"setting\": \"offline\"}, \"leaderboard_url\": \"https://huggingface.co/spaces/McGill-NLP/agent-reward-bench-leaderboard\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/McGill-NLP/agent-reward-bench/raw/main/README.md\"], \"reason\": \"The dataset card supplies custom Terms of Use and preserves third-party conditions. The repository has no root code license, and no dataset license is inherited from its benchmarks.\"}}}\n\nThe public repository contains judges, evaluation utilities and annotation files; the linked Hub hosts trajectory/screenshot manifests. No root code license is declared. Dataset Terms of Use are custom and retain third-party restrictions rather than declaring an SPDX license. The paper-reported 1,302 trajectories describe its reviewed study population, not a count recomputed from current Hub files.\n\nPrimary source: https://github.com/McGill-NLP/agent-reward-bench\nCatalog record: https://awesomecua.github.io/resources/github-mcgill-nlp-agent-reward-bench/"}
{"id": "github:mcgill-nlp/safearena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/McGill-NLP/safearena", "citation_url": "https://awesomecua.github.io/resources/github-mcgill-nlp-safearena/", "details": {"benchmark_protocol": {"actions": ["web interactions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "safe split", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2503.04957", "unit": "tasks", "value": 250}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "harmful split", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2503.04957", "unit": "tasks", "value": 250}], "evaluation": "Task outcomes and the Agent Risk Assessment categories evaluate agent behavior on benign and harmful requests.", "evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["web-agent safety", "task completion"], "notes": "Use SafeArena-specific containers rather than ordinary WebArena instances. The documented task splits require access approval; the profile does not imply unrestricted dataset access.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/McGill-NLP/safearena-leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:36:33Z", "role": "primary", "sha256": "ae50121aa60ac26b32d8d6ad6f578f9091a81a3cb9ca28b98050bddee16c929a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T09:36:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/safearena"}], "name": "SafeArena", "publication": null, "published_at": null, "record_sha256": "7a0fdfbc582169c9e6fb7c0d214d35b80c941cf3905f3be4703aacfcf3b37ce3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md", "target_id": "arxiv:2503.04957", "type": "documented_by"}], "resource_id": "github:mcgill-nlp/safearena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/McGill-NLP/safearena", "topics": ["safety-security", "evaluation"], "updated_at": null}, "text": "SafeArena\nResource type: benchmark\nReview status: curated\nScope: core\n\nWeb-agent benchmark pairing benign and harmful tasks to study deliberate misuse risks.\nPlatforms: web\nTopics: safety-security, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"web interactions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending\", \"population\": \"safe split\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2503.04957\", \"unit\": \"tasks\", \"value\": 250}, {\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending\", \"population\": \"harmful split\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2503.04957\", \"unit\": \"tasks\", \"value\": 250}], \"evaluation\": \"Task outcomes and the Agent Risk Assessment categories evaluate agent behavior on benign and harmful requests.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/McGill-NLP/safearena/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"focus\": [\"web-agent safety\", \"task completion\"], \"notes\": \"Use SafeArena-specific containers rather than ordinary WebArena instances. The documented task splits require access approval; the profile does not imply unrestricted dataset access.\", \"observations\": [], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://huggingface.co/spaces/McGill-NLP/safearena-leaderboard\"}\n\nPrimary source: https://github.com/McGill-NLP/safearena\nCatalog record: https://awesomecua.github.io/resources/github-mcgill-nlp-safearena/"}
{"id": "github:mcgill-nlp/weblinx", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/mcgill-nlp/weblinx", "citation_url": "https://awesomecua.github.io/resources/github-mcgill-nlp-weblinx/", "details": {"benchmark_protocol": {"actions": ["predicted browser actions"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "approximately", "unit": "interactions", "value": 100000}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "approximately", "unit": "expert demonstrations", "value": 2300}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "population": "original paper benchmark", "qualifier": "more-than", "unit": "websites", "value": 150}], "evaluation": "Offline evaluation compares action predictions against recorded human demonstrations.", "evidence_urls": ["https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50", "https://huggingface.co/datasets/McGill-NLP/WebLINX/raw/main/README.md"], "focus": ["conversational web navigation", "next-action prediction"], "notes": "The figures are rounded descriptions from the original paper. WebLINX 1.1 / BrowserGym removes some demonstrations and changes evaluated steps, including tab actions; do not mix those results or counts with the original release. This profile does not treat replayed demonstration steps as live task completion.", "observations": ["dialogue history", "HTML", "screenshots", "action history"], "setting": "offline"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "15c3f1656afd861894b180bca7d9214b4cf78a6ad9cd0601c80f5d67d2839528", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "15c3f1656afd861894b180bca7d9214b4cf78a6ad9cd0601c80f5d67d2839528", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:24Z", "role": "primary", "sha256": "3aac6e9748769efff3c475ede0da192776daa03431d231b5f67e2aa50f8a3be9", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:23Z", "role": "primary", "sha256": "9bd7512e257f5914d66b091530d7eb711a6f90bc21dc89cbba53015c7136c527", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/McGill-NLP/WebLINX/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "cc0156153933fb96f6f3bf8d736c15e24d78d0838945c2ed456e3ada96cf67c4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/mcgill-nlp/weblinx"}], "name": "WebLINX", "publication": null, "published_at": null, "record_sha256": "5fd394607a87373b133aa5ad5a7ab1ee7b7e978c2daec629581e85eeaf87d2bb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "resource_id": "github:mcgill-nlp/weblinx", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/mcgill-nlp/weblinx", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "WebLINX\nResource type: benchmark\nReview status: curated\nScope: core\n\nDemonstrations and evaluation for conversational web navigation.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"predicted browser actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50\", \"population\": \"original paper benchmark\", \"qualifier\": \"approximately\", \"unit\": \"interactions\", \"value\": 100000}, {\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50\", \"population\": \"original paper benchmark\", \"qualifier\": \"approximately\", \"unit\": \"expert demonstrations\", \"value\": 2300}, {\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50\", \"population\": \"original paper benchmark\", \"qualifier\": \"more-than\", \"unit\": \"websites\", \"value\": 150}], \"evaluation\": \"Offline evaluation compares action predictions against recorded human demonstrations.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/mcgill-nlp/weblinx/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2308.03688%2C2308.15272%2C2309.08172%2C2309.11436%2C2310.04716%2C2310.04869%2C2310.05126%2C2310.06770%2C2310.08740%2C2310.10634%2C2310.11441%2C2310.16042%2C2310.18331%2C2311.01767%2C2311.07562%2C2311.09278%2C2311.12983%2C2311.18751%2C2312.03003%2C2312.06677%2C2312.08914%2C2312.11671%2C2312.13108%2C2312.13771%2C2312.15820%2C2312.16062%2C2401.01614%2C2401.04124%2C2401.07781%2C2401.10935%2C2401.13178%2C2401.13649%2C2401.13919%2C2401.14268%2C2401.16158%2C2402.01030%2C2402.04476%2C2402.04615%2C2402.05930%2C2402.06596%2C2402.07456%2C2402.07939%2C2402.07945%2C2402.08567%2C2402.11941%2C2402.14073%2C2402.15057%2C2402.16965%2C2402.17553%2C2403.02713&max_results=50\", \"https://huggingface.co/datasets/McGill-NLP/WebLINX/raw/main/README.md\"], \"focus\": [\"conversational web navigation\", \"next-action prediction\"], \"notes\": \"The figures are rounded descriptions from the original paper. WebLINX 1.1 / BrowserGym removes some demonstrations and changes evaluated steps, including tab actions; do not mix those results or counts with the original release. This profile does not treat replayed demonstration steps as live task completion.\", \"observations\": [\"dialogue history\", \"HTML\", \"screenshots\", \"action history\"], \"setting\": \"offline\"}}\n\nPrimary source: https://github.com/mcgill-nlp/weblinx\nCatalog record: https://awesomecua.github.io/resources/github-mcgill-nlp-weblinx/"}
{"id": "github:mcgill-nlp/webllama", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/McGill-NLP/webllama", "citation_url": "https://awesomecua.github.io/resources/github-mcgill-nlp-webllama/", "details": {"availability": "Public fine-tuning/evaluation tooling, examples and results viewer.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE", "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md"], "reason": "README explicitly scopes MIT to code unless a file header states otherwise; models/data/images retain original licenses."}}, "repository_revision": "1b7b41327b2222312e07818ce239c04307beaf9b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "121546c7a73f4d7604ae5340bd78856e8170dc8aaa8e57f7e62efcfb85d8d4c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "6605b7b09e8be77cce9d5bfd1a6199c2256cbfc1f067b93b2b952c1d19e74f10", "supports": ["identity"], "url": "https://api.github.com/repos/McGill-NLP/webllama"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "915620d7a155f5c5fe099cdb0461d9614e1894b38559f0d2f9a6a138afe06212", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/McGill-NLP/webllama/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "e69268828e67ffd1d3393a2d3530f97df6deaa0b961039fc971abea28cec8715", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "README explicitly scopes MIT to code unless a file header states otherwise; models/data/images retain original licenses.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/McGill-NLP/webllama"}], "name": "WebLlama", "publication": null, "published_at": null, "record_sha256": "9081de3098e27aa7b3e6928c6ad96b07e23b2f2adf35dbc31b9a19605a2bd1c0", "relations": [{"evidence_url": "https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "resource_id": "github:mcgill-nlp/webllama", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/McGill-NLP/webllama", "topics": ["web-navigation", "imitation-learning", "evaluation", "tool-integration"], "updated_at": null}, "text": "WebLlama\nResource type: framework\nReview status: curated\nScope: core\n\nTraining, evaluation and integration tooling for dialogue-guided web agents, including Llama fine-tuning workflows and BrowserGym examples built around WebLINX.\nPlatforms: web\nTopics: web-navigation, imitation-learning, evaluation, tool-integration\n\nRecorded specifications: {\"availability\": \"Public fine-tuning/evaluation tooling, examples and results viewer.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/LICENSE\", \"https://raw.githubusercontent.com/McGill-NLP/webllama/HEAD/README.md\"], \"reason\": \"README explicitly scopes MIT to code unless a file header states otherwise; models/data/images retain original licenses.\"}}, \"repository_revision\": \"1b7b41327b2222312e07818ce239c04307beaf9b\"}\n\nThe manifest contains modeling/training code, BrowserGym examples, documentation and a results viewer. The README distinguishes currently available components from planned additional benchmarks and deployment integrations. It requests citation of WebLINX as the foundation for data, training and evaluation; this is not a claim that the WebLINX paper first introduced every later WebLlama component. MIT applies to repository code, subject to file-header exceptions; weights and datasets have their own terms.\n\nPrimary source: https://github.com/McGill-NLP/webllama\nCatalog record: https://awesomecua.github.io/resources/github-mcgill-nlp-webllama/"}
{"id": "github:meituan/evocua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/meituan/EvoCUA", "citation_url": "https://awesomecua.github.io/resources/github-meituan-evocua/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "b00a072dfde61f774db6acf94fcde9bd32337b1b4f16a178998378d842bbbe4d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/meituan/EvoCUA"}], "name": "EvoCUA", "publication": null, "published_at": null, "record_sha256": "807eed8cfdf155a9ecded21f9079bddc22e84e3a0500555750bdec7a42ca3ed7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/meituan/EvoCUA/HEAD/README.md", "target_id": "arxiv:2601.15876", "type": "documented_by"}], "resource_id": "github:meituan/evocua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/meituan/EvoCUA", "topics": ["reinforcement-learning", "planning"], "updated_at": null}, "text": "EvoCUA\nResource type: model\nReview status: curated\nScope: core\n\nComputer-use model and training resources from the EvoCUA project.\nPlatforms: desktop\nTopics: reinforcement-learning, planning\n\nPrimary source: https://github.com/meituan/EvoCUA\nCatalog record: https://awesomecua.github.io/resources/github-meituan-evocua/"}
{"id": "github:microsoft/fara", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/fara", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-fara/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "ce11bdfe8354eb9c3b8a985dd5ddd21f88f94fb5fb9d00c9b01a76226f0b0a36", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/fara"}], "name": "Fara", "publication": null, "published_at": null, "record_sha256": "3f87935b9df483ce48f37107b9497ae96c4064a8715ab21b31f94808a20271eb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md", "target_id": "arxiv:2511.19663", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/microsoft/fara/HEAD/README.md", "target_id": "arxiv:2606.20785", "type": "documented_by"}], "resource_id": "github:microsoft/fara", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/fara", "topics": ["planning", "web-navigation"], "updated_at": null}, "text": "Fara\nResource type: model\nReview status: curated\nScope: core\n\nMicrosoft browser-use agent models that act from screenshots by predicting click coordinates without accessibility trees; Fara1.5 comes in 4B, 9B and 27B, succeeding Fara-7B.\nPlatforms: web\nTopics: planning, web-navigation\n\nPrimary source: https://github.com/microsoft/fara\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-fara/"}
{"id": "github:microsoft/gui-actor", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/GUI-Actor", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-gui-actor/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/GUI-Actor"}], "name": "GUI-Actor", "publication": null, "published_at": null, "record_sha256": "3d1993d3009328cf6bb31ced907e7bf9e790a5b92a34e58732c116b9bf5419b5", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "resource_id": "github:microsoft/gui-actor", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/GUI-Actor", "topics": ["grounding"], "updated_at": null}, "text": "GUI-Actor\nResource type: model\nReview status: curated\nScope: core\n\nGUI grounding model with attention-based action localization.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/microsoft/GUI-Actor\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-gui-actor/"}
{"id": "github:microsoft/gui-agent-rl", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/GUI-Agent-RL", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-gui-agent-rl/", "details": {"availability": "Public training and evaluation source; trained checkpoints are generated by the documented workflow.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"], "reason": "Repository LICENSE grants MIT for code; linked training imagery and pretrained model licenses remain separate."}}, "repository_revision": "6d2381ef549d980452a828471221a2ce8023bf41"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "d21a1ae4359296c0f3034ac655888dce43448d10afe0b1909d6b6520678c2526", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:13Z", "role": "primary", "sha256": "791c6e56e62ce5e803a4ba313557bac0a39dfeba27499a3865ad444a860b426a", "supports": ["identity"], "url": "https://api.github.com/repos/microsoft/GUI-Agent-RL"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "5b799a5bdae78ff50652c0b464b6cad12d9dda265d4752e5d5f9eab65e9d5c6f", "supports": ["release_status"], "url": "https://api.github.com/repos/microsoft/GUI-Agent-RL/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:06Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T09:12:14Z", "licenses": {"code": {"basis": "Repository LICENSE grants MIT for code; linked training imagery and pretrained model licenses remain separate.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:06Z", "url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/GUI-Agent-RL"}], "name": "VEM GUI-Agent-RL", "publication": null, "published_at": null, "record_sha256": "974b48b7189af1c7ea637c535f8df475702a8b30945f5791bca0d69ecbbac10c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/README.md", "target_id": "arxiv:2502.18906", "type": "documented_by"}], "resource_id": "github:microsoft/gui-agent-rl", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/GUI-Agent-RL", "topics": ["reinforcement-learning", "reward-models", "mobile-navigation"], "updated_at": null}, "text": "VEM GUI-Agent-RL\nResource type: framework\nReview status: curated\nScope: core\n\nTraining code that learns a value environment model from offline Android interaction data and uses its frozen value estimates to guide GUI policy optimization.\nPlatforms: android\nTopics: reinforcement-learning, reward-models, mobile-navigation\n\nRecorded specifications: {\"availability\": \"Public training and evaluation source; trained checkpoints are generated by the documented workflow.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/microsoft/GUI-Agent-RL/HEAD/LICENSE\"], \"reason\": \"Repository LICENSE grants MIT for code; linked training imagery and pretrained model licenses remain separate.\"}}, \"repository_revision\": \"6d2381ef549d980452a828471221a2ce8023bf41\"}\n\nThe public tree includes critic-training configurations, policy training and offline/online evaluation scripts. The README obtains base models and AITW imagery separately, then instructs users to train critic and policy checkpoints locally; this entry does not claim that VEM-trained weights are released. MIT applies to repository code only.\n\nPrimary source: https://github.com/microsoft/GUI-Agent-RL\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-gui-agent-rl/"}
{"id": "github:microsoft/magentic-ui", "metadata": {"aliases": ["Magentic UI", "Magentic-UI", "MagenticLite"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/magentic-ui", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-magentic-ui/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "240e0c92b86c02d3790c0784d8c72193ffb4fd5f9c2b0ec618ef8121fe216b8b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "98a96c38494df4963951e32b81ec4effebe4c8a812ab8d9e7d185d59e86fe2fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/microsoft/magentic-ui/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/magentic-ui"}], "name": "MagenticLite / Magentic UI", "publication": null, "published_at": null, "record_sha256": "4e01c645ad40808edc69f1540245e9ddf173febcd34f978054e6451b08d32160", "relations": [], "resource_id": "github:microsoft/magentic-ui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/magentic-ui", "topics": ["human-in-the-loop"], "updated_at": null}, "text": "MagenticLite / Magentic UI\nResource type: agent\nReview status: curated\nScope: core\n\nMicrosoft agent application combining orchestration with browser interaction; successor to Magentic UI.\nRecorded aliases: Magentic UI, Magentic-UI, MagenticLite\nPlatforms: web\nTopics: human-in-the-loop\n\nPrimary source: https://github.com/microsoft/magentic-ui\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-magentic-ui/"}
{"id": "github:microsoft/magma", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/Magma", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-magma/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "8aaf8feff6e1d5a533c01afd8afc0cfc0700f0c1c73ba47949b56c11fb8ab5be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T17:18:56Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T17:18:56Z", "url": "https://raw.githubusercontent.com/microsoft/Magma/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Magma"}], "name": "Magma", "publication": null, "published_at": null, "record_sha256": "8c1c9e81bfbf57a7dfb82374d297e26ee6a41084403e1019e6594611f5fec52b", "relations": [], "resource_id": "github:microsoft/magma", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/Magma", "topics": ["grounding", "planning", "mobile-navigation"], "updated_at": null}, "text": "Magma\nResource type: model\nReview status: curated\nScope: core\n\nMicrosoft foundation model for multimodal agents, pretrained with Set-of-Mark and Trace-of-Mark objectives on UI, robotics and video data and fine-tunable for web and mobile UI navigation.\nPlatforms: web\nTopics: grounding, planning, mobile-navigation\n\nPrimary source: https://github.com/microsoft/Magma\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-magma/"}
{"id": "github:microsoft/omniparser", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/OmniParser", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-omniparser/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/OmniParser"}], "name": "OmniParser", "publication": null, "published_at": null, "record_sha256": "0c7cc75271e35816141aa4951d6aa3be1da46accc2029ef7bc94ebd3cc571742", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md", "target_id": "arxiv:2408.00203", "type": "documented_by"}], "resource_id": "github:microsoft/omniparser", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/OmniParser", "topics": ["screen-understanding", "grounding"], "updated_at": null}, "text": "OmniParser\nResource type: tool\nReview status: curated\nScope: core\n\nScreenshot parser that detects interactable regions and captions icons so vision-language models can ground actions, released as V1.5 and V2 weights alongside OmniTool for driving a Windows 11 VM.\nPlatforms: cross-platform\nTopics: screen-understanding, grounding\n\nPrimary source: https://github.com/microsoft/OmniParser\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-omniparser/"}
{"id": "github:microsoft/phi-ground", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/Phi-Ground", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-phi-ground/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Phi-Ground"}], "name": "Phi-Ground", "publication": null, "published_at": null, "record_sha256": "9fbf846af4cd8f51e07d8028e6bc4175dee48207d06f06b18fc86126a9da9515", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "arxiv:2507.23779", "type": "documented_by"}], "resource_id": "github:microsoft/phi-ground", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/Phi-Ground", "topics": ["grounding"], "updated_at": null}, "text": "Phi-Ground\nResource type: model\nReview status: curated\nScope: core\n\nMicrosoft GUI grounding models from an empirical study of grounding training, from data collection to model training, released as Phi-Ground-4B-7C with evaluation code.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/microsoft/Phi-Ground\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-phi-ground/"}
{"id": "github:microsoft/playwright", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/playwright", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-playwright/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "0a6a9dddcbf1295b61e83462ee7aa130289a4d4e60bac1a0deccd661f8fc8146", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "45873d00a0dd243596deb4aa23b2493b3d1f0671921bf2538ea431d7380220eb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/microsoft/playwright/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/playwright"}], "name": "Playwright", "publication": null, "published_at": null, "record_sha256": "fcb13e587b159ba0748ec69e65c4953042374b6a089871897e5a354d97fbefd0", "relations": [], "resource_id": "github:microsoft/playwright", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/microsoft/playwright", "topics": ["automation"], "updated_at": null}, "text": "Playwright\nResource type: infrastructure\nReview status: curated\nScope: adjacent\n\nBrowser automation library used as an execution layer by agents.\nPlatforms: web\nTopics: automation\n\nPrimary source: https://github.com/microsoft/playwright\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-playwright/"}
{"id": "github:microsoft/playwright-mcp", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/playwright-mcp", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-playwright-mcp/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "b33aa71a13d655e96e6b736d0c9de1e5ddea52f205916ca20b876387ff830e87", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "9a7110fc2d2f964038e5dc49128f908f29f47a574c961cba16085914e879cbda", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/microsoft/playwright-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/playwright-mcp"}], "name": "Playwright MCP", "publication": null, "published_at": null, "record_sha256": "21ba927e6731470594fe964a4c803dc3a5fe171caa6a0d9323e0e60d453bc92d", "relations": [], "resource_id": "github:microsoft/playwright-mcp", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/playwright-mcp", "topics": ["automation", "tool-integration"], "updated_at": null}, "text": "Playwright MCP\nResource type: tool\nReview status: curated\nScope: core\n\nMCP server that lets LLM clients drive Playwright browsers through structured accessibility snapshots instead of screenshots, with opt-in vision, PDF and devtools capabilities.\nPlatforms: web\nTopics: automation, tool-integration\n\nPrimary source: https://github.com/microsoft/playwright-mcp\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-playwright-mcp/"}
{"id": "github:microsoft/precision-cua-bench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/precision-cua-bench", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-precision-cua-bench/", "details": {"benchmark_protocol": {"actions": ["pixel (x,y) cursor coordinate"], "counts": [], "evaluation": "The inspected evaluator scales target boxes from a 1000×1000 reference grid into image pixels and tests pixel predictions, allowing ±10 pixels around the center along any target dimension narrower than 1 pixel after scaling. It reports cumulative hit accuracy by turn (earlier hits remain correct), miss distances, any-turn hit rate, and a correction rate whose denominator is all successfully processed samples. The default budget is five turns; retain prompt/feedback variant, turn budget, and processed/skipped coverage.", "evidence_urls": ["https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md", "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py", "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"], "focus": ["pixel-precise editor cursor placement", "multi-turn visual self-correction"], "notes": "The release has four editor/theme JSONL partitions (Cursor/VS Code × dark/light) and screenshot files. Partitions are not train/test splits or independent task totals. README’s initial field table calls bbox pixel coordinates, but the evaluation-format section and actual ratio_h/ratio_w=dimension/1000 code use normalized target boxes; profile follows the inspected evaluator and retains the inconsistency. Source/images were not executed or downloaded as a dataset. MIT is recorded for repository code only.", "observations": ["code-editor screenshot", "cursor-location instruction", "red-cross feedback marking previous incorrect prediction"], "setting": "offline"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "23d0dc8ea0a10c080cf8ae32121eade2785ae923a6b0d853d166c94c195e789b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://github.com/microsoft/precision-cua-bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "84ad4fb798abb776dce813cf4dab296d6e03248d8ae29388e1f06cccd27d1241", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:08Z", "role": "primary", "sha256": "1fda7c3980fe7a7c373478eda6ecfabff227e92c2b1f19c6fe903b408f583797", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:06Z", "role": "primary", "sha256": "ede84e3dda47f80fada4234caa2423db08e21ff7667a37599c03aae657b2f84b", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:32:00Z", "role": "primary", "sha256": "4d09ebfb093e43a4c44adef86c8fa8374f90d907fc33bd8e12a1315e62d5a583", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/precision-cua-bench"}], "name": "Precision CUA Bench", "publication": null, "published_at": null, "record_sha256": "c635c585d3fc0d6784288a9f3228e6e424298e1fb185042217051bbcd8d35c52", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2603.12823%2C2603.14707%2C2603.15039%2C2603.17357%2C2603.18429%2C2603.19191%2C2603.24440%2C2603.25226%2C2603.25864%2C2603.26211%2C2603.26648%2C2603.29318%2C2604.01535%2C2604.02947%2C2604.04399%2C2604.05477%2C2604.06126%2C2604.06182%2C2604.06367%2C2604.07429%2C2604.08455%2C2604.08516%2C2604.08523%2C2604.09155%2C2604.09574%2C2604.09587%2C2604.09815%2C2604.09937%2C2604.11259%2C2604.11784%2C2604.12284%2C2604.13019%2C2604.13822%2C2604.14113%2C2604.14262%2C2604.14956%2C2604.15093%2C2604.15309%2C2604.17817%2C2604.17821%2C2604.18860%2C2604.24964%2C2604.25380%2C2604.25562%2C2604.27151%2C2604.27253%2C2604.27776%2C2605.00642%2C2605.06393%2C2605.06664&max_results=50", "target_id": "arxiv:2604.13019", "type": "documented_by"}], "resource_id": "github:microsoft/precision-cua-bench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/precision-cua-bench", "topics": ["evaluation", "grounding"], "updated_at": null}, "text": "Precision CUA Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation resources for pixel-precise cursor grounding in code editors.\nPlatforms: cross-platform\nTopics: evaluation, grounding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"pixel (x,y) cursor coordinate\"], \"counts\": [], \"evaluation\": \"The inspected evaluator scales target boxes from a 1000×1000 reference grid into image pixels and tests pixel predictions, allowing ±10 pixels around the center along any target dimension narrower than 1 pixel after scaling. It reports cumulative hit accuracy by turn (earlier hits remain correct), miss distances, any-turn hit rate, and a correction rate whose denominator is all successfully processed samples. The default budget is five turns; retain prompt/feedback variant, turn budget, and processed/skipped coverage.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/README.md\", \"https://raw.githubusercontent.com/microsoft/precision-cua-bench/HEAD/eval/test_allturns_api.py\", \"https://api.github.com/repos/microsoft/precision-cua-bench/git/trees/main?recursive=1\"], \"focus\": [\"pixel-precise editor cursor placement\", \"multi-turn visual self-correction\"], \"notes\": \"The release has four editor/theme JSONL partitions (Cursor/VS Code × dark/light) and screenshot files. Partitions are not train/test splits or independent task totals. README’s initial field table calls bbox pixel coordinates, but the evaluation-format section and actual ratio_h/ratio_w=dimension/1000 code use normalized target boxes; profile follows the inspected evaluator and retains the inconsistency. Source/images were not executed or downloaded as a dataset. MIT is recorded for repository code only.\", \"observations\": [\"code-editor screenshot\", \"cursor-location instruction\", \"red-cross feedback marking previous incorrect prediction\"], \"setting\": \"offline\"}}\n\nPrimary source: https://github.com/microsoft/precision-cua-bench\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-precision-cua-bench/"}
{"id": "github:microsoft/som", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/SoM", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-som/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "b3fa8b7e24a01da023a5f21023c4853a90388dd7744055fbb5a2888938c68b0c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/SoM"}], "name": "Set of Mark", "publication": null, "published_at": null, "record_sha256": "fae8608aeea3d23891427e1231fd717f7d89155197ccdaf70de6db5abe858b84", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/SoM/HEAD/README.md", "target_id": "arxiv:2310.11441", "type": "documented_by"}], "resource_id": "github:microsoft/som", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/microsoft/SoM", "topics": ["grounding"], "updated_at": null}, "text": "Set of Mark\nResource type: tool\nReview status: curated\nScope: adjacent\n\nVisual-prompting toolbox that labels image regions for multimodal reference and grounding, including GUI applications.\nPlatforms: unspecified\nTopics: grounding\n\nPrimary source: https://github.com/microsoft/SoM\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-som/"}
{"id": "github:microsoft/ufo", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/UFO", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-ufo/", "details": {}, "environments": ["windows", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "40d9a4010ab122ab50242a0cd7a5320e2fd16b3e53494d988eecdbd94fd8e6e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/UFO/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/UFO"}], "name": "UFO", "publication": null, "published_at": null, "record_sha256": "dddd5dd8b5feb2ae1216cb85aca44e62eb4e640e5e00e73338afb2c79106fb18", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/UFO/HEAD/README.md", "target_id": "arxiv:2402.07939", "type": "documented_by"}], "resource_id": "github:microsoft/ufo", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/UFO", "topics": ["planning", "cross-app"], "updated_at": null}, "text": "UFO\nResource type: agent\nReview status: curated\nScope: core\n\nWindows agent framework for application and desktop automation.\nPlatforms: windows, desktop\nTopics: planning, cross-app\n\nPrimary source: https://github.com/microsoft/UFO\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-ufo/"}
{"id": "github:microsoft/webwright", "metadata": {"aliases": ["Webwright", "CUAWright"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/Webwright", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-webwright/", "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "7c0f87f684e5ed4d77b175b63a5dcaf0a268644c3662c723713de4cf0ac1de40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c2cfccb812fe482101a8f04597dfc5a9991a6b2748266c47ac91b6a5aae15383", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/Webwright"}], "name": "CUAWright (formerly Webwright)", "publication": null, "published_at": null, "record_sha256": "bcca465d8950dcd0d215453b70512119dff2b887a0e4ccbb082502eb12c02a37", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Webwright/HEAD/README.md", "target_id": "arxiv:2610.04116", "type": "documented_by"}], "resource_id": "github:microsoft/webwright", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/Webwright", "topics": ["tool-integration"], "updated_at": null}, "text": "CUAWright (formerly Webwright)\nResource type: framework\nReview status: curated\nScope: core\n\nAgent framework for browser and desktop tasks using automation code.\nRecorded aliases: Webwright, CUAWright\nPlatforms: web, desktop\nTopics: tool-integration\n\nPrimary source: https://github.com/microsoft/Webwright\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-webwright/"}
{"id": "github:microsoft/windowsagentarena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/microsoft/WindowsAgentArena", "citation_url": "https://awesomecua.github.io/resources/github-microsoft-windowsagentarena/", "details": {"benchmark_protocol": {"actions": ["agent-defined Windows interactions"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json", "population": "evaluation_examples_windows/test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 154}], "evaluation": "Task configuration selects an evaluator and expected application or system outcome; evaluators return a reward for the resulting state.", "evidence_urls": ["https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md", "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/docs/Develop-Tasks.md", "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json"], "focus": ["Windows task completion", "desktop workflows"], "notes": "Counted unique identifiers in the inspected manifest. Normal and hard modes differ in how much task setup the agent must perform. Record the difficulty and observation/backend configuration with each result; no VM was launched.", "observations": ["screenshots", "accessibility information", "screen-element detections"], "setting": "interactive"}}, "environments": ["windows", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e481abd422a4cc8b819ad84b32684e4af641e18a7c0d277dd608cfd5d3c5e6ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e481abd422a4cc8b819ad84b32684e4af641e18a7c0d277dd608cfd5d3c5e6ad", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "c565dc02ab0049bf9710b87e2760dda8d7c92268ea78f0354952695ab59c0de3", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/docs/Develop-Tasks.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "902fb24cf0a9e146bf6c29288e74a7254fe40294e49dd47363e0b15d65608fef", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "9906940f61b1f0b533fa7d99baf55178b2808fbe113ea51dfbfad8572ccd5f2b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/microsoft/WindowsAgentArena"}], "name": "Windows Agent Arena", "publication": null, "published_at": null, "record_sha256": "3cfaa0c768f90be0182bc7b052d05716abafd387c797ad97881491ce9180c4b7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md", "target_id": "arxiv:2409.08264", "type": "documented_by"}], "resource_id": "github:microsoft/windowsagentarena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/microsoft/WindowsAgentArena", "topics": ["evaluation"], "updated_at": null}, "text": "Windows Agent Arena\nResource type: benchmark\nReview status: curated\nScope: core\n\nWindows 11 benchmark for multimodal desktop agents in a reproducible VM that runs locally or in parallel on Azure ML, with a harder mode where agents must set tasks up themselves.\nPlatforms: windows, desktop\nTopics: evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"agent-defined Windows interactions\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json\", \"population\": \"evaluation_examples_windows/test_all.json\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 154}], \"evaluation\": \"Task configuration selects an evaluator and expected application or system outcome; evaluators return a reward for the resulting state.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/README.md\", \"https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/docs/Develop-Tasks.md\", \"https://raw.githubusercontent.com/microsoft/WindowsAgentArena/HEAD/src/win-arena-container/client/evaluation_examples_windows/test_all.json\"], \"focus\": [\"Windows task completion\", \"desktop workflows\"], \"notes\": \"Counted unique identifiers in the inspected manifest. Normal and hard modes differ in how much task setup the agent must perform. Record the difficulty and observation/backend configuration with each result; no VM was launched.\", \"observations\": [\"screenshots\", \"accessibility information\", \"screen-element detections\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/microsoft/WindowsAgentArena\nCatalog record: https://awesomecua.github.io/resources/github-microsoft-windowsagentarena/"}
{"id": "github:minorjerry/webvoyager", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MinorJerry/WebVoyager", "citation_url": "https://awesomecua.github.io/resources/github-minorjerry-webvoyager/", "details": {"benchmark_protocol": {"actions": ["Selenium-backed browser interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "WebVoyager task collection", "qualifier": "exact", "unit": "task queries", "value": 643}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "WebVoyager task collection", "qualifier": "exact", "unit": "websites", "value": 15}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "population": "separate GAIA validation-derived web subset", "qualifier": "exact", "unit": "tasks", "value": 90}], "evaluation": "The repository supplies an automated multimodal-model evaluation protocol.", "evidence_urls": ["https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md"], "focus": ["live-web task completion"], "notes": "The GAIA-derived subset is separate from the 643 WebVoyager queries. Live sites and time-sensitive travel tasks can change; record the task snapshot and evaluation date.", "observations": ["webpage screenshots", "textual task instructions"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "b5523941b2773634fdcf359b5b89bf0aca589f33ad9db30692c8b1912ed5dda5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MinorJerry/WebVoyager"}], "name": "WebVoyager", "publication": null, "published_at": null, "record_sha256": "4bb7a5de5b071d8aacf754be4545f45249e7988c70aa7f5f6dd41006db23ed1f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md", "target_id": "arxiv:2401.13919", "type": "documented_by"}], "resource_id": "github:minorjerry/webvoyager", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MinorJerry/WebVoyager", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "WebVoyager\nResource type: benchmark\nReview status: curated\nScope: core\n\nMultimodal web navigation agent and evaluation tasks on real websites.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"Selenium-backed browser interactions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md\", \"population\": \"WebVoyager task collection\", \"qualifier\": \"exact\", \"unit\": \"task queries\", \"value\": 643}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md\", \"population\": \"WebVoyager task collection\", \"qualifier\": \"exact\", \"unit\": \"websites\", \"value\": 15}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md\", \"population\": \"separate GAIA validation-derived web subset\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 90}], \"evaluation\": \"The repository supplies an automated multimodal-model evaluation protocol.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/MinorJerry/WebVoyager/HEAD/README.md\"], \"focus\": [\"live-web task completion\"], \"notes\": \"The GAIA-derived subset is separate from the 643 WebVoyager queries. Live sites and time-sensitive travel tasks can change; record the task snapshot and evaluation date.\", \"observations\": [\"webpage screenshots\", \"textual task instructions\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/MinorJerry/WebVoyager\nCatalog record: https://awesomecua.github.io/resources/github-minorjerry-webvoyager/"}
{"id": "github:mobile-next/mobile-mcp", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/mobile-next/mobile-mcp", "citation_url": "https://awesomecua.github.io/resources/github-mobile-next-mobile-mcp/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "1.0.11"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "bc24fa665d1a49fa806d58ec569d93e134058e0cb1a7bc9324339ba0e35ff107", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "80e8ad254f71f5c80b91a6b80edd09f96116ef4967ae4e6832e5960357f7b99a", "supports": ["identity"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "49684448428c28afd4bf191f695c3d45fbe46910fa93951d2a02f6b8afcf9f16", "supports": ["release_status"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:57Z", "role": "primary", "sha256": "43a5c02407afe86b6f1fedb4299f28420778ce21825cd87516deafe05507a7bc", "supports": ["release_status"], "url": "https://api.github.com/repos/mobile-next/mobile-mcp/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:58Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:00:58Z", "url": "https://raw.githubusercontent.com/mobile-next/mobile-mcp/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/mobile-next/mobile-mcp"}], "name": "Mobile Next MCP", "publication": null, "published_at": null, "record_sha256": "61d10a119c1946579de41dcbae24e998327eb3c9c76a3919a95f893e7fe37914", "relations": [], "resource_id": "github:mobile-next/mobile-mcp", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/mobile-next/mobile-mcp", "topics": ["mobile-navigation", "tool-integration", "automation"], "updated_at": null}, "text": "Mobile Next MCP\nResource type: tool\nReview status: curated\nScope: core\n\nAn MCP server exposing mobile accessibility snapshots, screenshots and device interactions across real devices, simulators and emulators.\nPlatforms: android, ios\nTopics: mobile-navigation, tool-integration, automation\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"1.0.11\"}\n\nSource and release 1.0.11 are public. The README specifies iOS and Android device setup and supports local devices as well as a separate cloud offering. Apache-2.0 is declared for this code; no cloud-service pricing or permissions are inferred.\n\nPrimary source: https://github.com/mobile-next/mobile-mcp\nCatalog record: https://awesomecua.github.io/resources/github-mobile-next-mobile-mcp/"}
{"id": "github:mobileagentbench/mobile-agent-bench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MobileAgentBench/mobile-agent-bench", "citation_url": "https://awesomecua.github.io/resources/github-mobileagentbench-mobile-agent-bench/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e1d75e8a196f19af6050c53694383f069561244c810ca259c5f8075ee91f72b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MobileAgentBench/mobile-agent-bench/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileAgentBench/mobile-agent-bench"}], "name": "MobileAgentBench", "publication": null, "published_at": null, "record_sha256": "bbcb9c79133af390765b1bb5e744fa0405196ad1a8db6715030d2f5b351f0c54", "relations": [], "resource_id": "github:mobileagentbench/mobile-agent-bench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MobileAgentBench/mobile-agent-bench", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "MobileAgentBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nAutomated benchmark for mobile LLM agents whose default tasks use ten Simple Mobile Tools Android apps, installed as a Python library that detects task completion.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nPrimary source: https://github.com/MobileAgentBench/mobile-agent-bench\nCatalog record: https://awesomecua.github.io/resources/github-mobileagentbench-mobile-agent-bench/"}
{"id": "github:mobilellm/agentprog", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MobileLLM/AgentProg", "citation_url": "https://awesomecua.github.io/resources/github-mobilellm-agentprog/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "ef5e2998ac2defa86e0dc4c13e7e00d595e1e7cf", "repository_url": "https://github.com/MobileLLM/AgentProg"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "becc4961d93f0170fab29120afa4b09b2999f2cd0e74b13ccb7f28cb9623761b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:01Z", "role": "primary", "sha256": "d14eecf7c041933cefecc03b62bbea45c1617a183fd8c4ec1334e22a644a7f67", "supports": ["identity"], "url": "https://api.github.com/repos/MobileLLM/AgentProg"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:22Z", "role": "primary", "sha256": "dd1cf9cd0aed31af63d7046ffa98d1be2c36258706951173ad8c28a64fcf9f7a", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/MobileLLM/AgentProg/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "751123fccb2fd04b5c08b8c712a68a8bc69f56391593711efbfa533eaebbdb10", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AgentProg"}], "name": "AgentProg", "publication": null, "published_at": null, "record_sha256": "9b790e765074e142f4dce7d07c260a1eec0d6f43ecaf2acfcd32bd925a879dc2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/README.md", "target_id": "arxiv:2512.10371", "type": "documented_by"}], "resource_id": "github:mobilellm/agentprog", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MobileLLM/AgentProg", "topics": ["long-horizon", "memory", "planning", "state-tracking"], "updated_at": null}, "text": "AgentProg\nResource type: framework\nReview status: curated\nScope: core\n\nAndroid GUI-agent framework that keeps long-running task context in structured programs, variables and control flow, with a CLI and AndroidWorld integration.\nPlatforms: android\nTopics: long-horizon, memory, planning, state-tracking\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/MobileLLM/AgentProg/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"ef5e2998ac2defa86e0dc4c13e7e00d595e1e7cf\", \"repository_url\": \"https://github.com/MobileLLM/AgentProg\"}\n\nPublic tree contains the agentprog package, CLI, program/context-management code and AndroidWorld evaluation integration. README announces AW-Extend, but this record is the framework rather than a separately audited benchmark release. External model services and Android device/emulator setup are required. Apache-2.0 covers code, not external model weights, apps or VM images.\n\nPrimary source: https://github.com/MobileLLM/AgentProg\nCatalog record: https://awesomecua.github.io/resources/github-mobilellm-agentprog/"}
{"id": "github:mobilellm/autodroid", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MobileLLM/AutoDroid", "citation_url": "https://awesomecua.github.io/resources/github-mobilellm-autodroid/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "4a79eb2e14b6da67ded536c127d41f865d5b43620d60d0d950d291c8956564fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AutoDroid"}], "name": "AutoDroid", "publication": null, "published_at": null, "record_sha256": "ca625b209ac86260ce124e4d93214e684f55dbee776becba7a13e7e9192bbc8d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid/HEAD/README.md", "target_id": "arxiv:2308.15272", "type": "documented_by"}], "resource_id": "github:mobilellm/autodroid", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MobileLLM/AutoDroid", "topics": ["mobile-navigation"], "updated_at": null}, "text": "AutoDroid\nResource type: agent\nReview status: curated\nScope: core\n\nLLM-powered Android task automation built on DroidBot, released with DroidTask, a dataset of per-app screenshots, view hierarchies, exploration records and ground-truth task files.\nPlatforms: android\nTopics: mobile-navigation\n\nPrimary source: https://github.com/MobileLLM/AutoDroid\nCatalog record: https://awesomecua.github.io/resources/github-mobilellm-autodroid/"}
{"id": "github:mobilellm/autodroid-v2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MobileLLM/AutoDroid-V2", "citation_url": "https://awesomecua.github.io/resources/github-mobilellm-autodroid-v2/", "details": {"availability": "Public source and minimal-experiment documentation are released; default README path is absent.", "repository_revision": "b1a02a52a6c94b96f63016ac12d3a4ab6a410b8e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "74508100843f8db6db6bb151e77ac1cc604a6b7ebd16d3225266d73434465136", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/README_minimal.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:10Z", "role": "primary", "sha256": "f349963f2ea3f7b81f1c04ee875654382929b1f93d99d0168d6e086155f41d72", "supports": ["identity"], "url": "https://api.github.com/repos/MobileLLM/AutoDroid-V2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:03Z", "role": "primary", "sha256": "70354b856b75bc4a2aca0a7274321637039ec5d083523ca2f677cd11d475b386", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/MobileLLM/AutoDroid-V2/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "8ba37983a2c2da53d96a52696c07e4bdbea4e987a1a07b557935d614c3be46d2", "supports": ["summary", "details", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "cfed016e038d09ce7c9e829273fffa4a6c618bf8e9043312e4690fbe1f40d247", "supports": ["environments", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_0_installation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "f5534771f68770ed2e45daf94fbf80304ba2904387571c34c5d80a152bef0a00", "supports": ["summary"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_1_doc_generation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:59Z", "role": "primary", "sha256": "0564f566fde97360496bea4a6eb023c4c84e03cdfdf5201c5b27235c36e32b18", "supports": ["summary", "details"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_4_accuracy_validation/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T09:14:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MobileLLM/AutoDroid-V2"}], "name": "AutoDroid V2", "publication": null, "published_at": null, "record_sha256": "920be8db5ba8bdae1f52adb9e8e00e501e1a76e25c1d460aa8433ef69e7cd927", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2412.18116", "type": "documented_by"}], "resource_id": "github:mobilellm/autodroid-v2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MobileLLM/AutoDroid-V2", "topics": ["mobile-navigation", "tool-integration", "data-generation"], "updated_at": null}, "text": "AutoDroid V2\nResource type: agent\nReview status: curated\nScope: core\n\nAn Android automation implementation that builds app documentation from exploration traces and evaluates language-model-generated execution code in emulator workflows.\nPlatforms: android\nTopics: mobile-navigation, tool-integration, data-generation\n\nRecorded specifications: {\"availability\": \"Public source and minimal-experiment documentation are released; default README path is absent.\", \"repository_revision\": \"b1a02a52a6c94b96f63016ac12d3a4ab6a410b8e\"}\n\nThe repository uses README_minimal.md rather than README.md and includes a documented App Launcher experiment, training setup and DroidTask/LlamaTouch evaluation scripts. Its supplied emulator setup targets x86_64 Linux and Android API 31. No repository-level code license was found; code, APK, data and weight terms are not inferred from one another. Dependencies, external evaluation files and model service are still required; no software was run.\n\nPrimary source: https://github.com/MobileLLM/AutoDroid-V2\nCatalog record: https://awesomecua.github.io/resources/github-mobilellm-autodroid-v2/"}
{"id": "github:murtyshikhar/nnetnav", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/MurtyShikhar/NNetnav", "citation_url": "https://awesomecua.github.io/resources/github-murtyshikhar-nnetnav/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/MurtyShikhar/NNetnav"}], "name": "NNetNav", "publication": null, "published_at": null, "record_sha256": "27fb211b9e90a73c55fc563c4504a4b08e9b7109eb6141a6e6076ee6351b9e7d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "arxiv:2410.02907", "type": "documented_by"}], "resource_id": "github:murtyshikhar/nnetnav", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/MurtyShikhar/NNetnav", "topics": ["data-generation", "web-navigation"], "updated_at": null}, "text": "NNetNav\nResource type: framework\nReview status: curated\nScope: core\n\nCollects and labels browser-exploration trajectories for agent training, with code for data processing, fine-tuning, and web evaluation.\nPlatforms: web\nTopics: data-generation, web-navigation\n\nPrimary source: https://github.com/MurtyShikhar/NNetnav\nCatalog record: https://awesomecua.github.io/resources/github-murtyshikhar-nnetnav/"}
{"id": "github:nanobrowser/nanobrowser", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/nanobrowser/nanobrowser", "citation_url": "https://awesomecua.github.io/resources/github-nanobrowser-nanobrowser/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.3.0"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "4d8a0b0d6286c4bab116c155f087948062166e7e169d6bbaf504fedb4f32dc67", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6b48386f32a971a724906de8b6d5ea53036f3813411ddc12931be3010c034400", "supports": ["identity"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "afd0b71b01d3d89033d8f32abaf9871f38e8fb0c50442ed3a8cbb3079d4a4fe0", "supports": ["release_status"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c7b28db1b1c8e78d100da066b08db713bc144baff99941b6c94489ef61baee8f", "supports": ["release_status"], "url": "https://api.github.com/repos/nanobrowser/nanobrowser/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:01:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:01:00Z", "url": "https://raw.githubusercontent.com/nanobrowser/nanobrowser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nanobrowser/nanobrowser"}], "name": "Nanobrowser", "publication": null, "published_at": null, "record_sha256": "6955ae4f4c535ce190d563e9f878c90d2f325f00e42ca12893041deb019a2acd", "relations": [], "resource_id": "github:nanobrowser/nanobrowser", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/nanobrowser/nanobrowser", "topics": ["web-navigation", "automation", "human-in-the-loop"], "updated_at": null}, "text": "Nanobrowser\nResource type: agent\nReview status: curated\nScope: core\n\nA browser-extension agent that reads pages, interacts with elements and moves between tabs from a side-panel task interface.\nPlatforms: web\nTopics: web-navigation, automation, human-in-the-loop\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"v0.3.0\"}\n\nThe official repository publishes extension source and a v0.3.0 ZIP release. It names Chrome and Edge as supported browsers. Apache-2.0 is declared for the code. Ask/Auto behavior and provider choices are documented, but no security guarantee, model performance or service-price claim is inferred.\n\nPrimary source: https://github.com/nanobrowser/nanobrowser\nCatalog record: https://awesomecua.github.io/resources/github-nanobrowser-nanobrowser/"}
{"id": "github:neulab/multiui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/neulab/multiui", "citation_url": "https://awesomecua.github.io/resources/github-neulab-multiui/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md", "population": "reported MultiUI training collection", "qualifier": "approximately", "unit": "samples", "value": 7300000}], "evidence_urls": ["https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"], "formats": [], "modalities": ["webpage UI content", "text", "accessibility-derived structure"], "notes": "The README provides a rounded sample count. It does not establish a unique-screenshot inventory or a normalized split breakdown. Synthetic examples and publicly sourced web content have provenance and quality limits; data terms must be checked separately from repository code.", "splits": ["training collection"], "supervision": ["UI understanding", "text-rich visual tasks"]}, "modalities": ["screenshot", "instruction"], "reported_scale": {"comparison": "approximately", "unit": "samples", "value": 7300000}, "sample_count": null, "sample_unit": "instructions"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/neulab/multiui"}], "name": "MultiUI", "publication": null, "published_at": null, "record_sha256": "be287bc057d52310e827050823e356d93420f8245805dce96fa932e91d607d98", "relations": [], "resource_id": "github:neulab/multiui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/neulab/multiui", "topics": ["screen-understanding", "data-generation"], "updated_at": null}, "text": "MultiUI\nResource type: dataset\nReview status: curated\nScope: core\n\n7.3 million training samples built from webpage UIs and structured with enhanced accessibility trees, for text-rich visual understanding, released with UIX-Qwen2 models.\nPlatforms: web\nTopics: screen-understanding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md\", \"population\": \"reported MultiUI training collection\", \"qualifier\": \"approximately\", \"unit\": \"samples\", \"value\": 7300000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md\"], \"formats\": [], \"modalities\": [\"webpage UI content\", \"text\", \"accessibility-derived structure\"], \"notes\": \"The README provides a rounded sample count. It does not establish a unique-screenshot inventory or a normalized split breakdown. Synthetic examples and publicly sourced web content have provenance and quality limits; data terms must be checked separately from repository code.\", \"splits\": [\"training collection\"], \"supervision\": [\"UI understanding\", \"text-rich visual tasks\"]}, \"modalities\": [\"screenshot\", \"instruction\"], \"reported_scale\": {\"comparison\": \"approximately\", \"unit\": \"samples\", \"value\": 7300000}, \"sample_unit\": \"instructions\"}\n\nPrimary source: https://github.com/neulab/multiui\nCatalog record: https://awesomecua.github.io/resources/github-neulab-multiui/"}
{"id": "github:night-chen/workforceagent-r1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/night-chen/WorkForceAgent-R1", "citation_url": "https://awesomecua.github.io/resources/github-night-chen-workforceagent-r1/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "2fb7c0d70ccfa93605d5250a8f6d5182324f0c53d70e1c35ba18cb00f4ada548", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/night-chen/WorkForceAgent-R1/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "0ad3ef1d900200e3ecab86cd44f79038000e8748fb285cdd5f32882455d2e36e", "supports": ["identity"], "url": "https://api.github.com/repos/night-chen/WorkForceAgent-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:39Z", "role": "primary", "sha256": "521d065188a6354416c1aafe88a0eb34486f2c2ec011c1cdf8ed008760dd9c55", "supports": ["release_status"], "url": "https://api.github.com/repos/night-chen/WorkForceAgent-R1/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/night-chen/WorkForceAgent-R1"}], "name": "WorkForceAgent-R1", "publication": null, "published_at": null, "record_sha256": "3f33b1a95db32525f063ea1988fffd5cde8c780738df4326054db8ef3d0443df", "relations": [{"evidence_url": "https://raw.githubusercontent.com/night-chen/WorkForceAgent-R1/HEAD/README.md", "target_id": "arxiv:2505.22942", "type": "documented_by"}], "resource_id": "github:night-chen/workforceagent-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/night-chen/WorkForceAgent-R1", "topics": ["reinforcement-learning", "web-navigation", "planning"], "updated_at": null}, "text": "WorkForceAgent-R1\nResource type: framework\nReview status: curated\nScope: core\n\nWeb-task reward functions, WorkArena data processing and GRPO training configurations for reasoning-oriented language agents.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, planning\n\nThe release contains WorkArena preprocessing and training source. The inspected root has no explicit project license; dependencies retain their own terms. Although the README discusses trained 7B/14B models, it does not provide a verified checkpoint release, so this is cataloged only as a framework.\n\nPrimary source: https://github.com/night-chen/WorkForceAgent-R1\nCatalog record: https://awesomecua.github.io/resources/github-night-chen-workforceagent-r1/"}
{"id": "github:niuzaisheng/screenagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/niuzaisheng/ScreenAgent", "citation_url": "https://awesomecua.github.io/resources/github-niuzaisheng-screenagent/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "2248b2048b9166ce0348b24e81b2587fdf373a958fd86ab870ba85825425b02f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/niuzaisheng/ScreenAgent/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/niuzaisheng/ScreenAgent"}], "name": "ScreenAgent", "publication": null, "published_at": null, "record_sha256": "4451bc25da97f51b1dd00ce32c402ac51fca7d9c7548e175b3f8a7d0c4e00ba3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/niuzaisheng/ScreenAgent/HEAD/README.md", "target_id": "arxiv:2402.07945", "type": "documented_by"}], "resource_id": "github:niuzaisheng/screenagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/niuzaisheng/ScreenAgent", "topics": ["planning"], "updated_at": null}, "text": "ScreenAgent\nResource type: agent\nReview status: curated\nScope: core\n\nEnvironment and agent in which a vision-language model controls real desktops over VNC with mouse and keyboard in plan, act and reflect stages, plus a hand-annotated dataset of daily computer tasks.\nPlatforms: desktop\nTopics: planning\n\nPrimary source: https://github.com/niuzaisheng/ScreenAgent\nCatalog record: https://awesomecua.github.io/resources/github-niuzaisheng-screenagent/"}
{"id": "github:njucckevin/seeclick", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/njucckevin/SeeClick", "citation_url": "https://awesomecua.github.io/resources/github-njucckevin-seeclick/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/njucckevin/SeeClick"}], "name": "SeeClick", "publication": null, "published_at": null, "record_sha256": "6817750fe935591caf7c1d0c256adcee56898a80eaacfb8cfde304b7ff5480d5", "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "resource_id": "github:njucckevin/seeclick", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/njucckevin/SeeClick", "topics": ["grounding"], "updated_at": null}, "text": "SeeClick\nResource type: model\nReview status: curated\nScope: core\n\nGUI grounding model trained to locate visual interface elements.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/njucckevin/SeeClick\nCatalog record: https://awesomecua.github.io/resources/github-njucckevin-seeclick/"}
{"id": "github:nottelabs/notte", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/nottelabs/notte", "citation_url": "https://awesomecua.github.io/resources/github-nottelabs-notte/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "5ecb73264cce10b75ec04dea59aaff77ffde552259fba6932507779a35681c5b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/nottelabs/notte/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nottelabs/notte"}], "name": "Notte", "publication": null, "published_at": null, "record_sha256": "201fc693a5ed89afd3814540b5d3cfa209e6146ecb91f889520e8f6ed68b3db0", "relations": [], "resource_id": "github:nottelabs/notte", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/nottelabs/notte", "topics": ["web-navigation"], "updated_at": null}, "text": "Notte\nResource type: framework\nReview status: curated\nScope: core\n\nWeb agent framework and cloud SDKs for managed Chrome sessions, adding a CLI, an MCP server for coding agents, serverless browser functions and an API that builds automations from English tasks.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/nottelabs/notte\nCatalog record: https://awesomecua.github.io/resources/github-nottelabs-notte/"}
{"id": "github:nut-tree/nut.js", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/nut-tree/nut.js", "citation_url": "https://awesomecua.github.io/resources/github-nut-tree-nut-js/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "41a6435d6c27a647b4d2170cba476fa6a941b12f8a4ffe41674a508df9e0ea1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/nut-tree/nut.js/HEAD/README.md"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/nut-tree/nut.js"}], "name": "Nut.js", "publication": null, "published_at": null, "record_sha256": "b9f5a0ee831edfe123818ab320f5d4b6fc8adcf1f75f78bbf323afa6e34fca0d", "relations": [], "resource_id": "github:nut-tree/nut.js", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/nut-tree/nut.js", "topics": ["automation"], "updated_at": null}, "text": "Nut.js\nResource type: infrastructure\nReview status: curated\nScope: adjacent\n\nDesktop automation library for keyboard, mouse, and screen interaction.\nPlatforms: desktop\nTopics: automation\n\nPrimary source: https://github.com/nut-tree/nut.js\nCatalog record: https://awesomecua.github.io/resources/github-nut-tree-nut-js/"}
{"id": "github:oculix-org/oculix", "metadata": {"aliases": ["Oculix", "SikuliX continuation"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/oculix-org/Oculix", "citation_url": "https://awesomecua.github.io/resources/github-oculix-org-oculix/", "details": {"availability": "Public source and official stable release assets.", "license_components_note": "Project source declares MIT; the retained third-party dependency notices describe additional component terms.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"], "reason": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution."}}, "release_tag": "v4.0.0", "repository_revision": "02ea8844483a83a2963db8016cd7ad421e15bc91"}, "environments": ["desktop", "windows", "macos", "linux", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "41d32f40bfb960f38b77ebb62a73ae20454be6c52829c9cbf0b8fedd27de2297", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "151ed1aaf544b9afcfd2d8f13d0fa682d8cf8f8a68496af20c3597b082070801", "supports": ["identity"], "url": "https://api.github.com/repos/oculix-org/Oculix"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:48Z", "role": "primary", "sha256": "cba4541e5d89210b570e42cc69c1529ca36e698622c69b4bb0eeffe56339a759", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/oculix-org/Oculix/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "1022eb9591a2f01833e88c2bae2869faeac9b92e535901d2c79e93a714278777", "supports": ["release_status", "details.release_tag", "environments", "notes"], "url": "https://api.github.com/repos/oculix-org/Oculix/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:14Z", "role": "primary", "sha256": "8994ff2000b0871aca819ecd7233430bb82cea3f0134a555cebbfa9e02e0f4da", "supports": ["identity", "relations", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/SikuliX1/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "102b86e9019fe5e936486b3213d2bbd9e80d6c7ad393284b1a4298b0be555ce7", "supports": ["identity"], "url": "https://api.github.com/repos/RaiMan/SikuliX1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "cea4b6292d59625e14968ec666faa7ae6e237b510d846a96875bd6a282677f3a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:50Z", "role": "primary", "sha256": "f0e0d8aa91c1c60bac2250688248c85503036aeac37ec02b8a75fe9d50b6ae4c", "supports": ["licenses.code", "notes"], "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/API/src/main/resources/Settings/THIRD-PARTY.txt"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:18:14Z", "licenses": {"code": {"basis": "Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/oculix-org/Oculix"}], "name": "OculiX (SikuliX continuation)", "publication": null, "published_at": null, "record_sha256": "85704e8a535bc3ea92d390ea08408204799d1dee32ab6dc2079d744db25070e5", "relations": [{"evidence_url": "https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/README.md", "target_id": "github:sikuli/sikuli", "type": "variant_of"}], "resource_id": "github:oculix-org/oculix", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/oculix-org/Oculix", "topics": ["automation", "grounding", "tool-integration"], "updated_at": null}, "text": "OculiX (SikuliX continuation)\nResource type: tool\nReview status: curated\nScope: core\n\nA maintained continuation of SikuliX that locates on-screen targets with image matching and OCR, drives desktop or Android interfaces, and exposes scripting and optional MCP interfaces.\nRecorded aliases: Oculix, SikuliX continuation\nPlatforms: desktop, windows, macos, linux, android\nTopics: automation, grounding, tool-integration\n\nRecorded specifications: {\"availability\": \"Public source and official stable release assets.\", \"license_components_note\": \"Project source declares MIT; the retained third-party dependency notices describe additional component terms.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/oculix-org/Oculix/HEAD/LICENSE\"], \"reason\": \"Publisher project-source LICENSE. Third-party dependency notices are separately retained and no blanket MIT assertion is made for the compiled distribution.\"}}, \"release_tag\": \"v4.0.0\", \"repository_revision\": \"02ea8844483a83a2963db8016cd7ad421e15bc91\"}\n\nThe official stable-release API lists v4.0.0 with platform-specific API/IDE JARs and an MCP server JAR; no binary was downloaded. The default README still advertises 3.0.3, so release-specific compatibility follows the release notes rather than that older badge. The RaiMan/SikuliX1 API resolves to oculix-org/SikuliX1, whose README calls itself a historical mirror and directs current users here; it is not separately counted. MIT describes the project source. Retained dependency notices also include GPL/LGPL, Apache, BSD and MPL declarations; a bundled artifact is not asserted to be uniformly MIT.\n\nPrimary source: https://github.com/oculix-org/Oculix\nCatalog record: https://awesomecua.github.io/resources/github-oculix-org-oculix/"}
{"id": "github:omxyz/lumen", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/omxyz/lumen", "citation_url": "https://awesomecua.github.io/resources/github-omxyz-lumen/", "details": {"availability": "Public browser-agent source and npm release metadata.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"], "reason": "Publisher project-source license; external model and browser-service terms are separate."}}, "package_version": "0.2.0", "repository_revision": "b1ad26a0784645ac3a97d402db99cd5d17f86334"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "7708f434142498959bb5ed5de35ac1b945f9ad6e070e436b4e73819b7db1df7f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "32b02ea43d19929d82e8b8c52bb8e448553aebdd96a8754d144927cb0f34c5b6", "supports": ["identity"], "url": "https://api.github.com/repos/omxyz/lumen"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "9911953e501a245d657f3835b22c6822ad7d990b96f03e3cba93a15ebe4f70a6", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/omxyz/lumen/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "34a03b6dd8cd1960d5592a6569e2a1c1f6c1b2697da1386a58d0ca64157f059e", "supports": ["identity", "release_status", "details.package_version", "notes"], "url": "https://registry.npmjs.org/@omxyz%2Flumen/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "30dd5336a8a45b548b9051e751d45923a9864b3810a36c836df9b07447ca3c5c", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": {"basis": "Publisher project-source license; external model and browser-service terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/omxyz/lumen"}], "name": "Lumen browser agent", "publication": null, "published_at": null, "record_sha256": "1bbfe29e4a2cce9b2f7c967a2be35e4681c62b1ad89607062d63e8b6044556e7", "relations": [], "resource_id": "github:omxyz/lumen", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/omxyz/lumen", "topics": ["web-navigation", "memory", "automation", "observability"], "updated_at": null}, "text": "Lumen browser agent\nResource type: agent\nReview status: curated\nScope: core\n\nA screenshot-driven browser agent over Chrome DevTools Protocol, with resumable sessions, context compression, action replay and configurable completion checks.\nPlatforms: web\nTopics: web-navigation, memory, automation, observability\n\nRecorded specifications: {\"availability\": \"Public browser-agent source and npm release metadata.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/omxyz/lumen/HEAD/LICENSE\"], \"reason\": \"Publisher project-source license; external model and browser-service terms are separate.\"}}, \"package_version\": \"0.2.0\", \"repository_revision\": \"b1ad26a0784645ac3a97d402db99cd5d17f86334\"}\n\nPublic TypeScript source and npm package metadata are available; the inspected package version is 0.2.0. The package declares Node ^20.19.0 or >=22.12.0 and requires a browser and model provider. Its README benchmark is explicitly preliminary and uses a small WebVoyager subset; this catalog makes no comparative performance or independent safety claim. Referenced research influences are not asserted as papers introducing Lumen.\n\nPrimary source: https://github.com/omxyz/lumen\nCatalog record: https://awesomecua.github.io/resources/github-omxyz-lumen/"}
{"id": "github:open-compass/mmbench-gui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/open-compass/MMBench-GUI", "citation_url": "https://awesomecua.github.io/resources/github-open-compass-mmbench-gui/", "details": {"benchmark_protocol": {"actions": [], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "population": "reported multi-level benchmark", "qualifier": "more-than", "unit": "tasks", "value": 8000}], "evaluation": "Level-specific evaluation; Efficiency–Quality Area (EQA) combines navigation quality and efficiency.", "evidence_urls": ["https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"], "focus": ["screen understanding", "grounding", "task automation", "task collaboration"], "notes": "The README documents four levels, but its release notes and TODOs distinguish released L1/L2 evaluation from pending L3/L4 integration. Do not treat every level as the same evaluator or a complete implemented interface.", "observations": ["screenshots"], "setting": "mixed"}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "35db9fca3ed1c1a3513924fefff4ccd224bc70cd7eb4bf8c54fcb9048851988e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/open-compass/MMBench-GUI"}], "name": "MMBench-GUI", "publication": null, "published_at": null, "record_sha256": "27da72eb791334ce95482b3723a2dc52d73b539f1ec9c39a77ca2afdfc18c18b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md", "target_id": "arxiv:2507.19478", "type": "documented_by"}], "resource_id": "github:open-compass/mmbench-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/open-compass/MMBench-GUI", "topics": ["evaluation", "grounding"], "updated_at": null}, "text": "MMBench-GUI\nResource type: benchmark\nReview status: curated\nScope: core\n\nFour-level GUI-agent benchmark across six platforms, from content understanding and element grounding to task automation and collaboration, scored with an efficiency-quality area metric.\nPlatforms: cross-platform\nTopics: evaluation, grounding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md\", \"population\": \"reported multi-level benchmark\", \"qualifier\": \"more-than\", \"unit\": \"tasks\", \"value\": 8000}], \"evaluation\": \"Level-specific evaluation; Efficiency–Quality Area (EQA) combines navigation quality and efficiency.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/open-compass/MMBench-GUI/HEAD/README.md\"], \"focus\": [\"screen understanding\", \"grounding\", \"task automation\", \"task collaboration\"], \"notes\": \"The README documents four levels, but its release notes and TODOs distinguish released L1/L2 evaluation from pending L3/L4 integration. Do not treat every level as the same evaluator or a complete implemented interface.\", \"observations\": [\"screenshots\"], \"setting\": \"mixed\"}}\n\nPrimary source: https://github.com/open-compass/MMBench-GUI\nCatalog record: https://awesomecua.github.io/resources/github-open-compass-mmbench-gui/"}
{"id": "github:openadaptai/openadapt", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenAdaptAI/OpenAdapt", "citation_url": "https://awesomecua.github.io/resources/github-openadaptai-openadapt/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "36a6e7d56a1d474ee83e5811babe8e11cb113ee12386f9621bba64ffd957a6bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "7f2bc0ccbf15854bbe46f5fa0e7cb26fe42f88ec9f8c9b151f0f93c4415820f9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/OpenAdaptAI/OpenAdapt/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenAdaptAI/OpenAdapt"}], "name": "OpenAdapt", "publication": null, "published_at": null, "record_sha256": "a1759be8b40b33aea8dc137847d70a157a0aa19842adb8f6abbec38a4dc16eaa", "relations": [], "resource_id": "github:openadaptai/openadapt", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenAdaptAI/OpenAdapt", "topics": ["recording", "workflow"], "updated_at": null}, "text": "OpenAdapt\nResource type: agent\nReview status: curated\nScope: core\n\nTools for learning and automating workflows from desktop demonstrations.\nPlatforms: desktop\nTopics: recording, workflow\n\nPrimary source: https://github.com/OpenAdaptAI/OpenAdapt\nCatalog record: https://awesomecua.github.io/resources/github-openadaptai-openadapt/"}
{"id": "github:openai/universe", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/openai/universe", "citation_url": "https://awesomecua.github.io/resources/github-openai-universe/", "details": {"availability": "Public archived client source; remote benchmark services and images are not verified usable.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"], "reason": "Official Universe client-source LICENSE; no license or availability claim is transferred to external remote applications or container contents."}}, "maintenance_status": "archived and deprecated", "repository_revision": "cc9ce6ec241821bfb0f3b85dd455bd36e4ee7a8c"}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "b1ca938480d4d36c76fa939c5b10cef11c24465cc1a6a8ee857cec4b939f9d61", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/openai/universe/HEAD/README.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:15Z", "role": "primary", "sha256": "94a7a4d383f754b96601bc0486b90e62760b8bfc35abc607077c8eb6e7934b2d", "supports": ["identity", "details.availability", "notes"], "url": "https://api.github.com/repos/openai/universe"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:47Z", "role": "primary", "sha256": "1808d7524506d1386dad507de8a8fa0f89080443a970ae032cea4c204699d113", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/openai/universe/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:05Z", "role": "primary", "sha256": "ecd4a4aacc4512a31bbdc3252ffc014d37ac7ef56ed34bc2a5d0f6337df6daa5", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T11:20:17Z", "licenses": {"code": {"basis": "Official Universe client-source LICENSE; no license or availability claim is transferred to external remote applications or container contents.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:05Z", "url": "https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/openai/universe"}], "name": "Universe legacy GUI interaction client", "publication": null, "published_at": null, "record_sha256": "703a9352dd97b477966e62b4a9b2095749f23cd0501d5f881f97b92087d05eef", "relations": [], "resource_id": "github:openai/universe", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/openai/universe", "topics": ["automation", "reinforcement-learning", "history"], "updated_at": null}, "text": "Universe legacy GUI interaction client\nResource type: infrastructure\nReview status: curated\nScope: adjacent\n\nAn archived Gym-compatible client that exposes screen observations and keyboard/mouse actions over VNC for remote application and browser environments, including the historical World of Bits integration.\nPlatforms: web, desktop\nTopics: automation, reinforcement-learning, history\n\nRecorded specifications: {\"availability\": \"Public archived client source; remote benchmark services and images are not verified usable.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/openai/universe/HEAD/LICENSE\"], \"reason\": \"Official Universe client-source LICENSE; no license or availability claim is transferred to external remote applications or container contents.\"}}, \"maintenance_status\": \"archived and deprecated\", \"repository_revision\": \"cc9ce6ec241821bfb0f3b85dd455bd36e4ee7a8c\"}\n\nGitHub marks this repository archived, and the README declares the library deprecated. It contains the client side of Universe; the README says remote-environment source repositories would be released separately. Original MiniWoB task-server or container availability was not established: alpha.openai.com no longer resolved in this audit and the official Quay tag metadata endpoint returned 401. This entry records the released client infrastructure, not a currently runnable original MiniWoB benchmark. Legacy host requirements are documented; no container or code was executed.\n\nPrimary source: https://github.com/openai/universe\nCatalog record: https://awesomecua.github.io/resources/github-openai-universe/"}
{"id": "github:openbmb/agentcpm-gui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenBMB/AgentCPM-GUI", "citation_url": "https://awesomecua.github.io/resources/github-openbmb-agentcpm-gui/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2cf79e21b5438bf6a2e6bd450ba8c4d86404560dfb61836f7e300112ce74a59d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenBMB/AgentCPM-GUI"}], "name": "AgentCPM-GUI", "publication": null, "published_at": null, "record_sha256": "2221c4f6bf1a75d012fdb05ff6b0ac588dcf0dffd2dc4a3b128e04e9d11fd247", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenBMB/AgentCPM-GUI/HEAD/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "resource_id": "github:openbmb/agentcpm-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenBMB/AgentCPM-GUI", "topics": ["grounding", "mobile-navigation"], "updated_at": null}, "text": "AgentCPM-GUI\nResource type: model\nReview status: curated\nScope: core\n\n8B on-device smartphone agent built on MiniCPM-V that operates Chinese and English apps from screenshots, trained with reinforcement fine-tuning and a compact JSON action space.\nPlatforms: android\nTopics: grounding, mobile-navigation\n\nPrimary source: https://github.com/OpenBMB/AgentCPM-GUI\nCatalog record: https://awesomecua.github.io/resources/github-openbmb-agentcpm-gui/"}
{"id": "github:opendfm/moba", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenDFM/MobA", "citation_url": "https://awesomecua.github.io/resources/github-opendfm-moba/", "details": {"availability": "Public Android executor, agents, prompts and setup documentation.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/OpenDFM/MobA/git/trees/main?recursive=1"], "reason": "No repository source-license declaration established from the current public tree."}}, "repository_revision": "05cd237d80fcd23ad0dd6096cdb144117ecb3471"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "49c096e5ee6a87205497d693b5248817ea98cad0f7e5254793660007ef4292f7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/OpenDFM/MobA/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "0ecd1833a1b6840441769e1cc26a55268f2de7beb61bc3e783a851dd7cff45f1", "supports": ["identity"], "url": "https://api.github.com/repos/OpenDFM/MobA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:04Z", "role": "primary", "sha256": "302f324fce2884c5948cf62c75f2f839cd40c301faff6ccf4b6eb5dd29f953a3", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/OpenDFM/MobA/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T12:30:53Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenDFM/MobA"}], "name": "MobA", "publication": null, "published_at": null, "record_sha256": "8bc4ca9fe3f7d084b83062cfc97071e4cbc96202f8c16d879b9587d02d6b02fa", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenDFM/MobA/HEAD/README.md", "target_id": "arxiv:2410.13757", "type": "documented_by"}], "resource_id": "github:opendfm/moba", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenDFM/MobA", "topics": ["mobile-navigation", "memory", "planning", "self-reflection"], "updated_at": null}, "text": "MobA\nResource type: agent\nReview status: curated\nScope: core\n\nAn Android agent that combines global task planning and memory with a local action executor, using reflection to revise mobile interaction steps.\nPlatforms: android\nTopics: mobile-navigation, memory, planning, self-reflection\n\nRecorded specifications: {\"availability\": \"Public Android executor, agents, prompts and setup documentation.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://api.github.com/repos/OpenDFM/MobA/git/trees/main?recursive=1\"], \"reason\": \"No repository source-license declaration established from the current public tree.\"}}, \"repository_revision\": \"05cd237d80fcd23ad0dd6096cdb144117ecb3471\"}\n\nThe repository releases a configurable Python executor and Android ADB integration; its setup documents a Windows host, but the controlled interface is Android. The source is marked under continuing development and requires external model credentials. The inspected repository has no license file; the separately hosted MobBench data declaration is not transferred to this code.\n\nPrimary source: https://github.com/OpenDFM/MobA\nCatalog record: https://awesomecua.github.io/resources/github-opendfm-moba/"}
{"id": "github:opengvlab/gui-odyssey", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenGVLab/GUI-Odyssey", "citation_url": "https://awesomecua.github.io/resources/github-opengvlab-gui-odyssey/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "population": "recommended hflqf88888/GUIOdyssey release", "qualifier": "exact", "unit": "episodes", "value": 8334}], "evidence_urls": ["https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"], "formats": ["JSON annotations", "PNG screenshots"], "modalities": ["screenshots", "instructions", "actions", "bounding boxes", "step context"], "notes": "The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.", "splits": ["random", "task", "device", "app"], "supervision": ["cross-app trajectories", "action grounding"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4585b765d7c8b3e9bb6187d36d576edde454d4c0158a6aa8b3fdbedf34c89a15", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "78c19ee712487af60df1913f11572958564e59a1a2cc663b1b30d971855d7a28", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/GUI-Odyssey"}], "name": "GUI-Odyssey", "publication": null, "published_at": null, "record_sha256": "a901fd7558fcf5f4842603b90925724801db35deaa9a51d436a33a164f3c49c7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "target_id": "arxiv:2406.08451", "type": "documented_by"}], "resource_id": "github:opengvlab/gui-odyssey", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenGVLab/GUI-Odyssey", "topics": ["cross-app", "data-generation"], "updated_at": null}, "text": "GUI-Odyssey\nResource type: dataset\nReview status: curated\nScope: core\n\nHuman-demonstrated Android episodes for cross-app navigation, spanning six task types, 212 apps and about 1,400 app combinations, for training and evaluating agents.\nPlatforms: android\nTopics: cross-app, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md\", \"population\": \"recommended hflqf88888/GUIOdyssey release\", \"qualifier\": \"exact\", \"unit\": \"episodes\", \"value\": 8334}], \"evidence_urls\": [\"https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md\", \"https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md\", \"https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md\"], \"formats\": [\"JSON annotations\", \"PNG screenshots\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\", \"bounding boxes\", \"step context\"], \"notes\": \"The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.\", \"splits\": [\"random\", \"task\", \"device\", \"app\"], \"supervision\": [\"cross-app trajectories\", \"action grounding\"]}}\n\nPrimary source: https://github.com/OpenGVLab/GUI-Odyssey\nCatalog record: https://awesomecua.github.io/resources/github-opengvlab-gui-odyssey/"}
{"id": "github:opengvlab/scalecua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenGVLab/ScaleCUA", "citation_url": "https://awesomecua.github.io/resources/github-opengvlab-scalecua/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "0538701c860afedd5d63c3839039699d8e5a14977d5c339b187ed13ad666c8fa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/ScaleCUA"}], "name": "ScaleCUA", "publication": null, "published_at": null, "record_sha256": "d431e23830242e264ac97e5de219c5580cdead9eb3426be81198155c5d798ca3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "resource_id": "github:opengvlab/scalecua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenGVLab/ScaleCUA", "topics": ["data-generation", "planning"], "updated_at": null}, "text": "ScaleCUA\nResource type: model\nReview status: curated\nScope: core\n\nComputer-use model family trained with cross-platform interaction data.\nPlatforms: cross-platform\nTopics: data-generation, planning\n\nPrimary source: https://github.com/OpenGVLab/ScaleCUA\nCatalog record: https://awesomecua.github.io/resources/github-opengvlab-scalecua/"}
{"id": "github:opengvlab/zerogui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenGVLab/ZeroGUI", "citation_url": "https://awesomecua.github.io/resources/github-opengvlab-zerogui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "033d2de815e11c208862c5c5601b78d58668b79cadd37a750cad9edd4dacabed", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenGVLab/ZeroGUI"}], "name": "ZeroGUI", "publication": null, "published_at": null, "record_sha256": "a544237550ad7b8177769d59c5daebb00c6cdc696263d993cda03573b79264ea", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ZeroGUI/HEAD/README.md", "target_id": "arxiv:2505.23762", "type": "documented_by"}], "resource_id": "github:opengvlab/zerogui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenGVLab/ZeroGUI", "topics": ["reinforcement-learning", "data-generation"], "updated_at": null}, "text": "ZeroGUI\nResource type: framework\nReview status: curated\nScope: core\n\nOnline RL framework for GUI agents in which VLMs generate the training tasks and judge success from trajectory screenshots, with OSWorld and AndroidLab checkpoints trained from UI-TARS-7B-DPO.\nPlatforms: cross-platform\nTopics: reinforcement-learning, data-generation\n\nPrimary source: https://github.com/OpenGVLab/ZeroGUI\nCatalog record: https://awesomecua.github.io/resources/github-opengvlab-zerogui/"}
{"id": "github:openinterpreter/open-interpreter", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenInterpreter/open-interpreter", "citation_url": "https://awesomecua.github.io/resources/github-openinterpreter-open-interpreter/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8f6c33487eca77769798677d32bce15335d0439f129c9a0413e12e0d9e70ab7a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "d17f227e4df5da1600391338865ce0f3055211760a36688f816941d58232d8dc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OpenInterpreter/open-interpreter/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenInterpreter/open-interpreter"}], "name": "Open Interpreter", "publication": null, "published_at": null, "record_sha256": "53334f5a480c7412b3e2b2969e3921017fd02b9fdcb4ea32b14202ca205746fa", "relations": [], "resource_id": "github:openinterpreter/open-interpreter", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenInterpreter/open-interpreter", "topics": ["tool-integration"], "updated_at": null}, "text": "Open Interpreter\nResource type: agent\nReview status: curated\nScope: core\n\nAssistant framework combining code execution and computer interaction.\nPlatforms: desktop\nTopics: tool-integration\n\nThe repository has evolved substantially; computer-use functionality is documented in its Computer Use section. Do not infer parity with historical releases.\n\nPrimary source: https://github.com/OpenInterpreter/open-interpreter\nCatalog record: https://awesomecua.github.io/resources/github-openinterpreter-open-interpreter/"}
{"id": "github:openmobile2/mobilegym-plusplus", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenMobile2/MobileGym-plusplus", "citation_url": "https://awesomecua.github.io/resources/github-openmobile2-mobilegym-plusplus/", "details": {"benchmark_protocol": {"actions": ["GUI gestures", "foreground-app tool calls in hybrid mode"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json", "population": "official.json", "qualifier": "exact", "unit": "task definitions", "value": 215}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "population": "complete simulator installation", "qualifier": "exact", "unit": "simulated clients", "value": 63}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "population": "new clients in this release", "qualifier": "exact", "unit": "simulated clients", "value": 35}], "evaluation": "Programmatic checkers inspect changes in application state; both modes share task definitions and initial states.", "evidence_urls": ["https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md", "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json"], "focus": ["task completion", "cross-application workflows", "GUI and app-native tool comparison"], "notes": "The task count was checked against unique task_id values in the published manifest. The 35 new clients are a subset of 63 total clients. This simulator runs in a browser; Android-emulator apps are a separate part of OpenMobile-2.", "observations": ["screenshots"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "d9f850ec196e715650e679cb81515adaa64ddc5a60a882bfe97b0f6a098ddc83", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["relations"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:50:11Z", "role": "primary", "sha256": "35c1641b349201c68917fbe4c4036814487678e9588d9b1c903af26917800316", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "7618cd7de039c264759ca8ca0b9130c6de63a4536dcf436baedd4cfbf639073e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenMobile2/MobileGym-plusplus"}], "name": "MobileGym++ / MobileGym++ Bench", "publication": null, "published_at": null, "record_sha256": "a3978a1f97bcaecad3f54da9cf565a73e1131e1b9099effdede561f48266c342", "relations": [{"evidence_url": "https://os-copilot.github.io/OpenMobile2-Home/", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "resource_id": "github:openmobile2/mobilegym-plusplus", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenMobile2/MobileGym-plusplus", "topics": ["mobile-navigation", "evaluation", "cross-app", "tool-integration"], "updated_at": null}, "text": "MobileGym++ / MobileGym++ Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\nBrowser-rendered mobile simulation with resettable app state and a shared task suite for GUI-only and GUI-plus-tool agents.\nPlatforms: web\nTopics: mobile-navigation, evaluation, cross-app, tool-integration\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"GUI gestures\", \"foreground-app tool calls in hybrid mode\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json\", \"population\": \"official.json\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 215}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md\", \"population\": \"complete simulator installation\", \"qualifier\": \"exact\", \"unit\": \"simulated clients\", \"value\": 63}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md\", \"population\": \"new clients in this release\", \"qualifier\": \"exact\", \"unit\": \"simulated clients\", \"value\": 35}], \"evaluation\": \"Programmatic checkers inspect changes in application state; both modes share task definitions and initial states.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/README.md\", \"https://raw.githubusercontent.com/OpenMobile2/MobileGym-plusplus/HEAD/bench_env/benchmark/official.json\"], \"focus\": [\"task completion\", \"cross-application workflows\", \"GUI and app-native tool comparison\"], \"notes\": \"The task count was checked against unique task_id values in the published manifest. The 35 new clients are a subset of 63 total clients. This simulator runs in a browser; Android-emulator apps are a separate part of OpenMobile-2.\", \"observations\": [\"screenshots\"], \"setting\": \"interactive\"}}\n\nThese clients simulate named commercial applications and are not those commercial products. The code and simulator data have distinct license files. No benchmark run was performed. The README describes a phone that runs in the browser, and the Android-emulator playground is a separate OpenMobile2 release, so the platform facet is web rather than a mobile operating system.\n\nPrimary source: https://github.com/OpenMobile2/MobileGym-plusplus\nCatalog record: https://awesomecua.github.io/resources/github-openmobile2-mobilegym-plusplus/"}
{"id": "github:openmobile2/openmobile2-code", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OpenMobile2/OpenMobile2-Code", "citation_url": "https://awesomecua.github.io/resources/github-openmobile2-openmobile2-code/", "details": {}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "485279038266d9454df75e27a80a0a0d71c24e90700126d2dcfa0da921b1273d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OpenMobile2/OpenMobile2-Code"}], "name": "OpenMobile-2 code", "publication": null, "published_at": null, "record_sha256": "1a735b81311144939125d25fe4c28e8f79fd04710faac0765f3679ec85bf1fd1", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "resource_id": "github:openmobile2/openmobile2-code", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OpenMobile2/OpenMobile2-Code", "topics": ["mobile-navigation", "tool-integration", "evaluation"], "updated_at": null}, "text": "OpenMobile-2 code\nResource type: framework\nReview status: curated\nScope: core\n\nEvaluation adapters and training guidance for OpenMobile-2 across Android emulators and browser-rendered mobile simulation.\nPlatforms: android, web\nTopics: mobile-navigation, tool-integration, evaluation\n\nThe README marks evaluation code, checkpoints, environments, and data as released; data-construction scripts and other resources remain pending. Each benchmark runtime is started separately. Code terms do not establish dataset permissions.\n\nPrimary source: https://github.com/OpenMobile2/OpenMobile2-Code\nCatalog record: https://awesomecua.github.io/resources/github-openmobile2-openmobile2-code/"}
{"id": "github:os-copilot/os-atlas", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/OS-Atlas", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-os-atlas/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "021cf14ec9482d1dab04dfeaadc3e39cdc405a284991cd540f5eeae7dea47a12", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Atlas"}], "name": "OS-Atlas", "publication": null, "published_at": null, "record_sha256": "983e8ca8b40f5959552a35161dec5e47697fa763eeecd49b04e09a50fc087228", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Atlas/HEAD/README.md", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "resource_id": "github:os-copilot/os-atlas", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/OS-Atlas", "topics": ["grounding"], "updated_at": null}, "text": "OS-Atlas\nResource type: model\nReview status: curated\nScope: core\n\nOS-Atlas GUI action models, including grounding bases fine-tuned from InternVL2-4B and Qwen2-VL-7B-Instruct that accept any image size and output 0-1000 relative coordinates.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/OS-Copilot/OS-Atlas\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-os-atlas/"}
{"id": "github:os-copilot/os-copilot", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/OS-Copilot", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-os-copilot/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "b840368441695a54137216c2ba6bc99a8d35df75ad4196a4f0f0a1f5b5a70a9d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "3704700ddeaef20ec20cad39ed105ea518b316bc80c04dbcf2e69c3dc768a34f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Copilot"}], "name": "OS-Copilot", "publication": null, "published_at": null, "record_sha256": "426d13df289e4d4c568fbf53959df191f7c59dec4a72529048ffea3c846be519", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Copilot/HEAD/README.md", "target_id": "arxiv:2402.07456", "type": "documented_by"}], "resource_id": "github:os-copilot/os-copilot", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/OS-Copilot", "topics": ["planning", "tool-integration"], "updated_at": null}, "text": "OS-Copilot\nResource type: agent\nReview status: curated\nScope: core\n\nOpen-source library for generalist agents acting on an operating system's web, terminals, files and third-party apps, with the FRIDAY assistant, which takes custom tools and self-learns Excel tasks.\nPlatforms: desktop\nTopics: planning, tool-integration\n\nPrimary source: https://github.com/OS-Copilot/OS-Copilot\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-os-copilot/"}
{"id": "github:os-copilot/os-genesis", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/OS-Genesis", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-os-genesis/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "formats": ["JSONL trajectories", "JSON transition triples"], "modalities": ["screenshots", "text", "actions"], "notes": "The repository links distinct AndroidControl-oriented, AndroidWorld-oriented, and web training files plus raw pre-state/action/post-state triples. These are training artifacts, not a documented train/test split, and related representations are not disjoint corpus totals. No numerical inventory is asserted from release names or linked archives.", "splits": [], "supervision": ["reverse task synthesis", "GUI action trajectories"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Genesis"}], "name": "OS-Genesis", "publication": null, "published_at": null, "record_sha256": "0b44cda7d0c2b41053a0450d925a19ca62481ebdfdcb801ad65d4df0199dba90", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "resource_id": "github:os-copilot/os-genesis", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/OS-Genesis", "topics": ["data-generation"], "updated_at": null}, "text": "OS-Genesis\nResource type: dataset\nReview status: curated\nScope: core\n\nInteraction-driven pipeline that synthesizes mobile and web GUI agent trajectories without human supervision or predefined tasks, via reverse task synthesis and a trajectory reward model.\nPlatforms: cross-platform\nTopics: data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"formats\": [\"JSONL trajectories\", \"JSON transition triples\"], \"modalities\": [\"screenshots\", \"text\", \"actions\"], \"notes\": \"The repository links distinct AndroidControl-oriented, AndroidWorld-oriented, and web training files plus raw pre-state/action/post-state triples. These are training artifacts, not a documented train/test split, and related representations are not disjoint corpus totals. No numerical inventory is asserted from release names or linked archives.\", \"splits\": [], \"supervision\": [\"reverse task synthesis\", \"GUI action trajectories\"]}}\n\nPrimary source: https://github.com/OS-Copilot/OS-Genesis\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-os-genesis/"}
{"id": "github:os-copilot/os-map", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/OS-Map", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-os-map/", "details": {"availability": "Public task definitions, execution harness and an ungated VM-file manifest.", "benchmark_protocol": {"actions": ["actions executed by the OSWorld-derived DesktopEnv"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json", "population": "evaluation_examples/all.json", "qualifier": "exact", "unit": "category/task references", "value": 416}], "evaluation": "Task configurations provide evaluators for application/system state, using an OSWorld-derived runtime.", "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json", "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/README.md"], "focus": ["breadth of desktop task coverage", "depth of desktop task execution"], "notes": "416 category/task pairs are unique, but only 412 bare ID strings are unique across categories. This is a manifest-reference count, not a claim of 416 semantically distinct tasks. Main runtime documentation specifies Ubuntu; the subdirectory README retains older Windows-only wording.", "observations": ["desktop observations from an Ubuntu virtual machine"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"], "reason": "Repository code LICENSE; no license is inferred for the VM image or bundled third-party applications."}}, "repository_revision": "191efd0e6766acab444a31685b437565e1a94ac2"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "fc9f3c8d9fffdff8b51e312b013be0d9fdfc10e57fdc2ec4a8a8d012c4b42475", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:14Z", "role": "primary", "sha256": "d6c91b4910e92f40cc694b4cb6467c1a20f42c0c14047b95a1e777aca0892677", "supports": ["identity"], "url": "https://api.github.com/repos/OS-Copilot/OS-Map"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "41342157c53f28ad2729137491930225ac11da922e7b16762e84ac958a2e9d28", "supports": ["release_status"], "url": "https://api.github.com/repos/OS-Copilot/OS-Map/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:07Z", "role": "primary", "sha256": "1dd924135cbf0a85ce72d75025259adb7b849970d400af4ed854327a34e7ee41", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:22Z", "role": "primary", "sha256": "4199af4c8bc60451079d8a2e7598ab7386bd82b0d5c9ad7e6927d6cb8e03ad6f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:17:22Z", "role": "primary", "sha256": "a6b60d5721cfdd26416e7d0ae3192360896427a97ccb3793812cc150e9cd7126", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:16Z", "role": "primary", "sha256": "d2da837872b123f2927a069a020aeb14fb4b22148599c9f2e8b6689ee444ec88", "supports": ["release_status", "details"], "url": "https://huggingface.co/api/datasets/os-map/OSMAP"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:14Z", "licenses": {"code": {"basis": "Repository code LICENSE; no license is inferred for the VM image or bundled third-party applications.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:07Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Map"}], "name": "OS-Map", "publication": null, "published_at": null, "record_sha256": "0e1fc44c1bafbe2a241356b9ff6a9933dea4d6046ee1277c84485fd87e0d4afd", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "target_id": "github:xlang-ai/osworld", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md", "target_id": "arxiv:2507.19132", "type": "documented_by"}], "resource_id": "github:os-copilot/os-map", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/OS-Map", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": null}, "text": "OS-Map\nResource type: benchmark\nReview status: curated\nScope: core\n\nAn OSWorld-derived desktop benchmark with an Ubuntu virtual-machine environment and expanded task configurations for measuring breadth and depth of computer interaction.\nPlatforms: desktop, linux\nTopics: evaluation, long-horizon, cross-app\n\nRecorded specifications: {\"availability\": \"Public task definitions, execution harness and an ungated VM-file manifest.\", \"benchmark_protocol\": {\"actions\": [\"actions executed by the OSWorld-derived DesktopEnv\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json\", \"population\": \"evaluation_examples/all.json\", \"qualifier\": \"exact\", \"unit\": \"category/task references\", \"value\": 416}], \"evaluation\": \"Task configurations provide evaluators for application/system state, using an OSWorld-derived runtime.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/README.md\", \"https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/all.json\", \"https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/evaluation_examples/README.md\"], \"focus\": [\"breadth of desktop task coverage\", \"depth of desktop task execution\"], \"notes\": \"416 category/task pairs are unique, but only 412 bare ID strings are unique across categories. This is a manifest-reference count, not a claim of 416 semantically distinct tasks. Main runtime documentation specifies Ubuntu; the subdirectory README retains older Windows-only wording.\", \"observations\": [\"desktop observations from an Ubuntu virtual machine\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/OS-Copilot/OS-Map/HEAD/LICENSE\"], \"reason\": \"Repository code LICENSE; no license is inferred for the VM image or bundled third-party applications.\"}}, \"repository_revision\": \"191efd0e6766acab444a31685b437565e1a94ac2\"}\n\nPublic task manifests, agent runner and evaluator code are present, and the linked ungated Hub manifest lists Ubuntu0.zip and task files. The main setup supports x86 VMware Workstation and says ARM/Docker alternatives are still in development. An inherited evaluation-examples README still mentions Windows 10 and under-construction status; those statements are retained as a documentation inconsistency, not interpreted as verified Windows support. The manifest has 416 category/task references but 412 globally distinct ID strings.\n\nPrimary source: https://github.com/OS-Copilot/OS-Map\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-os-map/"}
{"id": "github:os-copilot/os-sentinel", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/OS-Sentinel", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-os-sentinel/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "42bf4aea8369ac2909bf27bcb1078756c2c26567e2b46284c8f3a9e4c4f636c1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/OS-Sentinel"}], "name": "OS-Sentinel", "publication": null, "published_at": null, "record_sha256": "6381b0916b79a5052f644da69d4c22cce65f33d9a58b0eb22a945c733b2fd1f2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/OS-Sentinel/HEAD/README.md", "target_id": "arxiv:2510.24411", "type": "documented_by"}], "resource_id": "github:os-copilot/os-sentinel", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/OS-Sentinel", "topics": ["safety-security"], "updated_at": null}, "text": "OS-Sentinel\nResource type: framework\nReview status: curated\nScope: core\n\nSafety-checking environment and benchmark for mobile agents on AndroidWorld that judges proposed actions with rule-based and VLM-based checks, evaluated on the MobileRisk trajectory dataset.\nPlatforms: desktop\nTopics: safety-security\n\nPrimary source: https://github.com/OS-Copilot/OS-Sentinel\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-os-sentinel/"}
{"id": "github:os-copilot/scienceboard", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OS-Copilot/ScienceBoard", "citation_url": "https://awesomecua.github.io/resources/github-os-copilot-scienceboard/", "details": {"benchmark_protocol": {"actions": ["PyAutoGUI desktop actions", "code/tool commands and answer/stop actions according to configured agent mode"], "counts": [{"basis": "computed", "evidence_url": "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1", "population": "tasks/VM/**/*.json at tree 21b5eb0f; excludes tasks/Raw and tasks/Dual", "qualifier": "exact", "unit": "VM task definitions", "value": 169}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md", "population": "VM suite: Celestia, ChimeraX, GrassGIS, KAlgebra, Lean, TeXstudio", "qualifier": "exact", "unit": "application families", "value": 6}], "evaluation": "Application-specific verifiers inspect state, files, database values, equations, compilation or answer/stop conditions. The official audit records 242 verifier entries across its VM task suite. This profile covers the released VM task definitions; Docker smoke results and Raw/Dual alternatives are separate populations.", "evidence_urls": ["https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/known-issues.md", "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/sci/base/task.py"], "focus": ["scientific-software desktop workflows", "GUI and code-assisted scientific tasks"], "notes": "Manifest count 169 matches the official verifier audit. Other Raw/Dual definitions can represent alternative execution modes of the same problems and are not added. Official Docker migration documents a 20-task acceptance subset, not a full-suite performance result; legacy VMware remains. External service sign-in/rate limits, amd64 binaries and ChimeraX redistribution terms limit portability. Only code MIT is established; VM, scientific data and bundled apps need their own terms.", "observations": ["desktop screenshots", "accessibility-tree observations", "set-of-marks with accessibility tree"], "setting": "interactive"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c3fe93aa740dbf4a279a246ddd836f2590a2a36f39b820abe166e38cb484577", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "d276e11127abc5826bd1b2aa2caf543fcd45a8a60477774e8c8b1054067997ec", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:10Z", "role": "primary", "sha256": "4f90270617c0c1d58f523c0a7f2e4603fa67a49560d68e2f88cfa080d764313b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/known-issues.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:07Z", "role": "primary", "sha256": "3398670a517a5d1d71f420dd866493d3f8a77853e3933fd4928103f9a8be1e4f", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:35:24Z", "role": "primary", "sha256": "d6a1481b453c074e34739fd29def56efa35a1dc10e856418f557f067e9ff08c5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/sci/base/task.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "cbb1d28b515a39c294963d4bfc24077e669a483fc5ebba9b84a727afd4187426", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OS-Copilot/ScienceBoard"}], "name": "ScienceBoard", "publication": null, "published_at": null, "record_sha256": "302d793a67bad14d108af30cf5c086f1de0676cd937d63b812eac8f3207b81f3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "resource_id": "github:os-copilot/scienceboard", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OS-Copilot/ScienceBoard", "topics": ["evaluation", "science"], "updated_at": null}, "text": "ScienceBoard\nResource type: benchmark\nReview status: curated\nScope: core\n\nScientific-workflow tasks for multimodal agents in desktop software such as ChimeraX, KAlgebra, Celestia, Grass GIS and Lean 4, with apps modified to expose state for evaluation.\nPlatforms: desktop\nTopics: evaluation, science\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"PyAutoGUI desktop actions\", \"code/tool commands and answer/stop actions according to configured agent mode\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1\", \"population\": \"tasks/VM/**/*.json at tree 21b5eb0f; excludes tasks/Raw and tasks/Dual\", \"qualifier\": \"exact\", \"unit\": \"VM task definitions\", \"value\": 169}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md\", \"population\": \"VM suite: Celestia, ChimeraX, GrassGIS, KAlgebra, Lean, TeXstudio\", \"qualifier\": \"exact\", \"unit\": \"application families\", \"value\": 6}], \"evaluation\": \"Application-specific verifiers inspect state, files, database values, equations, compilation or answer/stop conditions. The official audit records 242 verifier entries across its VM task suite. This profile covers the released VM task definitions; Docker smoke results and Raw/Dual alternatives are separate populations.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md\", \"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/verifier-audit.md\", \"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/docs/known-issues.md\", \"https://api.github.com/repos/OS-Copilot/ScienceBoard/git/trees/main?recursive=1\", \"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/sci/base/task.py\"], \"focus\": [\"scientific-software desktop workflows\", \"GUI and code-assisted scientific tasks\"], \"notes\": \"Manifest count 169 matches the official verifier audit. Other Raw/Dual definitions can represent alternative execution modes of the same problems and are not added. Official Docker migration documents a 20-task acceptance subset, not a full-suite performance result; legacy VMware remains. External service sign-in/rate limits, amd64 binaries and ChimeraX redistribution terms limit portability. Only code MIT is established; VM, scientific data and bundled apps need their own terms.\", \"observations\": [\"desktop screenshots\", \"accessibility-tree observations\", \"set-of-marks with accessibility tree\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/OS-Copilot/ScienceBoard\nCatalog record: https://awesomecua.github.io/resources/github-os-copilot-scienceboard/"}
{"id": "github:osu-nlp-group/eia_against_webagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/EIA_against_webagent", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-eia-against-webagent/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "61eb6109404f266fd22c6fa8ca58ef773a9d502567403f42edc096958ab2e430", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "953ff47dbbc0c0563d1ec2d245b06d1206f2f38c4ed2f0da728e027ae2e5a973", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/EIA_against_webagent"}], "name": "EIA", "publication": null, "published_at": null, "record_sha256": "b0e1a3f940229d4974fc019cc777fbb0391cfcbaad25eab205110d5ba93fc09f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/EIA_against_webagent/HEAD/README.md", "target_id": "arxiv:2409.11295", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/eia_against_webagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/EIA_against_webagent", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null}, "text": "EIA\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation suite for Environmental Injection Attacks that embed invisible malicious HTML elements into benign websites to test privacy leakage in generalist web agents.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\n\nPrimary source: https://github.com/OSU-NLP-Group/EIA_against_webagent\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-eia-against-webagent/"}
{"id": "github:osu-nlp-group/mind2web", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/Mind2Web", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-mind2web/", "details": {"benchmark_protocol": {"actions": ["CLICK", "TYPE", "SELECT"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "train split", "qualifier": "exact", "unit": "task instances", "value": 1009}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-task test split", "qualifier": "exact", "unit": "task instances", "value": 252}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-website test split", "qualifier": "exact", "unit": "task instances", "value": 177}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "population": "cross-domain test split", "qualifier": "exact", "unit": "task instances", "value": 912}], "evaluation": "Offline candidate and action evaluation against recorded demonstrations; the repository distinguishes macro from micro averaging.", "evidence_urls": ["https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"], "focus": ["web action prediction", "generalization to unseen sites and domains"], "notes": "These are reported split inventories, not online task-success measurements. Original Mind2Web, Multimodal-Mind2Web, and Online-Mind2Web have different inputs/protocols. The authors require macro-average metrics for comparison with the paper; test archives remain upstream.", "observations": ["task instructions", "HTML", "recorded action history"], "setting": "offline"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "157dbd840264ea29f958c0b7383d63783845c573b3e19a1f39b773be22e0b32d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "157dbd840264ea29f958c0b7383d63783845c573b3e19a1f39b773be22e0b32d", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "953ff47dbbc0c0563d1ec2d245b06d1206f2f38c4ed2f0da728e027ae2e5a973", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Mind2Web"}], "name": "Mind2Web", "publication": null, "published_at": null, "record_sha256": "5ff3eb502e4f896644308376cfb09cc24763eacf0ef0c6dbb97df228356b59a2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md", "target_id": "arxiv:2306.06070", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/mind2web", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/Mind2Web", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "Mind2Web\nResource type: benchmark\nReview status: curated\nScope: core\n\nHuman demonstrations and tasks for generalist web navigation research.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"CLICK\", \"TYPE\", \"SELECT\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md\", \"population\": \"train split\", \"qualifier\": \"exact\", \"unit\": \"task instances\", \"value\": 1009}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md\", \"population\": \"cross-task test split\", \"qualifier\": \"exact\", \"unit\": \"task instances\", \"value\": 252}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md\", \"population\": \"cross-website test split\", \"qualifier\": \"exact\", \"unit\": \"task instances\", \"value\": 177}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md\", \"population\": \"cross-domain test split\", \"qualifier\": \"exact\", \"unit\": \"task instances\", \"value\": 912}], \"evaluation\": \"Offline candidate and action evaluation against recorded demonstrations; the repository distinguishes macro from micro averaging.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web/HEAD/README.md\"], \"focus\": [\"web action prediction\", \"generalization to unseen sites and domains\"], \"notes\": \"These are reported split inventories, not online task-success measurements. Original Mind2Web, Multimodal-Mind2Web, and Online-Mind2Web have different inputs/protocols. The authors require macro-average metrics for comparison with the paper; test archives remain upstream.\", \"observations\": [\"task instructions\", \"HTML\", \"recorded action history\"], \"setting\": \"offline\"}}\n\nPrimary source: https://github.com/OSU-NLP-Group/Mind2Web\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-mind2web/"}
{"id": "github:osu-nlp-group/mind2web-2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/Mind2Web-2", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-mind2web-2/", "details": {"leaderboard_url": "https://osu-nlp-group.github.io/Mind2Web-2/#leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "5cc72b992cfc11094dcf341c284d5baadc6c66a302fdfd474b27913fe1c91f6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T17:18:56Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T17:18:56Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Mind2Web-2/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Mind2Web-2"}], "name": "Mind2Web 2", "publication": null, "published_at": null, "record_sha256": "93d277479837a7155ef23658bb90f0df24ed8183cbb1690363b712a8833c810c", "relations": [], "resource_id": "github:osu-nlp-group/mind2web-2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/Mind2Web-2", "topics": ["evaluation", "web-navigation", "long-horizon"], "updated_at": null}, "text": "Mind2Web 2\nResource type: benchmark\nReview status: curated\nScope: core\n\nLong-horizon agentic web search benchmark whose Agent-as-a-Judge framework scores time-varying, citation-backed answers, with evaluation scripts for its dev and test sets.\nPlatforms: web\nTopics: evaluation, web-navigation, long-horizon\n\nRecorded specifications: {\"leaderboard_url\": \"https://osu-nlp-group.github.io/Mind2Web-2/#leaderboard\"}\n\nPrimary source: https://github.com/OSU-NLP-Group/Mind2Web-2\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-mind2web-2/"}
{"id": "github:osu-nlp-group/online-mind2web", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/Online-Mind2Web", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-online-mind2web/", "details": {"benchmark_protocol": {"actions": ["web interactions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md", "population": "documented Online-Mind2Web suite", "qualifier": "exact", "unit": "tasks", "value": 300}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md", "population": "documented site coverage", "qualifier": "exact", "unit": "websites", "value": 136}], "evaluation": "Human evaluation and WebJudge-based evaluation are documented as distinct assessment paths.", "evidence_urls": ["https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md"], "focus": ["live-web task completion"], "notes": "Tasks are actively replaced when websites or requirements become invalid. Record the task revision and submission schema; a constant headline task count does not imply an unchanged test set.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/osunlp/Online_Mind2Web_Leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "49d7cc317fd2ba0df65abf65a0357d1fa2384e59019a96d4e24b784888599eaa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "93675f9d13697ae98889d703b8fae86a7200623e7ae8d844c9ee2200ad480583", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/Online-Mind2Web"}], "name": "Online Mind2Web", "publication": null, "published_at": null, "record_sha256": "9bb7d2e586c0a125f1ac0513b20583d1cc6bf40c5d978aa657ce41cc08e20940", "relations": [], "resource_id": "github:osu-nlp-group/online-mind2web", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/Online-Mind2Web", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "Online Mind2Web\nResource type: benchmark\nReview status: curated\nScope: core\n\n300 tasks on 136 live websites for evaluating web agents online, with outdated tasks replaced over time, human-reviewed submissions and the WebJudge automatic evaluator.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"web interactions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md\", \"population\": \"documented Online-Mind2Web suite\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 300}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md\", \"population\": \"documented site coverage\", \"qualifier\": \"exact\", \"unit\": \"websites\", \"value\": 136}], \"evaluation\": \"Human evaluation and WebJudge-based evaluation are documented as distinct assessment paths.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/OSU-NLP-Group/Online-Mind2Web/HEAD/README.md\"], \"focus\": [\"live-web task completion\"], \"notes\": \"Tasks are actively replaced when websites or requirements become invalid. Record the task revision and submission schema; a constant headline task count does not imply an unchanged test set.\", \"observations\": [], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://huggingface.co/spaces/osunlp/Online_Mind2Web_Leaderboard\"}\n\nPrimary source: https://github.com/OSU-NLP-Group/Online-Mind2Web\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-online-mind2web/"}
{"id": "github:osu-nlp-group/redteamcua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/RedTeamCUA", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-redteamcua/", "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "782dea8635b417bc89d21489cacea215ee6878ba5a1d4b9e18ca7b2913545ab6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "10a9bb268e611da40ae6edba6777917abb6e97d43dcb480e772ae8169b687e10", "supports": ["identity"], "url": "https://api.github.com/repos/OSU-NLP-Group/RedTeamCUA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "712f1b7bc13b593cf23ecfab72cec33d1c256afe23a383e49d7d5b2996a0f55c", "supports": ["release_status"], "url": "https://api.github.com/repos/OSU-NLP-Group/RedTeamCUA/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/RedTeamCUA"}], "name": "RedTeamCUA", "publication": null, "published_at": null, "record_sha256": "62751a2bba158d26c1e6ac20eb11ef164e957f0133a15e5398b1798f0a795026", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/RedTeamCUA/HEAD/README.md", "target_id": "arxiv:2505.21936", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/redteamcua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/RedTeamCUA", "topics": ["safety-security", "evaluation", "sandbox"], "updated_at": null}, "text": "RedTeamCUA\nResource type: framework\nReview status: curated\nScope: core\n\nA controlled testing framework that combines desktop virtual machines and replicated websites to evaluate computer-use agents against indirect prompt injection.\nPlatforms: web, desktop\nTopics: safety-security, evaluation, sandbox\n\nThe release includes environment code, adversarial scenario configuration and RTC-Bench task-generation assets. The project declares Apache-2.0, while incorporated environments retain their licenses. It distinguishes full task runs from tests initialized at injection locations; those settings must not be treated as interchangeable results.\n\nPrimary source: https://github.com/OSU-NLP-Group/RedTeamCUA\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-redteamcua/"}
{"id": "github:osu-nlp-group/seeact", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/SeeAct", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-seeact/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "673b0e309540e57ec3b977689b33f89ef80e45ba1c0ffb05e7412e4862f6b99e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/SeeAct"}], "name": "SeeAct", "publication": null, "published_at": null, "record_sha256": "37035989f30e3d7495ff81e7941ee2806ac697d87e3faa0008bf943c831e9df7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/SeeAct/HEAD/README.md", "target_id": "arxiv:2401.01614", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/seeact", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/SeeAct", "topics": ["web-navigation", "grounding"], "updated_at": null}, "text": "SeeAct\nResource type: agent\nReview status: curated\nScope: core\n\nGeneralist web agent framework for large multimodal models such as GPT-4V, with a Playwright tool that runs agents on live websites, text-choice and set-of-mark grounding, and a crawler mode.\nPlatforms: web\nTopics: web-navigation, grounding\n\nPrimary source: https://github.com/OSU-NLP-Group/SeeAct\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-seeact/"}
{"id": "github:osu-nlp-group/skillweaver", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/SkillWeaver", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-skillweaver/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "e06dbc9d58e713a0019d5040aa21220de85266e80aacc7f2d320a0518c781496", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "4dfa648e22431d1a4903eb1e11f432fbe8a0f39d24e2e3324c300285ffe458fd", "supports": ["identity"], "url": "https://api.github.com/repos/OSU-NLP-Group/SkillWeaver"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "7a1c94c67131a97516d92632c00a2a20d09226fec57bb09e77ff6d84a4e6e79a", "supports": ["release_status"], "url": "https://api.github.com/repos/OSU-NLP-Group/SkillWeaver/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:21Z", "role": "primary", "sha256": "822a60969166874dd8a26273b8d749ef8c69c46137a0915ed7178b3265bd037f", "supports": ["relations"], "url": "https://osu-nlp-group.github.io/SkillWeaver/"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:44:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/SkillWeaver/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/SkillWeaver"}], "name": "SkillWeaver", "publication": null, "published_at": null, "record_sha256": "ab385a68d040d3379349a3dfb0657f006efb6797ebf8f34d005d40ac2ccd179b", "relations": [{"evidence_url": "https://osu-nlp-group.github.io/SkillWeaver/", "target_id": "arxiv:2504.07079", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/skillweaver", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/SkillWeaver", "topics": ["web-navigation", "tool-integration", "data-generation"], "updated_at": null}, "text": "SkillWeaver\nResource type: framework\nReview status: curated\nScope: core\n\nA web-agent framework that explores sites, synthesizes reusable interaction APIs and builds a skill library for later tasks.\nPlatforms: web\nTopics: web-navigation, tool-integration, data-generation\n\nExploration, task execution, browser integration and evaluation code are published under MIT. Synthesized skill files are included, but no trained model weights are implied. The source depends on external web environments and model services; no runtime or browser tasks were executed.\n\nPrimary source: https://github.com/OSU-NLP-Group/SkillWeaver\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-skillweaver/"}
{"id": "github:osu-nlp-group/uground", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/UGround", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-uground/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/LICENSE.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/UGround"}], "name": "UGround", "publication": null, "published_at": null, "record_sha256": "85cfb8591e37ef76bf4aeed55829bd340c8edfbe3ec1a027ffe30a59f94537a2", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "arxiv:2410.05243", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/uground", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/UGround", "topics": ["grounding"], "updated_at": null}, "text": "UGround\nResource type: model\nReview status: curated\nScope: core\n\nCode, training data and weights for universal visual grounding of GUI elements from screenshots, including Qwen2-VL-based UGround-V1 at 2B, 7B and 72B and offline and online evaluation suites.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/OSU-NLP-Group/UGround\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-uground/"}
{"id": "github:osu-nlp-group/webdreamer", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/WebDreamer", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-webdreamer/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "6602cb260b52ae5b2b93db231a51a38d0463c78d8e71321620e05b4fa31caacb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebDreamer/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/WebDreamer"}], "name": "WebDreamer", "publication": null, "published_at": null, "record_sha256": "dc85350fc1987892556d28714fe6a269dc611076d1ad04ee3fe339f10ae69e7b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebDreamer/HEAD/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/webdreamer", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/WebDreamer", "topics": ["world-models", "planning"], "updated_at": null}, "text": "WebDreamer\nResource type: framework\nReview status: curated\nScope: core\n\nWorld-model and planning components that simulate web-action outcomes, with linked training data and model checkpoints.\nPlatforms: web\nTopics: world-models, planning\n\nPrimary source: https://github.com/OSU-NLP-Group/WebDreamer\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-webdreamer/"}
{"id": "github:osu-nlp-group/webguard", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OSU-NLP-Group/WebGuard", "citation_url": "https://awesomecua.github.io/resources/github-osu-nlp-group-webguard/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:53Z", "role": "primary", "sha256": "be53561dcf69cb2295a26eaebf965eead10b3504fca30f71f73dc5e5728fc5b3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "93b43d692b033b76129504448695bfe76ef22d18b11e51352bfab4b5622e5aaa", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OSU-NLP-Group/WebGuard"}], "name": "WebGuard", "publication": null, "published_at": null, "record_sha256": "5031a4cc16bd43854260b9d2c4727d610dae8769434193908767fee902ac7470", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/WebGuard/HEAD/README.md", "target_id": "arxiv:2507.14293", "type": "documented_by"}], "resource_id": "github:osu-nlp-group/webguard", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OSU-NLP-Group/WebGuard", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null}, "text": "WebGuard\nResource type: benchmark\nReview status: curated\nScope: core\n\nCodebase, annotation toolkit, and action-level risk dataset for training and evaluating guardrail models that flag unsafe browser-agent actions before execution.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\n\nPrimary source: https://github.com/OSU-NLP-Group/WebGuard\nCatalog record: https://awesomecua.github.io/resources/github-osu-nlp-group-webguard/"}
{"id": "github:othersideai/self-operating-computer", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/OthersideAI/self-operating-computer", "citation_url": "https://awesomecua.github.io/resources/github-othersideai-self-operating-computer/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8a7f7199f7cf52cb31940bcea8ae797d4ab6dbd2029c3fc9066a17dde3d19095", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "7104ee457c53f2e8e12fd536850e9c2e4894e5e23100f0aa74c8fa5924cec6b2", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/OthersideAI/self-operating-computer/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/OthersideAI/self-operating-computer"}], "name": "Self Operating Computer", "publication": null, "published_at": null, "record_sha256": "b1aff34949cc1f0134940e6a1e844bdf8b889fbc8a3e76a436439e3c0cbddc8b", "relations": [], "resource_id": "github:othersideai/self-operating-computer", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/OthersideAI/self-operating-computer", "topics": ["grounding"], "updated_at": null}, "text": "Self Operating Computer\nResource type: agent\nReview status: curated\nScope: core\n\nFramework, released in November 2023, in which a multimodal model views the screen and picks mouse and keyboard actions toward a goal, with backends such as GPT-4o, Claude 3, Gemini and LLaVA.\nPlatforms: desktop\nTopics: grounding\n\nPrimary source: https://github.com/OthersideAI/self-operating-computer\nCatalog record: https://awesomecua.github.io/resources/github-othersideai-self-operating-computer/"}
{"id": "github:pathonaiorg/litewebagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/PathOnAIOrg/LiteWebAgent", "citation_url": "https://awesomecua.github.io/resources/github-pathonaiorg-litewebagent/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "75f99d9388aee3e0b124ebdb5f33e08fbe8ec999ce1e9eba5cd71c21e60522e0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/PathOnAI/LiteWebAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "ec6ef142e3900e9f5555009b825387e41da67a77a24ee72ef6963c99b065e6eb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/PathOnAIOrg/LiteWebAgent/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/PathOnAIOrg/LiteWebAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PathOnAIOrg/LiteWebAgent"}], "name": "LiteWebAgent", "publication": null, "published_at": null, "record_sha256": "b89b32b881cb6227d1d9b905c3de74b5e8e9445b5f2e6614c271d78f5f837556", "relations": [{"evidence_url": "https://raw.githubusercontent.com/PathOnAI/LiteWebAgent/HEAD/README.md", "target_id": "arxiv:2503.02950", "type": "documented_by"}], "resource_id": "github:pathonaiorg/litewebagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/PathOnAIOrg/LiteWebAgent", "topics": ["web-navigation", "planning", "memory", "tool-integration"], "updated_at": null}, "text": "LiteWebAgent\nResource type: framework\nReview status: curated\nScope: core\n\nModular web-agent framework with browser control, planning, workflow memory, and tree-search components, plus web and extension interfaces.\nPlatforms: web\nTopics: web-navigation, planning, memory, tool-integration\n\nThe repository is maintained by the PathOnAI.org research community; its README explicitly disclaims affiliation with a for-profit company.\n\nPrimary source: https://github.com/PathOnAIOrg/LiteWebAgent\nCatalog record: https://awesomecua.github.io/resources/github-pathonaiorg-litewebagent/"}
{"id": "github:patrickstarl/gui-grounding-eval", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/PatrickStarL/gui-grounding-eval", "citation_url": "https://awesomecua.github.io/resources/github-patrickstarl-gui-grounding-eval/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:51:38.123692Z", "role": "primary", "sha256": "3226df7608039522072e59d459094fc45be9c5a6be7ea0d7ad76ddf9723a3aea", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:04:53.470789+00:00", "role": "primary", "sha256": "3d76d934e98641746085772b98ee17059749f980a616435498a7ad483bd4cec1", "supports": ["id", "url"], "url": "https://api.github.com/search/repositories?q=topic%3Agui-grounding+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:51:05.389053Z", "role": "primary", "sha256": "4fe6f65a21586dde0bf8b64f8c49e5c583b320bf5e0c2b8253baf2cd3fcfc5d2", "supports": ["id", "url", "notes"], "url": "https://github.com/PatrickStarL/gui-grounding-eval"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:32.022035Z", "role": "primary", "sha256": "c3eb1708b7d57dcde7478ca2f5063241b1ab567b2ba2ba28b9956b043da2b6c2", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/metrics.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:35.461491Z", "role": "primary", "sha256": "66fa5940f0c4f6265cb162cb0596402848dd2a49fd48b07ee950dc866b8730a5", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/geometry.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:53:38.931130Z", "role": "primary", "sha256": "f83bc0632a616e7b01c8a46b0e435733f0cc47b4a19f0c709b221eaee8a0db5f", "supports": ["summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/PatrickStarL/gui-grounding-eval/d92b059caa0e0acf8db76f2164dd7be2617ccc72/ssp_eval/runner.py"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-20T11:51:38.123692Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PatrickStarL/gui-grounding-eval"}], "name": "GUI Grounding Evaluation Harness", "publication": null, "published_at": null, "record_sha256": "c9aa2a61f7563bd66236bef559cff49f42aeba0e1c593bb2ef4119faf7ac175b", "relations": [], "resource_id": "github:patrickstarl/gui-grounding-eval", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/PatrickStarL/gui-grounding-eval", "topics": ["grounding", "evaluation", "efficiency"], "updated_at": null}, "text": "GUI Grounding Evaluation Harness\nResource type: tool\nReview status: curated\nScope: core\n\nScreenSpot-Pro evaluation harness with explicit coordinate transforms, prompt and crop-based inference comparisons, and per-sample grounding accuracy and cost records.\nPlatforms: unspecified\nTopics: grounding, evaluation, efficiency\n\nThis is an independent harness using ScreenSpot-Pro, not a new dataset or verified leaderboard entry. The read code maps points back to original-image coordinates and scores point-in-box hits; documentation limits comparisons by prompt, model, subset and provider behavior. No OS support, code license or external data/weight license was established in this bounded review. Model calls, tests and reported results were not executed or reproduced.\n\nPrimary source: https://github.com/PatrickStarL/gui-grounding-eval\nCatalog record: https://awesomecua.github.io/resources/github-patrickstarl-gui-grounding-eval/"}
{"id": "github:pixiebrix/agent-browser-shield", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/pixiebrix/agent-browser-shield", "citation_url": "https://awesomecua.github.io/resources/github-pixiebrix-agent-browser-shield/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:05Z", "role": "primary", "sha256": "8682f83d71dd4d44772db3e531f2a9fcb168729211e0cc43493020a0773b0aba", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/pixiebrix/agent-browser-shield/HEAD/README.md"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:05Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/pixiebrix/agent-browser-shield"}], "name": "Agent Browser Shield", "publication": null, "published_at": null, "record_sha256": "64b606f0d5fe200a2fdf5a5d9bcf687ceb5abb44683878c6c785fef21e01e287", "relations": [], "resource_id": "github:pixiebrix/agent-browser-shield", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/pixiebrix/agent-browser-shield", "topics": ["safety-security"], "updated_at": null}, "text": "Agent Browser Shield\nResource type: tool\nReview status: curated\nScope: core\n\nBrowser extension for filtering potentially unsafe agent input.\nPlatforms: web\nTopics: safety-security\n\nPrimary source: https://github.com/pixiebrix/agent-browser-shield\nCatalog record: https://awesomecua.github.io/resources/github-pixiebrix-agent-browser-shield/"}
{"id": "github:posgnu/rci-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/posgnu/rci-agent", "citation_url": "https://awesomecua.github.io/resources/github-posgnu-rci-agent/", "details": {"availability": "Public agent and browser-environment source; legacy API compatibility is unverified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"], "reason": "Repository code license; not a license for the remote language-model service."}}, "repository_revision": "31bf737922fa0337d03367a844de5cf1aff4ac52"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "6a101c867546365d21b48527781bd3af324075929067fb465c6105872070f692", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:11Z", "role": "primary", "sha256": "b826fc1747a1bbe38a8df8da567a68f79936509c5325ec76a683706118c526f5", "supports": ["identity"], "url": "https://api.github.com/repos/posgnu/rci-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:11Z", "role": "primary", "sha256": "3dcb8cc06679490315f0f3636bb2795e2f717b77d4156ad0bf2429cc36e3273c", "supports": ["release_status"], "url": "https://api.github.com/repos/posgnu/rci-agent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ffa8b2a603abd9ede8eb686e87c4a433110f1be47d7cc98df4977b7fb2ee72fc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-17T09:14:12Z", "licenses": {"code": {"basis": "Repository code license; not a license for the remote language-model service.", "declared": "MIT", "retrieved_at": "2026-10-17T09:14:12Z", "url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/posgnu/rci-agent"}], "name": "RCI Agent", "publication": null, "published_at": null, "record_sha256": "a8466ed0ab2bfdf646c5d1b45220b398bac89364f1982523c3906e1e13d6b9da", "relations": [{"evidence_url": "https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/README.md", "target_id": "arxiv:2303.17491", "type": "documented_by"}], "resource_id": "github:posgnu/rci-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/posgnu/rci-agent", "topics": ["web-navigation", "planning", "self-reflection", "training-free"], "updated_at": null}, "text": "RCI Agent\nResource type: agent\nReview status: curated\nScope: core\n\nA MiniWoB++ agent that uses explicit and implicit recursive criticism-and-improvement prompts to refine action plans and ground actions to browser state.\nPlatforms: web\nTopics: web-navigation, planning, self-reflection, training-free\n\nRecorded specifications: {\"availability\": \"Public agent and browser-environment source; legacy API compatibility is unverified.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/posgnu/rci-agent/HEAD/LICENSE\"], \"reason\": \"Repository code license; not a license for the remote language-model service.\"}}, \"repository_revision\": \"31bf737922fa0337d03367a844de5cf1aff4ac52\"}\n\nThe public release includes the agent, prompting utilities and a bundled MiniWoB++ environment wrapper. Its README documents older hosted model names and an OpenAI-key setup; contemporary endpoint compatibility was not tested. The repository notes that its MiniWoB++ dependency is not officially supported on Windows. MIT applies to repository code; no new trained weights are claimed.\n\nPrimary source: https://github.com/posgnu/rci-agent\nCatalog record: https://awesomecua.github.io/resources/github-posgnu-rci-agent/"}
{"id": "github:prentisai/groundgui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/PrentisAI/GroundGUI", "citation_url": "https://awesomecua.github.io/resources/github-prentisai-groundgui/", "details": {}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:05Z", "role": "primary", "sha256": "a495bb4c86cf54cd5e2276433cfecb401fd133287406b1a7e04c3eaaac6305ac", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T01:52:51Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T01:46:05Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T01:52:51Z", "url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/PrentisAI/GroundGUI"}], "name": "GroundGUI training and evaluation code", "publication": null, "published_at": null, "record_sha256": "982855ce8f4bde76bcb12fe22d154b6edd4cf20d8032cbaa80e910ac20ed6429", "relations": [{"evidence_url": "https://raw.githubusercontent.com/PrentisAI/GroundGUI/HEAD/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "resource_id": "github:prentisai/groundgui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/PrentisAI/GroundGUI", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null}, "text": "GroundGUI training and evaluation code\nResource type: framework\nReview status: curated\nScope: core\n\nCode for screenshot grounding through supervised training, teacher-guided distillation, reinforcement learning, and benchmark evaluation.\nPlatforms: desktop, web, cross-platform\nTopics: grounding, reinforcement-learning, data-generation\n\nThe repository documents separate training-stage environments. Code, model weights, and ScreenRef data have distinct license declarations; no training or benchmark execution was reproduced here.\n\nPrimary source: https://github.com/PrentisAI/GroundGUI\nCatalog record: https://awesomecua.github.io/resources/github-prentisai-groundgui/"}
{"id": "github:princeton-nlp/webshop", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/princeton-nlp/WebShop", "citation_url": "https://awesomecua.github.io/resources/github-princeton-nlp-webshop/", "details": {"benchmark_protocol": {"actions": ["search queries", "click actions", "product option selection"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "population": "crowd-sourced instruction collection", "qualifier": "exact", "unit": "instructions", "value": 12087}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "population": "full product inventory", "qualifier": "approximately", "unit": "products", "value": 1180000}], "evaluation": "", "evidence_urls": ["https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"], "focus": ["instruction-conditioned product selection", "web navigation"], "notes": "This is a simulated shop. The default preview loads 1,000 products, while the full inventory is much larger; report the loaded subset and observation mode. Instructions and products are different populations. Detailed reward semantics are not normalized in this profile.", "observations": ["HTML pages", "simplified text views"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "79b7a90a1413a52deb53142b5bbd81ecf40f61743bd4601c2835e91375898ec0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "79b7a90a1413a52deb53142b5bbd81ecf40f61743bd4601c2835e91375898ec0", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "8872dbf8660b00890b5e07ce1ea1f7a44fa7ae4e1857da56caba172b51dab3cb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/LICENSE.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/princeton-nlp/WebShop"}], "name": "WebShop", "publication": null, "published_at": null, "record_sha256": "e22dbc44399c2ac22982b3b7e7a2053037e29659920fb1be65946f5c90694e5b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md", "target_id": "arxiv:2207.01206", "type": "documented_by"}], "resource_id": "github:princeton-nlp/webshop", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/princeton-nlp/WebShop", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "WebShop\nResource type: benchmark\nReview status: curated\nScope: core\n\nSimulated shopping environment for instruction-following web agents.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"search queries\", \"click actions\", \"product option selection\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md\", \"population\": \"crowd-sourced instruction collection\", \"qualifier\": \"exact\", \"unit\": \"instructions\", \"value\": 12087}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md\", \"population\": \"full product inventory\", \"qualifier\": \"approximately\", \"unit\": \"products\", \"value\": 1180000}], \"evaluation\": \"\", \"evidence_urls\": [\"https://raw.githubusercontent.com/princeton-nlp/WebShop/HEAD/README.md\"], \"focus\": [\"instruction-conditioned product selection\", \"web navigation\"], \"notes\": \"This is a simulated shop. The default preview loads 1,000 products, while the full inventory is much larger; report the loaded subset and observation mode. Instructions and products are different populations. Detailed reward semantics are not normalized in this profile.\", \"observations\": [\"HTML pages\", \"simplified text views\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/princeton-nlp/WebShop\nCatalog record: https://awesomecua.github.io/resources/github-princeton-nlp-webshop/"}
{"id": "github:purewhiter/mobilegym", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Purewhiter/mobilegym", "citation_url": "https://awesomecua.github.io/resources/github-purewhiter-mobilegym/", "details": {"availability": "Public simulator, task/judge and online-RL source, plus official release metadata for the companion data archive.", "benchmark_protocol": {"actions": ["tap", "type", "swipe", "back", "home", "drag", "complete"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "task templates", "value": 416}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "simulated apps", "value": 28}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt", "population": "bench_env/splits/train.txt", "qualifier": "exact", "unit": "task identifiers", "value": 160}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt", "population": "bench_env/splits/test.txt", "qualifier": "exact", "unit": "task identifiers", "value": 256}], "evaluation": "Deterministic task judges inspect state changes; reported metrics distinguish success, progress, false completion and unexpected side effects.", "evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt", "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt"], "focus": ["mobile GUI task completion in simulation", "state verification and parallel training"], "notes": "The split-ID inventory is not a count of sampled trajectories, live Android apps or real user accounts. Extra runtime/data dependencies and third-party content terms remain distinct.", "observations": ["simulated mobile screenshots", "structured simulator state for judges"], "setting": "interactive"}, "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"], "reason": "Core source LICENSE; mobilegym-rl has a separate matching Apache-2.0 license and retains vendored notices."}, "data": {"evidence_urls": ["https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"], "reason": "Limited to maintainers’ dataset compilation/curation and authored synthetic content; third-party user content remains subject to platform terms and its authors’ rights."}}, "release_tag": "v0.1.0", "repository_revision": "5fbd3f34be0e9780ae4b1e9d5089963ddbd05d38"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "815294dd2aa4bf8264978c3bfedee85a3a2c835c511860c732314a4c282a974a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "48df04a2fbf479f956d5c7f22d7006c22f591b8857cab8a1d64c196349045495", "supports": ["identity"], "url": "https://api.github.com/repos/Purewhiter/mobilegym"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:56Z", "role": "primary", "sha256": "68f1fa16c0999079bccc3aa1b6c3c22b72b570a5b971837c5026ea5408853375", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:08Z", "role": "primary", "sha256": "cb1057b0a0a2ef7ab781414c168bb41f0f3da88afc042ce266d6063493561bde", "supports": ["release_status", "details.release_tag"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/releases/tags/v0.1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:09Z", "role": "primary", "sha256": "ad0c86a6cdde2caba8bd1f98b121803817d93718319a863b0b833da2760c23aa", "supports": ["release_status"], "url": "https://api.github.com/repos/Purewhiter/mobilegym/releases/tags/data-v0.1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "7618cd7de039c264759ca8ca0b9130c6de63a4536dcf436baedd4cfbf639073e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:50Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/mobilegym-rl/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "e222fa518c5003d425ee9b371e987c95ba10e52b54fea078571254ae6a8eaf23", "supports": ["licenses.data", "notes"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:49Z", "role": "primary", "sha256": "0ecef460fecfd8aa91dcbb2cd4cca17b033f620f85a50d7f0b598ed19a27c0eb", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:27:48Z", "role": "primary", "sha256": "a3d485015d480f4f1ebbd12dbbf02c25c0aa667d1afcf5f6fa35f0263a82c567", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:19Z", "licenses": {"code": {"basis": "Core source LICENSE; mobilegym-rl has a separate matching Apache-2.0 license and retains vendored notices.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:22:59Z", "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE"}, "data": {"basis": "Limited to maintainers’ dataset compilation/curation and authored synthetic content; third-party user content remains subject to platform terms and its authors’ rights.", "declared": "CC-BY-NC-4.0", "retrieved_at": "2026-10-17T11:22:59Z", "url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Purewhiter/mobilegym"}], "name": "MobileGym", "publication": null, "published_at": null, "record_sha256": "34b507eb89bc567e5551d1c3a33761212346177872071819b84310d067f3d1df", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md", "target_id": "arxiv:2605.26114", "type": "documented_by"}], "resource_id": "github:purewhiter/mobilegym", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Purewhiter/mobilegym", "topics": ["evaluation", "mobile-navigation", "reinforcement-learning", "sandbox"], "updated_at": null}, "text": "MobileGym\nResource type: benchmark\nReview status: curated\nScope: core\n\nA browser-hosted mobile-app simulator with programmable state, parameterized tasks and deterministic state-checking judges for parallel GUI-agent evaluation and online reinforcement learning.\nPlatforms: web\nTopics: evaluation, mobile-navigation, reinforcement-learning, sandbox\n\nRecorded specifications: {\"availability\": \"Public simulator, task/judge and online-RL source, plus official release metadata for the companion data archive.\", \"benchmark_protocol\": {\"actions\": [\"tap\", \"type\", \"swipe\", \"back\", \"home\", \"drag\", \"complete\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md\", \"population\": \"publisher benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"task templates\", \"value\": 416}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md\", \"population\": \"publisher benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"simulated apps\", \"value\": 28}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt\", \"population\": \"bench_env/splits/train.txt\", \"qualifier\": \"exact\", \"unit\": \"task identifiers\", \"value\": 160}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt\", \"population\": \"bench_env/splits/test.txt\", \"qualifier\": \"exact\", \"unit\": \"task identifiers\", \"value\": 256}], \"evaluation\": \"Deterministic task judges inspect state changes; reported metrics distinguish success, progress, false completion and unexpected side effects.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/README.md\", \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/train.txt\", \"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/bench_env/splits/test.txt\"], \"focus\": [\"mobile GUI task completion in simulation\", \"state verification and parallel training\"], \"notes\": \"The split-ID inventory is not a count of sampled trajectories, live Android apps or real user accounts. Extra runtime/data dependencies and third-party content terms remain distinct.\", \"observations\": [\"simulated mobile screenshots\", \"structured simulator state for judges\"], \"setting\": \"interactive\"}, \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE\"], \"reason\": \"Core source LICENSE; mobilegym-rl has a separate matching Apache-2.0 license and retains vendored notices.\"}, \"data\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Purewhiter/mobilegym/HEAD/LICENSE-DATA\"], \"reason\": \"Limited to maintainers’ dataset compilation/curation and authored synthetic content; third-party user content remains subject to platform terms and its authors’ rights.\"}}, \"release_tag\": \"v0.1.0\", \"repository_revision\": \"5fbd3f34be0e9780ae4b1e9d5089963ddbd05d38\"}\n\nThe inspected code release is v0.1.0, with a separately listed data-v0.1.0 archive. The interaction target is a browser-hosted Android-like simulation; real-device transfer results are separate publisher-reported experiments. Source split manifests contain 160 training IDs and 256 test IDs with no duplicates or overlap. Platform/training code declares Apache-2.0. LICENSE-DATA applies CC-BY-NC-4.0 only to maintainer-created compilation, curation and synthetic content, and explicitly leaves third-party UGC/API terms and author rights separate. No runtime, archive or model was executed/downloaded.\n\nPrimary source: https://github.com/Purewhiter/mobilegym\nCatalog record: https://awesomecua.github.io/resources/github-purewhiter-mobilegym/"}
{"id": "github:radishsoftware/mobster", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/RadishSoftware/mobster", "citation_url": "https://awesomecua.github.io/resources/github-radishsoftware-mobster/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"], "reason": "Complete root MIT license text inspected at the pinned commit. Applies to repository software; data, weights, external WebDriverAgent code and separately distributed products do not inherit this declaration."}}, "repository_revision": "738cecdd3eacb9df88d48efb9efa53507bb4aeb4"}, "environments": ["ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:18.994340+00:00", "role": "primary", "sha256": "dfffdc4b5850ed6bfe0bcfe705fa5918d61420f8c80df2e38f61b7d1fb3819ec", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:06:00.276978+00:00", "role": "primary", "sha256": "01f97a863952373bd10a1f849d79706344be63f5774fe7e0cfe9ba774e41d213", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Amobile-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:17.260430+00:00", "role": "primary", "sha256": "fde31e2fdae142655f09fa14fb3114bf39a22cc1ff7994204a66622f5cbfe2c7", "supports": ["identity", "details.repository_revision"], "url": "https://api.github.com/repos/RadishSoftware/mobster/commits/main"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:37.457839+00:00", "role": "primary", "sha256": "8c2b9647b33a93b1a33e9256307e09b0f3e6aa62608019defa6dcdc5d7bff3dc", "supports": ["identity", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/RadishSoftware/mobster/git/trees/81f2c1990c07a3aaf001a5c3eb4f37a457f0fb3d?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:25.539870+00:00", "role": "primary", "sha256": "f9ad743798f16491100ea692069e346f82bbdb5a5df91e4e01c1d6168f2513cf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:05:06.711971+00:00", "role": "primary", "sha256": "4296a186cc2c00b9de93ababae490dc99cba596d858eff195e40283d58d73a0a", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/mobile_agent/drivers.py"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-20T11:57:18.994340+00:00", "licenses": {"code": {"basis": "Complete root MIT license text inspected at the pinned commit. Applies to repository software; data, weights, external WebDriverAgent code and separately distributed products do not inherit this declaration.", "declared": "MIT", "retrieved_at": "2026-10-20T11:57:25.539870+00:00", "url": "https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/RadishSoftware/mobster"}], "name": "Mobster engine", "publication": null, "published_at": null, "record_sha256": "c8b8345da5ab48862963324c61a59b332c29d7a65abbddbed97c5ad33c85b82e", "relations": [], "resource_id": "github:radishsoftware/mobster", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/RadishSoftware/mobster", "topics": ["mobile-navigation", "evaluation"], "updated_at": null}, "text": "Mobster engine\nResource type: agent\nReview status: curated\nScope: core\n\niOS agent engine organized around an observe, decide, verify and act loop, using WebDriverAgent screen state and phone actions such as taps, typing, swipes and app switches.\nPlatforms: ios\nTopics: mobile-navigation, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/RadishSoftware/mobster/738cecdd3eacb9df88d48efb9efa53507bb4aeb4/LICENSE\"], \"reason\": \"Complete root MIT license text inspected at the pinned commit. Applies to repository software; data, weights, external WebDriverAgent code and separately distributed products do not inherit this declaration.\"}}, \"repository_revision\": \"738cecdd3eacb9df88d48efb9efa53507bb4aeb4\"}\n\nThis entry covers the published engine, not the separate Mobster for Mac product. The documented reference example requires an already-serving WebDriverAgent runner for an iPhone and a model-service key; it performs a fixed read-only Settings task and declines approvals. The repository lists benchmark/evaluation tooling, but dated simulator subsets, real-device trials and earlier engine results are different evaluation populations. The README explicitly places several product modules outside this repository. Source inspection confirms WDA dispatch paths, not task success, security guarantees or reproducible performance; no phone, simulator, model or benchmark was run. MIT applies to the inspected repository software, not automatically to external WDA code, images or other data.\n\nPrimary source: https://github.com/RadishSoftware/mobster\nCatalog record: https://awesomecua.github.io/resources/github-radishsoftware-mobster/"}
{"id": "github:reallm-labs/infigui-r1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Reallm-Labs/InfiGUI-R1", "citation_url": "https://awesomecua.github.io/resources/github-reallm-labs-infigui-r1/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "5fd1399c88bdf406e7689765ccf0599b3b7e0ddbed082fa92bffdd7b120bd41a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Reallm-Labs/InfiGUI-R1"}], "name": "InfiGUI-R1", "publication": null, "published_at": null, "record_sha256": "48a38e4f64eb1fa59fa592da55a85f06b569137eb81c68d29d266fd3351e350a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md", "target_id": "arxiv:2504.14239", "type": "documented_by"}], "resource_id": "github:reallm-labs/infigui-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Reallm-Labs/InfiGUI-R1", "topics": ["reinforcement-learning", "planning"], "updated_at": null}, "text": "InfiGUI-R1\nResource type: model\nReview status: curated\nScope: core\n\n3B GUI agent trained in two stages: distilling spatial reasoning from teacher models, then reinforcement learning that rewards accurate sub-goals and trains error recovery.\nPlatforms: cross-platform\nTopics: reinforcement-learning, planning\n\nPrimary source: https://github.com/Reallm-Labs/InfiGUI-R1\nCatalog record: https://awesomecua.github.io/resources/github-reallm-labs-infigui-r1/"}
{"id": "github:reidbarber/webmarker", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/reidbarber/webmarker", "citation_url": "https://awesomecua.github.io/resources/github-reidbarber-webmarker/", "details": {"availability": "Public browser-library source and npm package metadata.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"], "reason": "Publisher source license for WebMarker."}}, "package_version": "0.7.0", "repository_revision": "03cf632eef3f76bfbc4fd9c23b31baa0355e41c7"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "b46edb85d41e033f84bdd9bb96122f801364448f28a36fa6deccc110a19b6ffc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "a44253aa866acd38fc3e2275aae61ecd051f956b18cb7b29346c699cf6241d6b", "supports": ["identity"], "url": "https://api.github.com/repos/reidbarber/webmarker"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "fc8de18814136ca3684749363727f263908c23e6ae274a0ed589923774aa3a95", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/reidbarber/webmarker/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1caae04ee16fcc72c48439677f5aba55128879cd99d6d55fd72caa85a7030e69", "supports": ["identity", "release_status", "details.package_version"], "url": "https://registry.npmjs.org/webmarker-js/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "62dcdb525872e3f49d90ff9b354b79341ac2574522bca7c7146f79040491a627", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T12:30:56Z", "licenses": {"code": {"basis": "Publisher source license for WebMarker.", "declared": "MIT", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/reidbarber/webmarker"}], "name": "WebMarker", "publication": null, "published_at": null, "record_sha256": "54580bbb96f051570075d018a4719ddd2958e6337a1ac82ea8379254cca35c65", "relations": [], "resource_id": "github:reidbarber/webmarker", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/reidbarber/webmarker", "topics": ["grounding", "screen-understanding", "tool-integration"], "updated_at": null}, "text": "WebMarker\nResource type: tool\nReview status: curated\nScope: core\n\nA browser library that labels interactive elements and their boxes for visual prompting, preserving element mappings that automation code can use to execute selected actions.\nPlatforms: web\nTopics: grounding, screen-understanding, tool-integration\n\nRecorded specifications: {\"availability\": \"Public browser-library source and npm package metadata.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/reidbarber/webmarker/HEAD/LICENSE\"], \"reason\": \"Publisher source license for WebMarker.\"}}, \"package_version\": \"0.7.0\", \"repository_revision\": \"03cf632eef3f76bfbc4fd9c23b31baa0355e41c7\"}\n\nSource and npm release metadata for webmarker-js 0.7.0 are public. The documented mark/unmark API supports screenshot prompts and Playwright interaction, but this utility is not itself a language model or autonomous agent. The WebVoyager/Set-of-Mark references are usage inspiration and are not asserted as papers introducing this library. No package was installed or executed.\n\nPrimary source: https://github.com/reidbarber/webmarker\nCatalog record: https://awesomecua.github.io/resources/github-reidbarber-webmarker/"}
{"id": "github:reworkd/tarsier", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/reworkd/tarsier", "citation_url": "https://awesomecua.github.io/resources/github-reworkd-tarsier/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "2a313a5a1c433c0609067c6d875d5ff69493e8792d2173e458eb5db8bf232977", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "f4ff2ba0f16e8abb7087071651c217ced6421e224313755c5d88433e324f19a1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/reworkd/tarsier/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/reworkd/tarsier"}], "name": "Tarsier", "publication": null, "published_at": null, "record_sha256": "f919f61ad282132c281ec742554ad709ab9c49723f308ca8833019ef07af63e8", "relations": [], "resource_id": "github:reworkd/tarsier", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/reworkd/tarsier", "topics": ["screen-understanding"], "updated_at": null}, "text": "Tarsier\nResource type: tool\nReview status: curated\nScope: core\n\nPerception utilities for web agents that tag interactable elements with typed IDs an LLM can act on, and use OCR to render page screenshots as whitespace-structured text for models without vision.\nPlatforms: web\nTopics: screen-understanding\n\nPrimary source: https://github.com/reworkd/tarsier\nCatalog record: https://awesomecua.github.io/resources/github-reworkd-tarsier/"}
{"id": "github:ritzz-ai/gui-r1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ritzz-ai/GUI-R1", "citation_url": "https://awesomecua.github.io/resources/github-ritzz-ai-gui-r1/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ritzz-ai/GUI-R1"}], "name": "GUI-R1", "publication": null, "published_at": null, "record_sha256": "267173a9a6a40ff8dde02af33a3f14172c388905e6285f0a05955695fe4df64c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "resource_id": "github:ritzz-ai/gui-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ritzz-ai/GUI-R1", "topics": ["reinforcement-learning", "grounding"], "updated_at": null}, "text": "GUI-R1\nResource type: model\nReview status: curated\nScope: core\n\nGUI action-model training and evaluation code with verifiable rewards, plus linked reinforcement-learning data and checkpoints.\nPlatforms: cross-platform\nTopics: reinforcement-learning, grounding\n\nPrimary source: https://github.com/ritzz-ai/GUI-R1\nCatalog record: https://awesomecua.github.io/resources/github-ritzz-ai-gui-r1/"}
{"id": "github:runamu/monday", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/runamu/monday", "citation_url": "https://awesomecua.github.io/resources/github-runamu-monday/", "details": {}, "environments": ["android", "ios", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "cb3d2329571a80b6edeab98c04f8b63c7f4e976929788d6a8283238b0b7fdadc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T11:03:47Z", "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/runamu/monday"}], "name": "MONDAY", "publication": null, "published_at": null, "record_sha256": "9e494cedb9cb77d4c8cc86e919fab1b58f2c63e494c3e4673ad7cd524e4f7f8a", "relations": [{"evidence_url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md", "target_id": "arxiv:2505.12632", "type": "documented_by"}], "resource_id": "github:runamu/monday", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/runamu/monday", "topics": ["data-generation", "mobile-navigation", "video"], "updated_at": null}, "text": "MONDAY\nResource type: dataset\nReview status: curated\nScope: core\n\nCross-platform mobile GUI trajectory dataset derived from instructional videos, with released data-processing helpers and example evaluation code for Android and iOS data.\nPlatforms: android, ios, cross-platform\nTopics: data-generation, mobile-navigation, video\n\nThe retained README marks release of the full data-generation pipeline, training scripts, and model checkpoints as pending. Linked video content and datasets remain subject to their original licenses; the repository code license does not establish a blanket data or weights license.\n\nPrimary source: https://github.com/runamu/monday\nCatalog record: https://awesomecua.github.io/resources/github-runamu-monday/"}
{"id": "github:saagent/mcpworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/SAAgent/MCPWorld", "citation_url": "https://awesomecua.github.io/resources/github-saagent-mcpworld/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5559ca58f8ee06d2812c6cbc8c3e49ef112f4ef41fdcf4e95568b268bb24bf14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:50Z", "role": "primary", "sha256": "26fb4922c4c0bb5c9069935ed0aef0a9f56938cf8ae1af9724894fa1409955fb", "supports": ["relations", "notes"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "5559ca58f8ee06d2812c6cbc8c3e49ef112f4ef41fdcf4e95568b268bb24bf14", "supports": ["notes"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "d7a04bf11ce24a7b79cd44b4a78dcc9971eaeb3b565821de437e0ecb7723e634", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/SAAgent/MCPWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SAAgent/MCPWorld"}], "name": "MCPWorld", "publication": null, "published_at": null, "record_sha256": "6cf930cd174630fb9de44cb12d2e85eb4af917cdb817d4c48de0a9e925e4ddf6", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=600&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2506.07672", "type": "documented_by"}], "resource_id": "github:saagent/mcpworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/SAAgent/MCPWorld", "topics": ["evaluation", "tool-integration"], "updated_at": null}, "text": "MCPWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation environment combining computer interaction with tool access.\nPlatforms: cross-platform\nTopics: evaluation, tool-integration\n\nThe retained paper reports 201 tasks, whereas the retained repository README describes approximately 170. These are separate source descriptions; their release/version difference has not been established.\n\nPrimary source: https://github.com/SAAgent/MCPWorld\nCatalog record: https://awesomecua.github.io/resources/github-saagent-mcpworld/"}
{"id": "github:salt-nlp/popupattack", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/SALT-NLP/PopupAttack", "citation_url": "https://awesomecua.github.io/resources/github-salt-nlp-popupattack/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "24251854f9481af5c8db65750197604546f000b54868930a6de79adc365db854", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SALT-NLP/PopupAttack/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SALT-NLP/PopupAttack"}], "name": "PopupAttack", "publication": null, "published_at": null, "record_sha256": "a34ff95b6b3e34b3dd5700f5e76482fdf79295ddb92bc6ed6eb5ae93718a55c4", "relations": [], "resource_id": "github:salt-nlp/popupattack", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/SALT-NLP/PopupAttack", "topics": ["safety-security"], "updated_at": null}, "text": "PopupAttack\nResource type: benchmark\nReview status: curated\nScope: core\n\nResources for evaluating GUI agents under malicious popup instructions.\nPlatforms: web\nTopics: safety-security\n\nPrimary source: https://github.com/SALT-NLP/PopupAttack\nCatalog record: https://awesomecua.github.io/resources/github-salt-nlp-popupattack/"}
{"id": "github:scaleapi/browser-art", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/scaleapi/browser-art", "citation_url": "https://awesomecua.github.io/resources/github-scaleapi-browser-art/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md", "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"], "reason": "The retained README states that the project is MIT licensed, while the retained LICENSE contains Creative Commons Attribution-NonCommercial-NoDerivatives 4.0. These declarations conflict, so a settled code license is not asserted."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "a9d33536fae38485eb6aac9698ae22ed3533510d43f783d6c0e9f9cf95d9c58f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code", "notes"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "f7f28b8c7a1af76b9874ca6d040e8b1eb6768fd1043106c9d315090ea96754e7", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "f7f28b8c7a1af76b9874ca6d040e8b1eb6768fd1043106c9d315090ea96754e7", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/scaleapi/browser-art"}], "name": "BrowserART", "publication": null, "published_at": null, "record_sha256": "20b977b6f96a8fec2375bad696228b3e3a0d884b8d2dd5dea3ac58bdf3178794", "relations": [{"evidence_url": "https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md", "target_id": "arxiv:2410.13886", "type": "documented_by"}], "resource_id": "github:scaleapi/browser-art", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/scaleapi/browser-art", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null}, "text": "BrowserART\nResource type: benchmark\nReview status: curated\nScope: core\n\nRed-teaming benchmark and test suite from Scale AI evaluating whether refusal-trained language models can be jailbroken when deployed as browser agents.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/README.md\", \"https://raw.githubusercontent.com/scaleapi/browser-art/HEAD/LICENSE\"], \"reason\": \"The retained README states that the project is MIT licensed, while the retained LICENSE contains Creative Commons Attribution-NonCommercial-NoDerivatives 4.0. These declarations conflict, so a settled code license is not asserted.\"}}}\n\nThe retained README and LICENSE contain conflicting code-license declarations. The catalog leaves the code-license value unknown pending an upstream clarification; it does not infer data or model permissions.\n\nPrimary source: https://github.com/scaleapi/browser-art\nCatalog record: https://awesomecua.github.io/resources/github-scaleapi-browser-art/"}
{"id": "github:segev-shlomov/st-webagentbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/segev-shlomov/ST-WebAgentBench", "citation_url": "https://awesomecua.github.io/resources/github-segev-shlomov-st-webagentbench/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:54Z", "role": "primary", "sha256": "e712feb15886d5a2aba1051d419dc69dc6e58c9a86a9d2d566090bbbf67492fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/segev-shlomov/ST-WebAgentBench/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/segev-shlomov/ST-WebAgentBench"}], "name": "ST-WebAgentBench", "publication": null, "published_at": null, "record_sha256": "a2162bc7a3e3f75dcff35b85feaad461d294825ee9304926639170fe3aead5dd", "relations": [{"evidence_url": "https://raw.githubusercontent.com/segev-shlomov/ST-WebAgentBench/HEAD/README.md", "target_id": "arxiv:2410.06703", "type": "documented_by"}], "resource_id": "github:segev-shlomov/st-webagentbench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/segev-shlomov/ST-WebAgentBench", "topics": ["safety-security", "evaluation", "web-navigation"], "updated_at": null}, "text": "ST-WebAgentBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nBrowserGym-based benchmark for evaluating the safety and policy compliance of enterprise web agents across six trustworthiness risk dimensions.\nPlatforms: web\nTopics: safety-security, evaluation, web-navigation\n\nPrimary source: https://github.com/segev-shlomov/ST-WebAgentBench\nCatalog record: https://awesomecua.github.io/resources/github-segev-shlomov-st-webagentbench/"}
{"id": "github:servicenow/agentlab", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ServiceNow/AgentLab", "citation_url": "https://awesomecua.github.io/resources/github-servicenow-agentlab/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "ca312434db1db852f96fb49bbc0a61f6a93cffe6c0dd3aa2b7ac814681ccc19e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/AgentLab"}], "name": "AgentLab", "publication": null, "published_at": null, "record_sha256": "be9543a4e292f42dcfb362f4648b812b905c837b64690ec16a32f010e2d56196", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/AgentLab/HEAD/README.md", "target_id": "arxiv:2412.05467", "type": "documented_by"}], "resource_id": "github:servicenow/agentlab", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ServiceNow/AgentLab", "topics": ["evaluation"], "updated_at": null}, "text": "AgentLab\nResource type: framework\nReview status: curated\nScope: core\n\nExperiment framework for developing and evaluating browser agents.\nPlatforms: web\nTopics: evaluation\n\nPrimary source: https://github.com/ServiceNow/AgentLab\nCatalog record: https://awesomecua.github.io/resources/github-servicenow-agentlab/"}
{"id": "github:servicenow/browsergym", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ServiceNow/BrowserGym", "citation_url": "https://awesomecua.github.io/resources/github-servicenow-browsergym/", "details": {"benchmark_protocol": {"actions": ["element-ID actions", "coordinate mouse and keyboard actions", "tab/navigation/scroll operations", "Python/Playwright action mode"], "counts": [], "evaluation": "Gymnasium reset/step interface delegates reward, completion, messages and task info to each adapter’s validate() method. There is no universal BrowserGym success metric or shared task total; retain benchmark/adapter version and action subset for comparisons.", "evidence_urls": ["https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md", "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py", "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst", "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"], "focus": ["shared browser-agent evaluation interface", "benchmark adapters with separate task definitions"], "notes": "A framework with interactive browser adapters and a static WebLINX evaluation integration, so the profile setting is mixed. README-listed integrations are not added as disjoint tasks or datasets. Core requires Playwright/Chromium; individual benchmarks have their own setup, credentials, licenses and validation rules. Apache-2.0 covers repository code only. No browser tasks or third-party project code were executed.", "observations": ["browser screenshot", "DOM snapshot", "accessibility tree", "tab/URL state", "previous action and its error", "chat messages and task goal"], "setting": "mixed"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "7b685115ef822989fac9093518d16c46f0202736d005eed56e7f9fc36b43f3ee", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "867cc6efdfead356839fef8a26b492502b8b5e3d165d1b3c17d24aaeec268d63", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "c4bbe6b7d0cb36a5d5effab1aa5568301cead97fe4015273e1b82c667da3823a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:08Z", "role": "primary", "sha256": "418efbe8c65ad7ebab61d3a066339876c820a9c957506eb456b8078ab352f6bf", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/BrowserGym"}], "name": "BrowserGym", "publication": null, "published_at": null, "record_sha256": "9de4a0fc9cefa4f43c23598b7aeb5cee98c0ff38c996660e10094068dc592bc4", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md", "target_id": "arxiv:2412.05467", "type": "documented_by"}], "resource_id": "github:servicenow/browsergym", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ServiceNow/BrowserGym", "topics": ["evaluation"], "updated_at": null}, "text": "BrowserGym\nResource type: benchmark\nReview status: curated\nScope: core\n\nUnified interfaces for browser-agent tasks and benchmark environments.\nPlatforms: web\nTopics: evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"element-ID actions\", \"coordinate mouse and keyboard actions\", \"tab/navigation/scroll operations\", \"Python/Playwright action mode\"], \"counts\": [], \"evaluation\": \"Gymnasium reset/step interface delegates reward, completion, messages and task info to each adapter’s validate() method. There is no universal BrowserGym success metric or shared task total; retain benchmark/adapter version and action subset for comparisons.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/README.md\", \"https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/browsergym/core/src/browsergym/core/env.py\", \"https://raw.githubusercontent.com/ServiceNow/BrowserGym/HEAD/docs/src/core/action_space.rst\", \"https://api.github.com/repos/ServiceNow/BrowserGym/git/trees/main?recursive=1\"], \"focus\": [\"shared browser-agent evaluation interface\", \"benchmark adapters with separate task definitions\"], \"notes\": \"A framework with interactive browser adapters and a static WebLINX evaluation integration, so the profile setting is mixed. README-listed integrations are not added as disjoint tasks or datasets. Core requires Playwright/Chromium; individual benchmarks have their own setup, credentials, licenses and validation rules. Apache-2.0 covers repository code only. No browser tasks or third-party project code were executed.\", \"observations\": [\"browser screenshot\", \"DOM snapshot\", \"accessibility tree\", \"tab/URL state\", \"previous action and its error\", \"chat messages and task goal\"], \"setting\": \"mixed\"}}\n\nPrimary source: https://github.com/ServiceNow/BrowserGym\nCatalog record: https://awesomecua.github.io/resources/github-servicenow-browsergym/"}
{"id": "github:servicenow/workarena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ServiceNow/WorkArena", "citation_url": "https://awesomecua.github.io/resources/github-servicenow-workarena/", "details": {"benchmark_protocol": {"actions": ["BrowserGym browser actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena-L1 atomic tasks", "qualifier": "exact", "unit": "task definitions", "value": 33}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena-L1 documented instances", "qualifier": "exact", "unit": "task instances", "value": 19912}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "population": "WorkArena++ compositional tasks", "qualifier": "exact", "unit": "task definitions", "value": 682}], "evaluation": "Task-specific checks on ServiceNow workflows; levels and task configurations define different evaluations.", "evidence_urls": ["https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "focus": ["enterprise web workflows", "compositional tasks"], "notes": "Definitions and instantiated tasks are different units. ServiceNow instance access is gated. WorkArena-L1 and WorkArena++ should be reported separately.", "observations": ["multimodal browser observations"], "setting": "interactive"}, "leaderboard_url": "https://huggingface.co/spaces/ServiceNow/browsergym-leaderboard"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "834a48f0346036a0fc591a36f9eb6f513b44a2777a66afa0917f28fc63f29334", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "b192c58991e8ff585cc574615d40e74185404d4b96c1109d423071ab1367344b", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ServiceNow/WorkArena"}], "name": "WorkArena", "publication": null, "published_at": null, "record_sha256": "887d56f58f86a84754172dff5651813647e7ba3a5faa14675726bcbf10f11361", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "target_id": "arxiv:2403.07718", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md", "target_id": "arxiv:2407.05291", "type": "documented_by"}], "resource_id": "github:servicenow/workarena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ServiceNow/WorkArena", "topics": ["evaluation", "workflow"], "updated_at": null}, "text": "WorkArena\nResource type: benchmark\nReview status: curated\nScope: core\n\nBrowser-based enterprise workflow tasks built around ServiceNow.\nPlatforms: web\nTopics: evaluation, workflow\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"BrowserGym browser actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md\", \"population\": \"WorkArena-L1 atomic tasks\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 33}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md\", \"population\": \"WorkArena-L1 documented instances\", \"qualifier\": \"exact\", \"unit\": \"task instances\", \"value\": 19912}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md\", \"population\": \"WorkArena++ compositional tasks\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 682}], \"evaluation\": \"Task-specific checks on ServiceNow workflows; levels and task configurations define different evaluations.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/ServiceNow/WorkArena/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"focus\": [\"enterprise web workflows\", \"compositional tasks\"], \"notes\": \"Definitions and instantiated tasks are different units. ServiceNow instance access is gated. WorkArena-L1 and WorkArena++ should be reported separately.\", \"observations\": [\"multimodal browser observations\"], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://huggingface.co/spaces/ServiceNow/browsergym-leaderboard\"}\n\nPrimary source: https://github.com/ServiceNow/WorkArena\nCatalog record: https://awesomecua.github.io/resources/github-servicenow-workarena/"}
{"id": "github:showlab/assistgui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/assistgui", "citation_url": "https://awesomecua.github.io/resources/github-showlab-assistgui/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9c43f97560c3781b28f97c894ff65bbea7121c3df8c183ddf15bb38842bd1e39", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/assistgui/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/assistgui"}], "name": "AssistGUI", "publication": null, "published_at": null, "record_sha256": "b58ba2e2f605bace0d5eef9cdb8cecfdadcab19beca093b6f078b56d75433a72", "relations": [], "resource_id": "github:showlab/assistgui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/assistgui", "topics": ["planning"], "updated_at": null}, "text": "AssistGUI\nResource type: agent\nReview status: curated\nScope: core\n\nLLM-based assistant for automating work in desktop productivity software by watching the screen and issuing mouse and keyboard commands; its repository says the code will be released later.\nPlatforms: desktop\nTopics: planning\n\nPrimary source: https://github.com/showlab/assistgui\nCatalog record: https://awesomecua.github.io/resources/github-showlab-assistgui/"}
{"id": "github:showlab/computer_use_ootb", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/computer_use_ootb", "citation_url": "https://awesomecua.github.io/resources/github-showlab-computer-use-ootb/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "7b53e58d3efb35d23e58bd75bfa6d36f1e57af9f63556c34935726cead47e322", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "cebb89bae373fa4f9a210ed5b664fbbfcbdb7ac8b5d8c34bfa7e8905d1f63b4e", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/showlab/computer_use_ootb/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/computer_use_ootb"}], "name": "Computer Use OOTB", "publication": null, "published_at": null, "record_sha256": "c43df342213d74edaaa06a3e4f9488ff963efc4205a43545d2aa8d2c2fca2afd", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2411.10323", "type": "documented_by"}], "resource_id": "github:showlab/computer_use_ootb", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/computer_use_ootb", "topics": ["workflow"], "updated_at": null}, "text": "Computer Use OOTB\nResource type: framework\nReview status: curated\nScope: core\n\nGradio desktop GUI agent for Windows and macOS that needs no Docker, using Claude 3.5 Computer Use over the API or ShowUI and UI-TARS run locally, with multi-display and phone remote control.\nPlatforms: desktop\nTopics: workflow\n\nPrimary source: https://github.com/showlab/computer_use_ootb\nCatalog record: https://awesomecua.github.io/resources/github-showlab-computer-use-ootb/"}
{"id": "github:showlab/gui-action-narrator", "metadata": {"aliases": ["GUI-Narrator"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/GUI-Action-Narrator", "citation_url": "https://awesomecua.github.io/resources/github-showlab-gui-action-narrator/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "Act2Cap benchmark described in the paper", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2406.13719", "unit": "video-captioning samples", "value": 4189}], "evidence_urls": ["https://github.com/showlab/GUI-Action-Narrator", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "formats": [], "modalities": ["GUI video", "action captions"], "notes": "A captioning sample is not necessarily a unique video or complete task trajectory. Package layout and split sizes were not established by the inspected sources.", "splits": [], "supervision": ["GUI action captioning"]}}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:52Z", "role": "primary", "sha256": "379aae1a8890a0c3d3b5999589b8338338cf1ea63883c2b77df678fd3bb367a2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://github.com/showlab/GUI-Action-Narrator"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.dataset_profile"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:40:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/GUI-Action-Narrator"}], "name": "GUI-Action Narrator", "publication": null, "published_at": null, "record_sha256": "86dbb8c589bf102f6667049f693ce77f53943c458f38d8ba67ce84d3634200e3", "relations": [], "resource_id": "github:showlab/gui-action-narrator", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/GUI-Action-Narrator", "topics": ["data-generation", "video"], "updated_at": null}, "text": "GUI-Action Narrator\nResource type: dataset\nReview status: curated\nScope: core\n\nAction narration resources for GUI interaction trajectories.\nRecorded aliases: GUI-Narrator\nPlatforms: cross-platform\nTopics: data-generation, video\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\", \"population\": \"Act2Cap benchmark described in the paper\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2406.13719\", \"unit\": \"video-captioning samples\", \"value\": 4189}], \"evidence_urls\": [\"https://github.com/showlab/GUI-Action-Narrator\", \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\"], \"formats\": [], \"modalities\": [\"GUI video\", \"action captions\"], \"notes\": \"A captioning sample is not necessarily a unique video or complete task trajectory. Package layout and split sizes were not established by the inspected sources.\", \"splits\": [], \"supervision\": [\"GUI action captioning\"]}}\n\nOriginal repository URL redirects to showlab/GUI-Narrator.\n\nPrimary source: https://github.com/showlab/GUI-Action-Narrator\nCatalog record: https://awesomecua.github.io/resources/github-showlab-gui-action-narrator/"}
{"id": "github:showlab/showui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/ShowUI", "citation_url": "https://awesomecua.github.io/resources/github-showlab-showui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "730e51422c0ebe00cb2a1a8f4510d9914b1ea735fdb06445fde2b243da80bdd4", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "1b4977f976211c2644742e3a56cf75ecffb4c7c31d1bce456875123ba8bcd342", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/ShowUI"}], "name": "ShowUI", "publication": null, "published_at": null, "record_sha256": "24c2815821da5371162d43a13e1a1ce5b2d2711135c3c86a40e450320064c3de", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/ShowUI/HEAD/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "resource_id": "github:showlab/showui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/ShowUI", "topics": ["grounding", "efficiency"], "updated_at": null}, "text": "ShowUI\nResource type: model\nReview status: curated\nScope: core\n\nLightweight GUI-agent vision-language-action model from Show Lab and Microsoft, released with UI-guided token selection, training code, the ShowUI-web and ShowUI-desktop datasets and vLLM inference.\nPlatforms: cross-platform\nTopics: grounding, efficiency\n\nPrimary source: https://github.com/showlab/ShowUI\nCatalog record: https://awesomecua.github.io/resources/github-showlab-showui/"}
{"id": "github:showlab/showui-aloha", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/ShowUI-Aloha", "citation_url": "https://awesomecua.github.io/resources/github-showlab-showui-aloha/", "details": {}, "environments": ["desktop", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "ad5553270b67c335df4d4e909f1549f43af8e74ad687f9cbcc812761affcad5c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T11:15:55Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T11:15:55Z", "url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/ShowUI-Aloha"}], "name": "ShowUI-Aloha", "publication": null, "published_at": null, "record_sha256": "90b62263731df8cf52c61b2564bc52b45e59db75be8646e9d4f610e0b852585b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/ShowUI-Aloha/HEAD/README.md", "target_id": "arxiv:2601.07181", "type": "documented_by"}], "resource_id": "github:showlab/showui-aloha", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/ShowUI-Aloha", "topics": ["recording", "workflow", "automation"], "updated_at": null}, "text": "ShowUI-Aloha\nResource type: agent\nReview status: curated\nScope: core\n\nDesktop agent that records demonstrations, derives semantic action traces, and uses them to plan and execute related workflows.\nPlatforms: desktop, windows, macos\nTopics: recording, workflow, automation\n\nPrimary source: https://github.com/showlab/ShowUI-Aloha\nCatalog record: https://awesomecua.github.io/resources/github-showlab-showui-aloha/"}
{"id": "github:showlab/showui-pi", "metadata": {"aliases": ["ShowUI-pi", "ShowUI Pi"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/showui-pi", "citation_url": "https://awesomecua.github.io/resources/github-showlab-showui-pi/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/showui-pi"}], "name": "ShowUI-π", "publication": null, "published_at": null, "record_sha256": "9ae5cf36d29e02b1419464b09fc84faa25d69a359a6643f76af662fdd2cd3c9c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "resource_id": "github:showlab/showui-pi", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/showui-pi", "topics": ["grounding", "automation"], "updated_at": null}, "text": "ShowUI-π\nResource type: model\nReview status: curated\nScope: core\n\nVision-language-action model and training code for GUI dragging, with links to checkpoints and continuous-action evaluation data.\nRecorded aliases: ShowUI-pi, ShowUI Pi\nPlatforms: desktop\nTopics: grounding, automation\n\nPrimary source: https://github.com/showlab/showui-pi\nCatalog record: https://awesomecua.github.io/resources/github-showlab-showui-pi/"}
{"id": "github:showlab/videogui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/videogui", "citation_url": "https://awesomecua.github.io/resources/github-showlab-videogui/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "1a21631fb312cc7eac35547792611a84f13dcf756a3cf92209676ccdb92cbc6b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/videogui/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/videogui"}], "name": "VideoGUI", "publication": null, "published_at": null, "record_sha256": "d23cfbc46582dd0ed31303f6d248c091f7a375e2fb324179ba0f5459bbb0c649", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/videogui/HEAD/README.md", "target_id": "arxiv:2406.10227", "type": "documented_by"}], "resource_id": "github:showlab/videogui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/videogui", "topics": ["evaluation", "video"], "updated_at": null}, "text": "VideoGUI\nResource type: benchmark\nReview status: curated\nScope: core\n\nBenchmark for desktop GUI tasks derived from instructional videos.\nPlatforms: desktop\nTopics: evaluation, video\n\nPrimary source: https://github.com/showlab/videogui\nCatalog record: https://awesomecua.github.io/resources/github-showlab-videogui/"}
{"id": "github:showlab/worldgui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/showlab/WorldGUI", "citation_url": "https://awesomecua.github.io/resources/github-showlab-worldgui/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "ea54ea82b86548ff38913ef328ec18152ace71a3fe9bb497bcabcaad24e6a678", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/showlab/WorldGUI/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/showlab/WorldGUI"}], "name": "WorldGUI", "publication": null, "published_at": null, "record_sha256": "f7a7cf152cd17bc72a80aab5e1cd67d77caf5600aa582d23c6cbf040d689a15f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/WorldGUI/HEAD/README.md", "target_id": "arxiv:2502.08047", "type": "documented_by"}], "resource_id": "github:showlab/worldgui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/showlab/WorldGUI", "topics": ["evaluation", "robustness"], "updated_at": null}, "text": "WorldGUI\nResource type: benchmark\nReview status: curated\nScope: core\n\n611 desktop GUI tasks that agents must complete from varied starting points, shipped with instructional videos and project files, plus the self-reflective WorldGUI-Agent baseline.\nPlatforms: desktop\nTopics: evaluation, robustness\n\nPrimary source: https://github.com/showlab/WorldGUI\nCatalog record: https://awesomecua.github.io/resources/github-showlab-worldgui/"}
{"id": "github:shulin16/mmina", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/shulin16/MMInA", "citation_url": "https://awesomecua.github.io/resources/github-shulin16-mmina/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "abc848e7eb5181be22168dbc076446c5eb0f065fd57c7e10367284ca6e50bf5c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/shulin16/MMInA/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/shulin16/MMInA"}], "name": "MMInA", "publication": null, "published_at": null, "record_sha256": "51fd755c9a3be171d88c15e8be4460e2a753559926f496ac717a0c66baee48d5", "relations": [{"evidence_url": "https://raw.githubusercontent.com/shulin16/MMInA/HEAD/README.md", "target_id": "arxiv:2404.09992", "type": "documented_by"}], "resource_id": "github:shulin16/mmina", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/shulin16/MMInA", "topics": ["evaluation", "long-horizon"], "updated_at": null}, "text": "MMInA\nResource type: benchmark\nReview status: curated\nScope: core\n\nMultimodal web tasks in a WebArena-based environment that chain from 1 to 10 hops across sites such as a shopping store and Wikipedia, testing long-chain reasoning.\nPlatforms: web\nTopics: evaluation, long-horizon\n\nPrimary source: https://github.com/shulin16/MMInA\nCatalog record: https://awesomecua.github.io/resources/github-shulin16-mmina/"}
{"id": "github:sikuli/sikuli", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/sikuli/sikuli", "citation_url": "https://awesomecua.github.io/resources/github-sikuli-sikuli/", "details": {"availability": "Public legacy source and original documentation; maintainers explicitly state it is no longer supported.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"], "reason": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion."}}, "maintenance_status": "unsupported legacy research code", "repository_revision": "4adaab7880d2f3e14702ca7287ae9c9e4f4de9ab"}, "environments": ["desktop", "macos", "linux", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:12Z", "role": "primary", "sha256": "e752b2226af679a778e21506e660742376be3d7effe621f63c0b8816d3323729", "supports": ["name", "kind", "summary", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:12Z", "role": "primary", "sha256": "90d840ab98d35b90c0f91f93dc2dc752eac631a4bdc299601bffe8856610dd4c", "supports": ["identity"], "url": "https://api.github.com/repos/sikuli/sikuli"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:49Z", "role": "primary", "sha256": "8ee5f1e75659bfbce0d9b22967ff2b51335be5f19a3dc9e645273cc51333153e", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/sikuli/sikuli/git/trees/develop?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:13Z", "role": "primary", "sha256": "ed7659fff11dac079b99ad3c264f7e1e0a97f313a75379067e1c7fccca23ef88", "supports": ["notes", "details.availability", "relations"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "bab2f2dae13092e9c94f5d4b077bc96a6aa7da5257b54089f76ff5c0b0f61402", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/tutorials/helloworld/helloworld-mac.rst"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "28354a06091997d233cc5124958d85b75a5d815677c58b4fbabd492a6b0038d3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-linux.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:28:13Z", "role": "primary", "sha256": "b7c66064d08c6a768a4ed93fee0f86f837b41d51e60785d44a59d7a13bd2452a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/build-win32.txt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:03Z", "role": "primary", "sha256": "717e5c54ef8201b2307c9828a6bacb7710385e5014adba93e5f1dc444339fdab", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T11:28:12Z", "licenses": {"code": {"basis": "Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion.", "declared": "MIT", "retrieved_at": "2026-10-17T11:23:03Z", "url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sikuli/sikuli"}], "name": "Sikuli legacy research implementation", "publication": null, "published_at": null, "record_sha256": "b07868e9e043fbd8c9803efbfce4d486951fa0018dcc75cfbdb461f7bf123402", "relations": [{"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:a4cdb2ee0d07d3980651", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/sikuli/sikuli/HEAD/docs/source/index.rst", "target_id": "url:6262c757080f8838e62c", "type": "documented_by"}], "resource_id": "github:sikuli/sikuli", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/sikuli/sikuli", "topics": ["automation", "grounding", "screen-understanding", "history"], "updated_at": null}, "text": "Sikuli legacy research implementation\nResource type: tool\nReview status: curated\nScope: core\n\nThe original Sikuli source distribution combines screenshot targets with a Jython scripting library and IDE for visual GUI automation and testing.\nPlatforms: desktop, macos, linux, windows\nTopics: automation, grounding, screen-understanding, history\n\nRecorded specifications: {\"availability\": \"Public legacy source and original documentation; maintainers explicitly state it is no longer supported.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/sikuli/sikuli/HEAD/copyright.txt\"], \"reason\": \"Original Sikuli copyright/license file for project source; bundled third-party libraries are not relicensed by this assertion.\"}}, \"maintenance_status\": \"unsupported legacy research code\", \"repository_revision\": \"4adaab7880d2f3e14702ca7287ae9c9e4f4de9ab\"}\n\nThis is the retained Sikuli X-1.0rc3 source line, explicitly marked unsupported by its maintainers. Its documented build targets include historical macOS, Linux and Windows setups; present-day build compatibility was not tested. The official documentation cites the 2009 search/automation and 2010 visual-testing papers. Follow-up SikuliX and current OculiX are distinguished from this legacy research implementation. MIT applies to project code; external libraries keep their own terms.\n\nPrimary source: https://github.com/sikuli/sikuli\nCatalog record: https://awesomecua.github.io/resources/github-sikuli-sikuli/"}
{"id": "github:simular-ai/agent-s", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/simular-ai/Agent-S", "citation_url": "https://awesomecua.github.io/resources/github-simular-ai-agent-s/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "9ea381fdb78eb326d988ea384b6f5074ef0ff4c0d5fab555a148eded385e88f1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "43070e2d4e532684de521b885f385d0841030efa2b1a20bafb76133a5e1379c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/simular-ai/Agent-S"}], "name": "Agent S", "publication": null, "published_at": null, "record_sha256": "752610da986b4ec6f8a521538b40a0afbce2bcd5227bb604b087a3c1f9fd7d95", "relations": [{"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "arxiv:2410.08164", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "arxiv:2504.00906", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/simular-ai/Agent-S/HEAD/README.md", "target_id": "organization:simular", "type": "published_by"}], "resource_id": "github:simular-ai/agent-s", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/simular-ai/Agent-S", "topics": ["planning", "memory"], "updated_at": null}, "text": "Agent S\nResource type: agent\nReview status: curated\nScope: core\n\nSimular's open-source framework for computer-use agents that click, type and scroll through desktop and web apps on macOS, Windows and Linux; its third generation, Agent S3, reports 72.6% on OSWorld.\nPlatforms: desktop\nTopics: planning, memory\n\nPrimary source: https://github.com/simular-ai/Agent-S\nCatalog record: https://awesomecua.github.io/resources/github-simular-ai-agent-s/"}
{"id": "github:skyvern-ai/skyvern", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Skyvern-AI/skyvern", "citation_url": "https://awesomecua.github.io/resources/github-skyvern-ai-skyvern/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "68fda945255825907aeada13d19de19209708c14504ac975e63a235f0cd9343b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Skyvern-AI/skyvern/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Skyvern-AI/skyvern"}], "name": "Skyvern", "publication": null, "published_at": null, "record_sha256": "64db7335548f55fc6dc095fb358a62a9ed15df62987eeab5683949284a7b7eb1", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Skyvern-AI/skyvern/HEAD/README.md", "target_id": "organization:skyvern", "type": "published_by"}], "resource_id": "github:skyvern-ai/skyvern", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Skyvern-AI/skyvern", "topics": ["workflow"], "updated_at": null}, "text": "Skyvern\nResource type: framework\nReview status: curated\nScope: core\n\nBrowser workflow automation that relies on vision LLMs and a swarm of agents instead of fixed XPath selectors, offered as a Playwright-compatible SDK and a no-code workflow builder.\nPlatforms: web\nTopics: workflow\n\nPrimary source: https://github.com/Skyvern-AI/skyvern\nCatalog record: https://awesomecua.github.io/resources/github-skyvern-ai-skyvern/"}
{"id": "github:skyworkai/agent-studio", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/skyworkai/agent-studio", "citation_url": "https://awesomecua.github.io/resources/github-skyworkai-agent-studio/", "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:56Z", "role": "primary", "sha256": "5653db27f736c0c77afb63f81ca051bf31f1e1f0579ac77cf00dca4ea751c159", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:47Z", "role": "primary", "sha256": "8486a10c4393cee1c25392769ddd3b2d6c242d6ec7928e1414efff7dfb2f07ef", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:56Z", "licenses": {"code": {"basis": "repository LICENSE header; GNU text does not say -only or -or-later; not a weights/data license", "declared": "AGPL-3.0", "retrieved_at": "2026-10-18T11:03:47Z", "url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/skyworkai/agent-studio"}], "name": "AgentStudio", "publication": null, "published_at": null, "record_sha256": "bd486171e14da42fe9745d2491cab572531b996ebb1372f39198291500d52668", "relations": [{"evidence_url": "https://raw.githubusercontent.com/skyworkai/agent-studio/HEAD/README.md", "target_id": "arxiv:2403.17918", "type": "documented_by"}], "resource_id": "github:skyworkai/agent-studio", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/skyworkai/agent-studio", "topics": ["sandbox", "evaluation", "grounding", "recording"], "updated_at": null}, "text": "AgentStudio\nResource type: framework\nReview status: curated\nScope: core\n\nToolkit and Ubuntu-based environment for building and evaluating generalist computer-use agents through GUI and API interaction.\nPlatforms: desktop, linux\nTopics: sandbox, evaluation, grounding, recording\n\nPrimary source: https://github.com/skyworkai/agent-studio\nCatalog record: https://awesomecua.github.io/resources/github-skyworkai-agent-studio/"}
{"id": "github:solplum/yunkai", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/solplum/yunkai", "citation_url": "https://awesomecua.github.io/resources/github-solplum-yunkai/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"], "reason": "Complete root MIT license text inspected at the pinned commit. Repository software only; data, weights and third-party assets retain separate terms."}}, "repository_revision": "2aa6cd83165bfc201b0128c250e80892895d63af"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:12.431753+00:00", "role": "primary", "sha256": "eea31b35aa0b55c97f8ccda3438382ad49d2453671b49bb817efb5bc3384e900", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T11:06:00.276978+00:00", "role": "primary", "sha256": "01f97a863952373bd10a1f849d79706344be63f5774fe7e0cfe9ba774e41d213", "supports": ["identity"], "url": "https://api.github.com/search/repositories?q=topic%3Amobile-agent+is%3Apublic+fork%3Afalse+created%3A2008-01-01..2026-10-20&per_page=100&page=1&sort=updated&order=asc"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:54:02.419278+00:00", "role": "primary", "sha256": "e2797712132e69cc50fa7ad2ae815a48339f77eebec1e85a5dc0cd75702574f1", "supports": ["identity", "details.repository_revision"], "url": "https://api.github.com/repos/solplum/yunkai/commits/main"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:56:30.226860+00:00", "role": "primary", "sha256": "08f48cb5c15b8d505e76a91d996c656d87765946cc7137bce088e73e6e64cf06", "supports": ["identity", "details.repository_revision", "notes"], "url": "https://api.github.com/repos/solplum/yunkai/git/trees/4ca94f2f7d8f13099ae2c351d9578ed5ef0cb509?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:57:22.238029+00:00", "role": "primary", "sha256": "c39441f92c303832d42f1f153c4442533ac925007e12d70f3199c568a0baac88", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:59:51.424713+00:00", "role": "primary", "sha256": "964aad706676fc7b74e532f8bc441fa766f3ce0ca3e4d0e115c7a90c70656883", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-android/app/src/main/java/com/zhuolin/yunkai/service/screen/ScreenSenseService.kt"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T11:59:54.745065+00:00", "role": "primary", "sha256": "628cc3015999176052a09e3f3bbb8635d87163a9430d98cb6a6a7c09117dfd37", "supports": ["notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/entry/src/main/ets/service/tools/M2Tools.ets"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:04:53.632582+00:00", "role": "primary", "sha256": "ba8c9b0f4b5cea5ac36a4b49f3cf27e77e5c6beb49022e43c667e359be2c6fca", "supports": ["environments", "notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/entry/src/main/ets/service/screen/ScreenSenseExtension.ets"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:04:56.914360+00:00", "role": "primary", "sha256": "423a77e1fd57f3e105898f1adfc94c493de5c0de97fae4379f9cf540c89d4640", "supports": ["notes"], "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/yunkai-harmony/tests/fullflow/drive_harmony.sh"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-20T11:57:12.431753+00:00", "licenses": {"code": {"basis": "Complete root MIT license text inspected at the pinned commit. Repository software only; data, weights and third-party assets retain separate terms.", "declared": "MIT", "retrieved_at": "2026-10-20T11:57:22.238029+00:00", "url": "https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/solplum/yunkai"}], "name": "Yunkai", "publication": null, "published_at": null, "record_sha256": "6625f8af32a67872f8f7677d7e81894e3b9357a90fea0b5e577967fe161b7d9e", "relations": [], "resource_id": "github:solplum/yunkai", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/solplum/yunkai", "topics": ["mobile-navigation", "screen-understanding", "human-in-the-loop"], "updated_at": null}, "text": "Yunkai\nResource type: agent\nReview status: curated\nScope: core\n\nAndroid phone assistant combining model-driven tool use with accessibility-tree reading, screenshots, and tap, swipe and text-entry primitives. Its documented interaction flow presents operation plans for user confirmation.\nPlatforms: android\nTopics: mobile-navigation, screen-understanding, human-in-the-loop\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/solplum/yunkai/2aa6cd83165bfc201b0128c250e80892895d63af/LICENSE\"], \"reason\": \"Complete root MIT license text inspected at the pinned commit. Repository software only; data, weights and third-party assets retain separate terms.\"}}, \"repository_revision\": \"2aa6cd83165bfc201b0128c250e80892895d63af\"}\n\nThe Android AccessibilityService contains concrete screen-reading and action paths. The inspected HarmonyOS accessibility extension is a lifecycle/event scaffold with its screen-content probe disabled; HDC appears in an external UI-test driver, so Harmony GUI-control parity is not established by this review. Setup requires a configured OpenAI-compatible model endpoint and API key: local storage does not establish offline inference or a no-transmission guarantee. Documented approval and privacy mechanisms are not independently tested security guarantees. No app, device test or model inference was run.\n\nPrimary source: https://github.com/solplum/yunkai\nCatalog record: https://awesomecua.github.io/resources/github-solplum-yunkai/"}
{"id": "github:squeezeailab/plan-and-act", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/SqueezeAILab/plan-and-act", "citation_url": "https://awesomecua.github.io/resources/github-squeezeailab-plan-and-act/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "feee31aa4d0801cb364578099a13e4f9cdcaa4eb5fbf4312b3fc3f8edd479a14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "eada6ee9aa180f55daca059c5e7a4922b320cccb0514864108eba9040b52415c", "supports": ["identity"], "url": "https://api.github.com/repos/SqueezeAILab/plan-and-act"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "10ff0a1c1abe25454541949fdaac1869d68b5d4426fbb5ef1d487657afcd26f5", "supports": ["release_status"], "url": "https://api.github.com/repos/SqueezeAILab/plan-and-act/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "f86e78202a51ab013e1cc47ba37386134601317f5a0c4e4ed287981b52159122", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:02:15Z", "role": "primary", "sha256": "f86e78202a51ab013e1cc47ba37386134601317f5a0c4e4ed287981b52159122", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:02:15Z", "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/SqueezeAILab/plan-and-act"}], "name": "Plan-and-Act", "publication": null, "published_at": null, "record_sha256": "f55becd69165ead5cfaaef16785a72c6de5afe2b791389c06807ea52888e7067", "relations": [{"evidence_url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "resource_id": "github:squeezeailab/plan-and-act", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/SqueezeAILab/plan-and-act", "topics": ["planning", "web-navigation", "data-generation"], "updated_at": null}, "text": "Plan-and-Act\nResource type: framework\nReview status: curated\nScope: core\n\nA web-agent framework that separates high-level planning from action execution, with replanning and synthetic planner-data generation components.\nPlatforms: web\nTopics: planning, web-navigation, data-generation\n\nRunnable entry points, framework modules and synthetic-data tooling are published. The repository code declares MIT; separate model and dataset cards have their own declarations. Its setup targets WebArena-Lite and requires that external environment.\n\nPrimary source: https://github.com/SqueezeAILab/plan-and-act\nCatalog record: https://awesomecua.github.io/resources/github-squeezeailab-plan-and-act/"}
{"id": "github:sqzhang-lazy/d-pot", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/sqzhang-lazy/D-PoT", "citation_url": "https://awesomecua.github.io/resources/github-sqzhang-lazy-d-pot/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sqzhang-lazy/D-PoT/HEAD/README.md"], "reason": "No project-wide code license is established by the inspected README/API/tree; third-party code presence does not establish a license for this package."}}, "repository_tree_sha": "fc764f19802f4f59afa635bfeec5a50e4c3d5465", "repository_url": "https://github.com/sqzhang-lazy/D-PoT"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "c9c21102db512270079666531dbab71c1ab2ec2073add16be56f8ab3c6fe8abf", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.code"], "url": "https://raw.githubusercontent.com/sqzhang-lazy/D-PoT/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "d62cfecbfccd120e53a7fc60e8d4b90b157e998cc5ddac467514c30c0cae1363", "supports": ["identity"], "url": "https://api.github.com/repos/sqzhang-lazy/D-PoT"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:24Z", "role": "primary", "sha256": "f75444a0252df5d6ec1ab70d805584cb490cdc38efa995aae55ffbde43ab5fb1", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/sqzhang-lazy/D-PoT/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:35Z", "role": "primary", "sha256": "b3e67984f59a1b05ccd946eb54e2d37b0bb78ab0ca796fbb6d544793dafea3e0", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sqzhang-lazy/D-PoT"}], "name": "D-PoT evaluation code", "publication": null, "published_at": null, "record_sha256": "26fb88589661c4728b852587959ca29c86839468a0dc0a84f07897e3442be4c4", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2410.00467%2C2410.00689%2C2410.02810%2C2410.02907%2C2410.05243%2C2410.06703%2C2410.08164%2C2410.09006%2C2410.09024%2C2410.11843%2C2410.11871%2C2410.11872%2C2410.13232%2C2410.13757%2C2410.13824%2C2410.13825%2C2410.13886%2C2410.14803%2C2410.15164%2C2410.16464%2C2410.17236%2C2410.17401%2C2410.17520%2C2410.17883%2C2410.18603%2C2410.18963%2C2410.18967%2C2410.19054%2C2410.19100%2C2410.19461%2C2410.19609%2C2410.20745%2C2410.22552%2C2410.23218%2C2410.24024%2C2411.00820%2C2411.02337%2C2411.02391%2C2411.04468%2C2411.04890%2C2411.06559%2C2411.07618%2C2411.10323%2C2411.13451%2C2411.13591%2C2411.15004%2C2411.17465%2C2411.18279%2C2412.00402%2C2412.01268&max_results=50", "target_id": "arxiv:2410.00467", "type": "documented_by"}], "resource_id": "github:sqzhang-lazy/d-pot", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/sqzhang-lazy/D-PoT", "topics": ["planning", "action-prediction", "evaluation"], "updated_at": null}, "text": "D-PoT evaluation code\nResource type: framework\nReview status: curated\nScope: core\n\nOffline prompting and evaluation code for Dynamic Planning of Thoughts on Android-in-the-Wild, including action extraction and comparison with reference actions.\nPlatforms: android\nTopics: planning, action-prediction, evaluation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/sqzhang-lazy/D-PoT/HEAD/README.md\"], \"reason\": \"No project-wide code license is established by the inspected README/API/tree; third-party code presence does not establish a license for this package.\"}}, \"repository_tree_sha\": \"fc764f19802f4f59afa635bfeec5a50e4c3d5465\", \"repository_url\": \"https://github.com/sqzhang-lazy/D-PoT\"}\n\nREADME describes AITW test sampling and preprocessing reused from Auto-GUI, followed by prediction/action-extraction/evaluation scripts. This is not claimed as a ready-to-run interactive phone controller. No project-level license declaration appears in the inspected README/API/non-truncated tree; code, dataset and model rights remain unestablished here.\n\nPrimary source: https://github.com/sqzhang-lazy/D-PoT\nCatalog record: https://awesomecua.github.io/resources/github-sqzhang-lazy-d-pot/"}
{"id": "github:stanfordnlp/wge", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/stanfordnlp/wge", "citation_url": "https://awesomecua.github.io/resources/github-stanfordnlp-wge/", "details": {"availability": "Public historical training and environment-wrapper source; legacy dependencies documented.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"], "reason": "Stanford WGE source license; not a blanket license for all third-party tasks, embeddings or demonstration resources."}}, "repository_revision": "bed526657a62a64a163f80deb4a8627cb63871a3"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:16Z", "role": "primary", "sha256": "50ee5234bf62320d096fbe4c11c7be21f50c9e09dae97754a16a5bbde2f1ab19", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:15Z", "role": "primary", "sha256": "66fcd6577d1867abf84e62fe9edf2e3f223292724cbe6b9d0ef02b582db5c8b2", "supports": ["identity"], "url": "https://api.github.com/repos/stanfordnlp/wge"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:19:50Z", "role": "primary", "sha256": "926aff11e14c529d05d9bc51fe12aadb007818119da55f99c2fa47d0387f466d", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/stanfordnlp/wge/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:20:17Z", "role": "primary", "sha256": "249b23ff62e21bd446b3e411b1d7ea7f173c7e19e42ed75e97800ed32201bb3a", "supports": ["details", "notes"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/third-party/miniwob-sandbox/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:05Z", "role": "primary", "sha256": "3df3dd1e3149cd6cc1a7b052a237a6653449a7c5f18edf19a3a79294c7d541c1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:16Z", "licenses": {"code": {"basis": "Stanford WGE source license; not a blanket license for all third-party tasks, embeddings or demonstration resources.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T11:23:05Z", "url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/stanfordnlp/wge"}], "name": "Workflow-Guided Exploration research code", "publication": null, "published_at": null, "record_sha256": "99ba0d3ee58e256d5ebe2641884f68fd4f06aeac1a1d365ebd5469a0ddd2aa50", "relations": [{"evidence_url": "https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/README.md", "target_id": "arxiv:1802.08802", "type": "documented_by"}], "resource_id": "github:stanfordnlp/wge", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/stanfordnlp/wge", "topics": ["reinforcement-learning", "web-navigation", "workflow", "history"], "updated_at": null}, "text": "Workflow-Guided Exploration research code\nResource type: framework\nReview status: curated\nScope: core\n\nResearch code that learns browser-control policies by constraining exploration with workflows induced from demonstrations, with a Selenium-based MiniWoB environment and experiment harness.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, workflow, history\n\nRecorded specifications: {\"availability\": \"Public historical training and environment-wrapper source; legacy dependencies documented.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/stanfordnlp/wge/HEAD/LICENSE\"], \"reason\": \"Stanford WGE source license; not a blanket license for all third-party tasks, embeddings or demonstration resources.\"}}, \"repository_revision\": \"bed526657a62a64a163f80deb4a8627cb63871a3\"}\n\nThe released repository includes training policies, experiment configuration and a MiniWoB sandbox extended for the 2018 paper. This is the WGE implementation, not another copy of the already cataloged Farama MiniWoB++ benchmark and not the original OpenAI MiniWoB release. Its documented dependency is PyTorch 0.1.12, and current runtime compatibility was not tested. Stanford code declares Apache-2.0; bundled third-party browser tasks and externally downloaded demonstration/embedding data retain their own terms.\n\nPrimary source: https://github.com/stanfordnlp/wge\nCatalog record: https://awesomecua.github.io/resources/github-stanfordnlp-wge/"}
{"id": "github:steel-dev/steel-browser", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/steel-dev/steel-browser", "citation_url": "https://awesomecua.github.io/resources/github-steel-dev-steel-browser/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "668949f121cb5a8ac262a8d3222a31eaee6692bfc0de9b86e60ae04f885ccc9b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/steel-dev/steel-browser"}], "name": "Steel Browser", "publication": null, "published_at": null, "record_sha256": "148fcadfad929cf99a5512bd862d5ee882a93d0138e051725fecdf251b8c6056", "relations": [{"evidence_url": "https://raw.githubusercontent.com/steel-dev/steel-browser/HEAD/README.md", "target_id": "organization:steel", "type": "published_by"}], "resource_id": "github:steel-dev/steel-browser", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/steel-dev/steel-browser", "topics": ["sandbox", "web-navigation"], "updated_at": null}, "text": "Steel Browser\nResource type: infrastructure\nReview status: curated\nScope: core\n\nOpen-source browser API managing Chrome sessions for agents, reachable through Puppeteer, Playwright or Selenium over CDP, with persisted cookies, proxy rotation, stealth and a session debugging UI.\nPlatforms: web\nTopics: sandbox, web-navigation\n\nPrimary source: https://github.com/steel-dev/steel-browser\nCatalog record: https://awesomecua.github.io/resources/github-steel-dev-steel-browser/"}
{"id": "github:steel-dev/surf.new", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/steel-dev/surf.new", "citation_url": "https://awesomecua.github.io/resources/github-steel-dev-surf-new/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "b92abb5fdf714e569cf9f0d2c85ac3fb91ed67365e371a4b322aafe6fab39a1d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "efae611b1201878994c77b110867dcb65fb24a098c398bf5dec6138a47662bbc", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/steel-dev/surf.new/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/steel-dev/surf.new"}], "name": "Surf New", "publication": null, "published_at": null, "record_sha256": "6eabd8166de180572ce03856040b38dac177c3100ebc86f5363dd534dff488e7", "relations": [], "resource_id": "github:steel-dev/surf.new", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/steel-dev/surf.new", "topics": ["web-navigation"], "updated_at": null}, "text": "Surf New\nResource type: agent\nReview status: curated\nScope: core\n\nNext.js and FastAPI playground from Steel for trying web agents in a controlled browser environment, with live views of browser interaction and support for providers such as Claude and GPT-4.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/steel-dev/surf.new\nCatalog record: https://awesomecua.github.io/resources/github-steel-dev-surf-new/"}
{"id": "github:stepfun-ai/gelab-zero", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/stepfun-ai/gelab-zero", "citation_url": "https://awesomecua.github.io/resources/github-stepfun-ai-gelab-zero/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "462e2b627989a969f7e859e22040a83d38e7c3eac12305d844563b0925f8a787", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "08095ca875809a84572af7129caef5d4f4b052cc32ced7bdeb010fca125e07f0", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/stepfun-ai/gelab-zero/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/stepfun-ai/gelab-zero"}], "name": "GELab Zero", "publication": null, "published_at": null, "record_sha256": "87980df01e399076d976e9d3d83f5a788789bc9b2d75c5fea0aaa565f032da37", "relations": [], "resource_id": "github:stepfun-ai/gelab-zero", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/stepfun-ai/gelab-zero", "topics": ["mobile-navigation"], "updated_at": null}, "text": "GELab Zero\nResource type: model\nReview status: curated\nScope: core\n\nOpen mobile GUI agent stack pairing a 4B model that runs on consumer hardware with local multi-device infrastructure, an MCP server, and AndroidDaily benchmark tasks.\nPlatforms: android\nTopics: mobile-navigation\n\nPrimary source: https://github.com/stepfun-ai/gelab-zero\nCatalog record: https://awesomecua.github.io/resources/github-stepfun-ai-gelab-zero/"}
{"id": "github:stonybrooknlp/appworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/StonyBrookNLP/appworld", "citation_url": "https://awesomecua.github.io/resources/github-stonybrooknlp-appworld/", "details": {"benchmark_protocol": {"actions": ["Python code", "application API calls"], "counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "original AppWorld benchmark (2024 paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.18901", "unit": "tasks", "value": 750}, {"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50", "population": "original AppWorld environment (2024 paper)", "qualifier": "exact", "reference_url": "https://arxiv.org/abs/2407.18901", "unit": "applications", "value": 9}], "evaluation": "State-based unit tests check task outcomes and unintended changes.", "evidence_urls": ["https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"], "focus": ["interactive coding", "cross-app tool use"], "notes": "This is an adjacent API/coding benchmark, not a screenshot-only task suite. Preserve the named release; task counts here refer to the original paper.", "observations": ["API documentation", "execution feedback"], "setting": "interactive"}, "leaderboard_url": "https://appworld.dev/appworld/leaderboard"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:45:03Z", "role": "primary", "sha256": "f4049141b6e520217395b7ccd9f27048c8e0f90331690dcbef300d2bc4ff7bdd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:32Z", "role": "primary", "sha256": "b9a83994d8ac36ccdb606a5219511f8bb35ad3948ade4f9f3e28d386221aedfb", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/StonyBrookNLP/appworld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:45:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/StonyBrookNLP/appworld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/StonyBrookNLP/appworld"}], "name": "AppWorld", "publication": null, "published_at": null, "record_sha256": "e0a73704d715d454dda3fb1b63dc34bb95f9ff87194f41b1c03aa72e3b466784", "relations": [{"evidence_url": "https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md", "target_id": "arxiv:2407.18901", "type": "documented_by"}], "resource_id": "github:stonybrooknlp/appworld", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/StonyBrookNLP/appworld", "topics": ["evaluation", "tool-integration", "cross-app"], "updated_at": null}, "text": "AppWorld\nResource type: benchmark\nReview status: curated\nScope: adjacent\n\nSimulated applications, user data, and state-based task checks for evaluating agents that interact through APIs and code.\nPlatforms: unspecified\nTopics: evaluation, tool-integration, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"Python code\", \"application API calls\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\", \"population\": \"original AppWorld benchmark (2024 paper)\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2407.18901\", \"unit\": \"tasks\", \"value\": 750}, {\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\", \"population\": \"original AppWorld environment (2024 paper)\", \"qualifier\": \"exact\", \"reference_url\": \"https://arxiv.org/abs/2407.18901\", \"unit\": \"applications\", \"value\": 9}], \"evaluation\": \"State-based unit tests check task outcomes and unintended changes.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/stonybrooknlp/appworld/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2406.12814%2C2406.13719%2C2406.14056%2C2406.14250%2C2406.14314%2C2406.18082%2C2406.19263%2C2407.00993%2C2407.01476%2C2407.01489%2C2407.01511%2C2407.01558%2C2407.03037%2C2407.03913%2C2407.04346%2C2407.05291%2C2407.07061%2C2407.08850%2C2407.09018%2C2407.09295%2C2407.10956%2C2407.13032%2C2407.13063%2C2407.15673%2C2407.15711%2C2407.16741%2C2407.17490%2C2407.18901%2C2407.19056%2C2407.20183%2C2408.00203%2C2408.02544%2C2408.06327%2C2408.07199%2C2408.11824%2C2408.15978%2C2409.02453%2C2409.04566%2C2409.07429%2C2409.08264%2C2409.09354%2C2409.10741%2C2409.11295%2C2409.13711%2C2409.14337%2C2409.14818%2C2409.15441%2C2409.15637%2C2409.17140%2C2409.20566&max_results=50\"], \"focus\": [\"interactive coding\", \"cross-app tool use\"], \"notes\": \"This is an adjacent API/coding benchmark, not a screenshot-only task suite. Preserve the named release; task counts here refer to the original paper.\", \"observations\": [\"API documentation\", \"execution feedback\"], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://appworld.dev/appworld/leaderboard\"}\n\nPrimarily API and coding interaction; included as an adjacent comparison for GUI and hybrid agents.\n\nPrimary source: https://github.com/StonyBrookNLP/appworld\nCatalog record: https://awesomecua.github.io/resources/github-stonybrooknlp-appworld/"}
{"id": "github:sunblaze-ucb/agentsynth", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/sunblaze-ucb/AgentSynth", "citation_url": "https://awesomecua.github.io/resources/github-sunblaze-ucb-agentsynth/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"], "reason": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate."}}, "repository_tree_sha": "cd42a4bd158fe3f0a1b0f376b9428d6a5fcb81b2", "repository_url": "https://github.com/sunblaze-ucb/AgentSynth"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:00Z", "role": "primary", "sha256": "e08f08a99b8cb3a4bb857a9fac976b2820f9ad406d66675b0286e880dfeeceaa", "supports": ["identity"], "url": "https://api.github.com/repos/sunblaze-ucb/AgentSynth"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:20Z", "role": "primary", "sha256": "b0998652a9b6b465444c2bbeef015bdbef50e83984a3bd15725f0e811645ca62", "supports": ["identity", "release_status", "details.repository_tree_sha"], "url": "https://api.github.com/repos/sunblaze-ucb/AgentSynth/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": {"basis": "Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/sunblaze-ucb/AgentSynth"}], "name": "AgentSynth task-generation pipeline", "publication": null, "published_at": null, "record_sha256": "70c753b1f58ded356a2d878b35d5d8c3f430a93fbf6b49dd2d61a4d3d34f4165", "relations": [{"evidence_url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md", "target_id": "arxiv:2506.14205", "type": "documented_by"}], "resource_id": "github:sunblaze-ucb/agentsynth", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/sunblaze-ucb/AgentSynth", "topics": ["data-generation", "long-horizon", "automation"], "updated_at": null}, "text": "AgentSynth task-generation pipeline\nResource type: framework\nReview status: curated\nScope: core\n\nGenerates synthetic computer-use tasks and records trajectories in OSWorld desktop environments and InSTA browser environments.\nPlatforms: desktop, web\nTopics: data-generation, long-horizon, automation\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/LICENSE\"], \"reason\": \"Reviewed repository license declaration for code; third-party dependencies, data, weights and executable environment images are separate.\"}}, \"repository_tree_sha\": \"cd42a4bd158fe3f0a1b0f376b9428d6a5fcb81b2\", \"repository_url\": \"https://github.com/sunblaze-ucb/AgentSynth\"}\n\nReleased code includes task-generation and training-format conversion scripts. OSWorld/InSTA environments and model services require separate setup; the repository is not a bundled VM or model checkpoint. Code and the separately published dataset have distinct license declarations.\n\nPrimary source: https://github.com/sunblaze-ucb/AgentSynth\nCatalog record: https://awesomecua.github.io/resources/github-sunblaze-ucb-agentsynth/"}
{"id": "github:tencent/points-gui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Tencent/POINTS-GUI", "citation_url": "https://awesomecua.github.io/resources/github-tencent-points-gui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tencent/POINTS-GUI"}], "name": "POINTS-GUI", "publication": null, "published_at": null, "record_sha256": "d504136c11252c9514d613180b4d60646dcb2f6c20eda3188e4940c9449fca24", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "resource_id": "github:tencent/points-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Tencent/POINTS-GUI", "topics": ["grounding"], "updated_at": null}, "text": "POINTS-GUI\nResource type: model\nReview status: curated\nScope: core\n\nPOINTS-GUI-G-8B GUI grounding model, trained from a POINTS-1.5 base without native grounding through a data pipeline that unifies coordinates, filters noise and adds hard cases.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/Tencent/POINTS-GUI\nCatalog record: https://awesomecua.github.io/resources/github-tencent-points-gui/"}
{"id": "github:tencentqqgylab/appagent", "metadata": {"aliases": ["github:mnotgod96/appagent", "mnotgod96/AppAgent", "https://github.com/mnotgod96/AppAgent"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/TencentQQGYLab/AppAgent", "citation_url": "https://awesomecua.github.io/resources/github-tencentqqgylab-appagent/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "ac94cca1f69807d2d31d30bd5521fbf1bda6155996881054df87a85522cd8f5b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "d9e2b107cbaa3599336510dbb6703cb30210f1b4345221e76cf75cba85c2d113", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:57Z", "role": "primary", "sha256": "cd15c021e5023f43c17cc6365a58177c08ae91ac7e854c64c0e531f4ed3c5d5f", "supports": ["aliases"], "url": "https://api.github.com/repos/mnotgod96/AppAgent"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:13:58Z", "role": "primary", "sha256": "cd15c021e5023f43c17cc6365a58177c08ae91ac7e854c64c0e531f4ed3c5d5f", "supports": ["aliases"], "url": "https://api.github.com/repos/TencentQQGYLab/AppAgent"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TencentQQGYLab/AppAgent"}], "name": "AppAgent", "publication": null, "published_at": null, "record_sha256": "5b1c6276542c98088a7f10dbd2830d8217e2f260fa892205e73549c5be37b088", "relations": [{"evidence_url": "https://raw.githubusercontent.com/TencentQQGYLab/AppAgent/HEAD/README.md", "target_id": "arxiv:2312.13771", "type": "documented_by"}], "resource_id": "github:tencentqqgylab/appagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/TencentQQGYLab/AppAgent", "topics": ["mobile-navigation", "memory"], "updated_at": null}, "text": "AppAgent\nResource type: agent\nReview status: curated\nScope: core\n\nMultimodal agent that operates Android apps through taps and swipes without back-end access, building a reusable knowledge base from autonomous exploration or from watching human demonstrations.\nRecorded aliases: github:mnotgod96/appagent, mnotgod96/AppAgent, https://github.com/mnotgod96/AppAgent\nPlatforms: android\nTopics: mobile-navigation, memory\n\nPrimary source: https://github.com/TencentQQGYLab/AppAgent\nCatalog record: https://awesomecua.github.io/resources/github-tencentqqgylab-appagent/"}
{"id": "github:theagentcompany/theagentcompany", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/TheAgentCompany/TheAgentCompany", "citation_url": "https://awesomecua.github.io/resources/github-theagentcompany-theagentcompany/", "details": {"benchmark_protocol": {"actions": ["web browsing", "code execution", "coworker communication"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md", "population": "workspace images documented by the README", "qualifier": "exact", "unit": "task images", "value": 175}], "evaluation": "Task-specific evaluator scripts run within the prepared workplace environment.", "evidence_urls": ["https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md"], "focus": ["professional workflows", "multi-tool tasks"], "notes": "Each task is packaged as a Docker image. Record task-image and server versions; the environment includes simulated coworkers and external application state.", "observations": [], "setting": "interactive"}, "leaderboard_url": "https://the-agent-company.com/#/leaderboard"}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "d1646f67d49d38ef1e81bf7b5057dfc3d6f172b4c8a7413a2cc78741ccb08723", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "2d6976a9b482bb14d9cd4f0d97ce6e14305496f1649b8667cb91d027c670930f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:34Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TheAgentCompany/TheAgentCompany"}], "name": "TheAgentCompany", "publication": null, "published_at": null, "record_sha256": "f79ee40b4d309499f6abcf4fbd7d4b6550c4c26cb4b0e6c19e411108789961da", "relations": [{"evidence_url": "https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md", "target_id": "arxiv:2412.14161", "type": "documented_by"}], "resource_id": "github:theagentcompany/theagentcompany", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/TheAgentCompany/TheAgentCompany", "topics": ["evaluation", "workflow", "cross-app"], "updated_at": null}, "text": "TheAgentCompany\nResource type: benchmark\nReview status: curated\nScope: core\n\nSelf-contained workplace environment for evaluating agents on professional tasks involving websites, files, code, and simulated coworkers.\nPlatforms: web, desktop\nTopics: evaluation, workflow, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"web browsing\", \"code execution\", \"coworker communication\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md\", \"population\": \"workspace images documented by the README\", \"qualifier\": \"exact\", \"unit\": \"task images\", \"value\": 175}], \"evaluation\": \"Task-specific evaluator scripts run within the prepared workplace environment.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/TheAgentCompany/TheAgentCompany/HEAD/README.md\"], \"focus\": [\"professional workflows\", \"multi-tool tasks\"], \"notes\": \"Each task is packaged as a Docker image. Record task-image and server versions; the environment includes simulated coworkers and external application state.\", \"observations\": [], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://the-agent-company.com/#/leaderboard\"}\n\nPrimary source: https://github.com/TheAgentCompany/TheAgentCompany\nCatalog record: https://awesomecua.github.io/resources/github-theagentcompany-theagentcompany/"}
{"id": "github:thu-ml/mla-trust", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/thu-ml/MLA-Trust", "citation_url": "https://awesomecua.github.io/resources/github-thu-ml-mla-trust/", "details": {}, "environments": ["web", "android", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "d69da19edaed185ad3071a97db3bb3fcfbdb14c0b514b7d27794b39fbe8e2541", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-18T11:03:46Z", "role": "primary", "sha256": "240fb6820e7bcb0753617bcad6ffbd4a9ba83ca98e1e6aaa06c7b197b244bf16", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-18T11:03:46Z", "url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/thu-ml/MLA-Trust"}], "name": "MLA-Trust", "publication": null, "published_at": null, "record_sha256": "f2a2406013e9ae6630ff4dc4c0b41319fd1db4e5ab642d8f586a21df3706cb66", "relations": [{"evidence_url": "https://raw.githubusercontent.com/thu-ml/MLA-Trust/HEAD/README.md", "target_id": "arxiv:2506.01616", "type": "documented_by"}], "resource_id": "github:thu-ml/mla-trust", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/thu-ml/MLA-Trust", "topics": ["safety-security", "evaluation", "robustness"], "updated_at": null}, "text": "MLA-Trust\nResource type: benchmark\nReview status: curated\nScope: core\n\nUnified evaluation framework measuring truthfulness, controllability, safety, and privacy of multimodal LLM agents across interactive web and mobile GUI tasks.\nPlatforms: web, android, cross-platform\nTopics: safety-security, evaluation, robustness\n\nPrimary source: https://github.com/thu-ml/MLA-Trust\nCatalog record: https://awesomecua.github.io/resources/github-thu-ml-mla-trust/"}
{"id": "github:thudm/agentbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/AgentBench", "citation_url": "https://awesomecua.github.io/resources/github-thudm-agentbench/", "details": {"leaderboard_url": "https://docs.google.com/spreadsheets/d/e/2PACX-1vRR3Wl7wsCgHpwUw1_eUXW_fptAPLL3FkhnW_rua0O1Ji_GIVrpTjY5LaKAhwO-WeARjnY_KNw0SYNJ/pubhtml"}, "environments": ["cross-platform", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "1b7d1f1cd7a74ac9497c639324ec342dd3f8769e9c0c4cd2cd92a4eb6740ac98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/AgentBench"}], "name": "AgentBench", "publication": null, "published_at": null, "record_sha256": "394e30cd6939007aef24e38da08115d7d23ffb68c44b2cfa728a0fe49ab18b50", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/AgentBench/HEAD/README.md", "target_id": "arxiv:2308.03688", "type": "documented_by"}], "resource_id": "github:thudm/agentbench", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/THUDM/AgentBench", "topics": ["evaluation", "tool-integration"], "updated_at": null}, "text": "AgentBench\nResource type: benchmark\nReview status: curated\nScope: adjacent\n\nAgent evaluation suite with containerized tasks and a function-calling interface, including operating-system and shopping environments.\nPlatforms: cross-platform, web\nTopics: evaluation, tool-integration\n\nRecorded specifications: {\"leaderboard_url\": \"https://docs.google.com/spreadsheets/d/e/2PACX-1vRR3Wl7wsCgHpwUw1_eUXW_fptAPLL3FkhnW_rua0O1Ji_GIVrpTjY5LaKAhwO-WeARjnY_KNw0SYNJ/pubhtml\"}\n\nPrimary source: https://github.com/THUDM/AgentBench\nCatalog record: https://awesomecua.github.io/resources/github-thudm-agentbench/"}
{"id": "github:thudm/android-lab", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/Android-Lab", "citation_url": "https://awesomecua.github.io/resources/github-thudm-android-lab/", "details": {"benchmark_protocol": {"actions": ["Android operations through the agent framework"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "population": "benchmark task collection", "qualifier": "exact", "unit": "tasks", "value": 138}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "population": "benchmark application set", "qualifier": "exact", "unit": "applications", "value": 9}], "evaluation": "The evaluation pipeline produces success, subgoal success, reversed redundancy, and reasonable-operation metrics; result generation requires a configured judge model.", "evidence_urls": ["https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"], "focus": ["Android task completion", "subgoal completion", "operation efficiency"], "notes": "The benchmark uses predefined virtual-device images and apps chosen to work without network/login requirements. XML and SoM modes should be reported separately. These counts concern evaluation tasks, not the separately released Android Instruct training set.", "observations": ["UI XML", "set-of-marks views"], "setting": "interactive"}, "evaluation_type": "benchmark task evaluation", "leaderboard_url": "https://docs.google.com/spreadsheets/d/1Zv6mBfd4Ibt8mke24K6zAFrXe4AvEFaBLW3hpJqSgjw/edit?gid=0#gid=0", "metrics": ["task success"], "task_count": 138}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "340680a9a7833d103234b8367195deebd29bfec7e25ebc50c2dc04fc3627e55b", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:58Z", "role": "primary", "sha256": "1bbbee919161337e2456c587c6be809214406482cd9f1292a9eb62428f6961f3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:41:58Z", "url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/Android-Lab"}], "name": "AndroidLab", "publication": null, "published_at": null, "record_sha256": "4974e5bbcb848e665dbfb0d8d69d4e8aa78a3201a37eaf73f9bf7eca00728fdf", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md", "target_id": "arxiv:2410.24024", "type": "documented_by"}], "resource_id": "github:thudm/android-lab", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/Android-Lab", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "AndroidLab\nResource type: benchmark\nReview status: curated\nScope: core\n\nReproducible Android benchmark of 138 tasks across nine apps on predefined virtual devices, run via AVD on Mac or Docker on Linux, released with the Android Instruct training set.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"Android operations through the agent framework\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md\", \"population\": \"benchmark task collection\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 138}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md\", \"population\": \"benchmark application set\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 9}], \"evaluation\": \"The evaluation pipeline produces success, subgoal success, reversed redundancy, and reasonable-operation metrics; result generation requires a configured judge model.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/THUDM/Android-Lab/HEAD/README.md\"], \"focus\": [\"Android task completion\", \"subgoal completion\", \"operation efficiency\"], \"notes\": \"The benchmark uses predefined virtual-device images and apps chosen to work without network/login requirements. XML and SoM modes should be reported separately. These counts concern evaluation tasks, not the separately released Android Instruct training set.\", \"observations\": [\"UI XML\", \"set-of-marks views\"], \"setting\": \"interactive\"}, \"evaluation_type\": \"benchmark task evaluation\", \"leaderboard_url\": \"https://docs.google.com/spreadsheets/d/1Zv6mBfd4Ibt8mke24K6zAFrXe4AvEFaBLW3hpJqSgjw/edit?gid=0#gid=0\", \"metrics\": [\"task success\"], \"task_count\": 138}\n\nPrimary source: https://github.com/THUDM/Android-Lab\nCatalog record: https://awesomecua.github.io/resources/github-thudm-android-lab/"}
{"id": "github:thudm/autowebglm", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/AutoWebGLM", "citation_url": "https://awesomecua.github.io/resources/github-thudm-autowebglm/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "9d18b958cfefd47100aa32ab14374c22ababb997a637a80da98b1977dec83fe0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:04Z", "role": "primary", "sha256": "0d97778d2d5e7439d1fbb0ce628565e7bd49886b2e86e6e6ca3d481c7fdc3256", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:04Z", "url": "https://raw.githubusercontent.com/THUDM/AutoWebGLM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/AutoWebGLM"}], "name": "AutoWebGLM", "publication": null, "published_at": null, "record_sha256": "351eb0bf9c7cec41cb86eabf9934281d15801126f0db0aa5c83d24b1a817dee6", "relations": [], "resource_id": "github:thudm/autowebglm", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/AutoWebGLM", "topics": ["web-navigation"], "updated_at": null}, "text": "AutoWebGLM\nResource type: agent\nReview status: curated\nScope: core\n\nWeb navigation agent built on ChatGLM3-6B that simplifies HTML for the model, is trained on hybrid human and AI data with RL and rejection sampling, and comes with the bilingual AutoWebBench.\nPlatforms: web\nTopics: web-navigation\n\nPrimary source: https://github.com/THUDM/AutoWebGLM\nCatalog record: https://awesomecua.github.io/resources/github-thudm-autowebglm/"}
{"id": "github:thudm/cogagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/CogAgent", "citation_url": "https://awesomecua.github.io/resources/github-thudm-cogagent/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f025e4bc9d920ec982525394c69b87043aae87fbe719de4b9e973408a44211e2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "c6a77f53ba7933f07b95807e972e50259921344e777e52a6a8bc690373df2d74", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/CogAgent"}], "name": "CogAgent", "publication": null, "published_at": null, "record_sha256": "3907e81b53e162f649baebb1f5f51f0fc5a750b5c52b442bb042c32ec35c7978", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/CogAgent/HEAD/README.md", "target_id": "arxiv:2312.08914", "type": "documented_by"}], "resource_id": "github:thudm/cogagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/CogAgent", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "CogAgent\nResource type: model\nReview status: curated\nScope: core\n\nVisual language model for GUI understanding and interaction.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\n\nPrimary source: https://github.com/THUDM/CogAgent\nCatalog record: https://awesomecua.github.io/resources/github-thudm-cogagent/"}
{"id": "github:thudm/computerrl", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/ComputerRL", "citation_url": "https://awesomecua.github.io/resources/github-thudm-computerrl/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "8b327dcd58c18e48e1e7f5b19ebf5673da4b5ff9320aa960e391844adf3a8a55", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/ComputerRL"}], "name": "ComputerRL", "publication": null, "published_at": null, "record_sha256": "4588c7bb89723fd46a6f904f89d6f80a6c8f93b115402d17aa544a98f714cca4", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/ComputerRL/HEAD/README.md", "target_id": "arxiv:2508.14040", "type": "documented_by"}], "resource_id": "github:thudm/computerrl", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/ComputerRL", "topics": ["reinforcement-learning"], "updated_at": null}, "text": "ComputerRL\nResource type: framework\nReview status: curated\nScope: core\n\nReinforcement-learning infrastructure for computer-use agents.\nPlatforms: desktop\nTopics: reinforcement-learning\n\nPrimary source: https://github.com/THUDM/ComputerRL\nCatalog record: https://awesomecua.github.io/resources/github-thudm-computerrl/"}
{"id": "github:thudm/mobilerl", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/MobileRL", "citation_url": "https://awesomecua.github.io/resources/github-thudm-mobilerl/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "752fad8ebd8b8eda03648329a6fc2585e5bcb2b96875b8aa2ce7d9f3fff78153", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "1bbbee919161337e2456c587c6be809214406482cd9f1292a9eb62428f6961f3", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/MobileRL"}], "name": "MobileRL", "publication": null, "published_at": null, "record_sha256": "a7e583c01705351809c89a1e70c667fdf1208fefb9e8069275fe6220020f73db", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md", "target_id": "arxiv:2509.18119", "type": "documented_by"}], "resource_id": "github:thudm/mobilerl", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/MobileRL", "topics": ["reinforcement-learning"], "updated_at": null}, "text": "MobileRL\nResource type: framework\nReview status: curated\nScope: core\n\nOnline agentic RL framework for mobile GUI agents built around the difficulty-adaptive ADAGRPO algorithm, with Dockerized AndroidWorld and AndroidLab evaluation and a MobileRL-9B checkpoint.\nPlatforms: android\nTopics: reinforcement-learning\n\nPrimary source: https://github.com/THUDM/MobileRL\nCatalog record: https://awesomecua.github.io/resources/github-thudm-mobilerl/"}
{"id": "github:thudm/scale-cua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/SCALE-CUA", "citation_url": "https://awesomecua.github.io/resources/github-thudm-scale-cua/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "d4a68b00acc4becf2a25688520595babea9b694b56541e709f066413b5b80aea", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/SCALE-CUA/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/SCALE-CUA"}], "name": "SCALE-CUA (THUDM)", "publication": null, "published_at": null, "record_sha256": "234134760c7334c4009f1f8282d89709bfaaf3887efcc6a68e1604bd573371bf", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/SCALE-CUA/HEAD/README.md", "target_id": "arxiv:2607.11185", "type": "documented_by"}], "resource_id": "github:thudm/scale-cua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/SCALE-CUA", "topics": ["reinforcement-learning", "data-generation", "efficiency"], "updated_at": null}, "text": "SCALE-CUA (THUDM)\nResource type: framework\nReview status: curated\nScope: core\n\nComputer-use training framework with VeriGen task synthesis, adaptive rollout sampling, and visual-context segmentation.\nPlatforms: desktop\nTopics: reinforcement-learning, data-generation, efficiency\n\nA distinct project from OpenGVLab/ScaleCUA, which documents arXiv:2509.15221. Similar names do not establish a shared project identity.\n\nPrimary source: https://github.com/THUDM/SCALE-CUA\nCatalog record: https://awesomecua.github.io/resources/github-thudm-scale-cua/"}
{"id": "github:thudm/visualagentbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/VisualAgentBench", "citation_url": "https://awesomecua.github.io/resources/github-thudm-visualagentbench/", "details": {}, "environments": ["cross-platform", "web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "21afbc4086441dee131fe4b84986870f9b280413dd2a38df6e99406f1fc45586", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/VisualAgentBench"}], "name": "VisualAgentBench", "publication": null, "published_at": null, "record_sha256": "ac8cd2c8f4463ef4fd3adf5a4147c7ec4d16050b6596fd4a91a33bf687b43505", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/VisualAgentBench/HEAD/README.md", "target_id": "arxiv:2408.06327", "type": "documented_by"}], "resource_id": "github:thudm/visualagentbench", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/THUDM/VisualAgentBench", "topics": ["evaluation"], "updated_at": null}, "text": "VisualAgentBench\nResource type: benchmark\nReview status: curated\nScope: adjacent\n\nEvaluation framework spanning visual GUI, embodied, and design tasks, including mobile and WebArena-based environments.\nPlatforms: cross-platform, web, android\nTopics: evaluation\n\nPrimary source: https://github.com/THUDM/VisualAgentBench\nCatalog record: https://awesomecua.github.io/resources/github-thudm-visualagentbench/"}
{"id": "github:thudm/webrl", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/THUDM/WebRL", "citation_url": "https://awesomecua.github.io/resources/github-thudm-webrl/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "eb217b96c54c3c2ccfd3a27255c6fd740df40c2d6743ae811858339bf7f740b0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "42496cf0cff06a7f9e794f4313a10a670bc48df9909ec6d57763cb902bb229ea", "supports": ["identity"], "url": "https://api.github.com/repos/THUDM/WebRL"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:05Z", "role": "primary", "sha256": "bb1683c6fb5ff658c89f0ebf0f8baf8026bb703840e174c9e46b223c85a16856", "supports": ["release_status"], "url": "https://api.github.com/repos/THUDM/WebRL/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/THUDM/WebRL"}], "name": "WebRL", "publication": null, "published_at": null, "record_sha256": "d06ae42c3f1edcc40d2d9f8ca9ceea69e8707dea0aefc14782de924a749913cb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "resource_id": "github:thudm/webrl", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/THUDM/WebRL", "topics": ["reinforcement-learning", "web-navigation", "data-generation"], "updated_at": null}, "text": "WebRL\nResource type: framework\nReview status: curated\nScope: core\n\nTraining pipeline that evolves web tasks from prior attempts and uses an outcome reward model to refine browser policies.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, data-generation\n\nTraining code and checkpoint links are released. No root code-license declaration was found in the inspected repository metadata; bundled dependency licenses do not establish a license for the whole project. Weights have separate Hub terms.\n\nPrimary source: https://github.com/THUDM/WebRL\nCatalog record: https://awesomecua.github.io/resources/github-thudm-webrl/"}
{"id": "github:tml-epfl/os-harm", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/tml-epfl/os-harm", "citation_url": "https://awesomecua.github.io/resources/github-tml-epfl-os-harm/", "details": {"benchmark_protocol": {"actions": ["OSWorld/PyAutoGUI desktop interaction"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "population": "reported benchmark total; not a recount of unique released task IDs", "qualifier": "exact", "unit": "tasks as reported by the paper/README", "value": 150}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json", "population": "released misuse task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 50}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json", "population": "released injection task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 10}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json", "population": "released misbehavior task-ID list; before injection/goal expansion", "qualifier": "exact", "unit": "task IDs", "value": 50}], "evaluation": "Automated LLM judge scores both task accuracy and safety from saved execution traces. Keep judge model/prompt version and harm category with a result; the scripts separately compute predicted metrics. This is an interactive desktop safety benchmark, not a single offline grounding score.", "evidence_urls": ["https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/run.py", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json", "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json"], "focus": ["deliberate misuse", "prompt-injection safety", "unintended harmful behavior during normal tasks"], "notes": "README/paper reports 150 tasks, whereas retained misuse/injection/misbehavior lists contain 50/10/50 IDs. run.py expands injection vectors/goals, so the released ID inventory is not a direct recount of paper settings; the complete expanded total was not reconstructed. Judge agreement F1 values in the README measure agreement with human labels, not agent success. Experiments use OSWorld desktop apps, with VMware documented as tested and screenshot/a11y/SOM observation choices. Code Apache-2.0 does not establish VM/app/data rights. Publisher asks that the benchmark not be used directly as agent training data.", "observations": ["screenshot", "accessibility tree", "screenshot plus accessibility tree", "set-of-marks"], "setting": "interactive"}, "evaluation_type": "automated judgment of accuracy and safety", "metrics": ["task completion", "safety"], "task_count": 150, "task_unit": "tasks reported by the paper/README; not a recount of released task IDs"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:36:33Z", "role": "primary", "sha256": "d001158f58ac37b01c8470224db1567a52feb8010877305f7966637cf7400706", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:23:14Z", "role": "primary", "sha256": "86dd59a252fd654fc3244268a75de37b2c7f97e16f6b1de8c4cf885458a422f7", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/run.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "8ff6092a51cc924c5225ace01c011c6481a2461eb0a47b7360eddaf6619b898c", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "b4c451236bb4ff36eb7e790dfa8cb517bffb245602bc441080314b13c8aba54f", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "8eb6bade72696c3c0b53160a05b006a773c9e553ab5e401418b54d9acc98c0f2", "supports": ["details.benchmark_protocol", "notes"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T09:39:34Z", "role": "primary", "sha256": "7c3ab1eeeb073d92238ecb65ffaebe58ab5922929b115a7c0b64eb4f9a9afc2f", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T09:36:33Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T09:39:34Z", "url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/tml-epfl/os-harm"}], "name": "OS-Harm", "publication": null, "published_at": null, "record_sha256": "f3ccc02a277b486dd3a80947c956667e106c91c22482a0bb03513f127e89741e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md", "target_id": "arxiv:2506.14866", "type": "documented_by"}], "resource_id": "github:tml-epfl/os-harm", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/tml-epfl/os-harm", "topics": ["safety-security", "evaluation"], "updated_at": null}, "text": "OS-Harm\nResource type: benchmark\nReview status: curated\nScope: core\n\nBenchmark for safety of computer-use agents across misuse, prompt injection, and unintended model behavior.\nPlatforms: desktop\nTopics: safety-security, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"OSWorld/PyAutoGUI desktop interaction\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md\", \"population\": \"reported benchmark total; not a recount of unique released task IDs\", \"qualifier\": \"exact\", \"unit\": \"tasks as reported by the paper/README\", \"value\": 150}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json\", \"population\": \"released misuse task-ID list; before injection/goal expansion\", \"qualifier\": \"exact\", \"unit\": \"task IDs\", \"value\": 50}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json\", \"population\": \"released injection task-ID list; before injection/goal expansion\", \"qualifier\": \"exact\", \"unit\": \"task IDs\", \"value\": 10}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json\", \"population\": \"released misbehavior task-ID list; before injection/goal expansion\", \"qualifier\": \"exact\", \"unit\": \"task IDs\", \"value\": 50}], \"evaluation\": \"Automated LLM judge scores both task accuracy and safety from saved execution traces. Keep judge model/prompt version and harm category with a result; the scripts separately compute predicted metrics. This is an interactive desktop safety benchmark, not a single offline grounding score.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/README.md\", \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/run.py\", \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misuse.json\", \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_injection.json\", \"https://raw.githubusercontent.com/tml-epfl/os-harm/HEAD/evaluation_examples/test_misbehavior.json\"], \"focus\": [\"deliberate misuse\", \"prompt-injection safety\", \"unintended harmful behavior during normal tasks\"], \"notes\": \"README/paper reports 150 tasks, whereas retained misuse/injection/misbehavior lists contain 50/10/50 IDs. run.py expands injection vectors/goals, so the released ID inventory is not a direct recount of paper settings; the complete expanded total was not reconstructed. Judge agreement F1 values in the README measure agreement with human labels, not agent success. Experiments use OSWorld desktop apps, with VMware documented as tested and screenshot/a11y/SOM observation choices. Code Apache-2.0 does not establish VM/app/data rights. Publisher asks that the benchmark not be used directly as agent training data.\", \"observations\": [\"screenshot\", \"accessibility tree\", \"screenshot plus accessibility tree\", \"set-of-marks\"], \"setting\": \"interactive\"}, \"evaluation_type\": \"automated judgment of accuracy and safety\", \"metrics\": [\"task completion\", \"safety\"], \"task_count\": 150, \"task_unit\": \"tasks reported by the paper/README; not a recount of released task IDs\"}\n\nThe retained task-ID manifests have 50 misuse, 10 injection and 50 misbehavior IDs. Prompt-injection execution expands vectors/goals; the reported 150-task total must not be read as 150 unique released IDs.\n\nPrimary source: https://github.com/tml-epfl/os-harm\nCatalog record: https://awesomecua.github.io/resources/github-tml-epfl-os-harm/"}
{"id": "github:tongui-agent/tongui-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/TongUI-agent/TongUI-agent", "citation_url": "https://awesomecua.github.io/resources/github-tongui-agent-tongui-agent/", "details": {"availability": "Public training configurations, preprocessing and inference examples; full experiment-script release incomplete.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"], "reason": "No project-level source-license declaration in the inspected tree; upstream dependency licenses do not establish this project’s grant."}}, "repository_revision": "82631b583180cdc870b4774d2495dd94f1558c46"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "90c209a9758f7b527def072cfa78ebc87f1b49b436fad57c4c7bb5a6b66ae790", "supports": ["identity"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "0c955a89dcda18652b09ce836c51c6ccaa007be6f6afeb48d21d01813cef1ec2", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/TongUI-agent/TongUI-agent"}], "name": "TongUI training and inference framework", "publication": null, "published_at": null, "record_sha256": "d3277084f313bf87a1abf1b40680b8a4f8b76c73ff8ef5a18574f5c26acbe235", "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "target_id": "url:8c9aaafb918aaf8b4e92", "type": "uses"}, {"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "resource_id": "github:tongui-agent/tongui-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/TongUI-agent/TongUI-agent", "topics": ["grounding", "imitation-learning", "data-generation"], "updated_at": null}, "text": "TongUI training and inference framework\nResource type: framework\nReview status: curated\nScope: core\n\nTraining configurations, preprocessing and inference examples for tutorial-trained GUI models, with a LLaMA-Factory fine-tuning workflow and screenshot-grounding interfaces.\nPlatforms: cross-platform, desktop, web\nTopics: grounding, imitation-learning, data-generation\n\nRecorded specifications: {\"availability\": \"Public training configurations, preprocessing and inference examples; full experiment-script release incomplete.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://api.github.com/repos/TongUI-agent/TongUI-agent/git/trees/main?recursive=1\"], \"reason\": \"No project-level source-license declaration in the inspected tree; upstream dependency licenses do not establish this project’s grant.\"}}, \"repository_revision\": \"82631b583180cdc870b4774d2495dd94f1558c46\"}\n\nThe public tree includes training configs, preprocessing, inference and serving examples. The project still marks release of all experiment/evaluation scripts as work in progress. This entry covers the implementation rather than counting its 3B/7B/32B model-size variants as separate new projects. The inspected project tree has no root source license; linked frameworks, data and checkpoints retain their own terms.\n\nPrimary source: https://github.com/TongUI-agent/TongUI-agent\nCatalog record: https://awesomecua.github.io/resources/github-tongui-agent-tongui-agent/"}
{"id": "github:tongyi-mai/mai-ui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Tongyi-MAI/MAI-UI", "citation_url": "https://awesomecua.github.io/resources/github-tongyi-mai-mai-ui/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE", "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"], "reason": "The implementation and LICENSE live under MAI-UI/. A root-level missing license or the follow-up Qwen-UI-Agent documentation should not change the scoped declaration."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:14Z", "role": "primary", "sha256": "f644a5c0d0eb29b0ff8d8552dd505c5ecedf7c9cd9ab8cf2c617a1a14ecd0c14", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "fd90bc766d7e73889d14e8d3c5c92d384fa86e163222808f6e492509f00a3656", "supports": ["identity"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:46Z", "role": "primary", "sha256": "4340b3d36bc56963fb8173079da8a0f7e91f195994d29fc47a9f7087cb56ffba", "supports": ["release_status"], "url": "https://api.github.com/repos/Tongyi-MAI/MAI-UI/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T05:29:15Z", "role": "primary", "sha256": "bf989559906a7926b77390082e4267ff7b2b494e8f599a9dc5640319eb45c2ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T05:29:14Z", "licenses": {"code": {"basis": "Retained MAI-UI/LICENSE and MAI-UI README; applies to the inspected implementation, subject to third-party components.", "declared": "Apache-2.0 (MAI-UI subdirectory)", "retrieved_at": "2026-10-17T05:29:15Z", "url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tongyi-MAI/MAI-UI"}], "name": "MAI-UI inference and evaluation code", "publication": null, "published_at": null, "record_sha256": "09293aa87e9162b8e921eb8ab271df4769c10b3893b777a6fdca484e47d6e238", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "resource_id": "github:tongyi-mai/mai-ui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Tongyi-MAI/MAI-UI", "topics": ["mobile-navigation", "grounding", "tool-integration"], "updated_at": null}, "text": "MAI-UI inference and evaluation code\nResource type: framework\nReview status: curated\nScope: core\n\nPublished MAI-UI grounding and navigation-agent implementations with inference demos and evaluation scripts.\nPlatforms: android\nTopics: mobile-navigation, grounding, tool-integration\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/LICENSE\", \"https://raw.githubusercontent.com/Tongyi-MAI/MAI-UI/HEAD/MAI-UI/README.md\"], \"reason\": \"The implementation and LICENSE live under MAI-UI/. A root-level missing license or the follow-up Qwen-UI-Agent documentation should not change the scoped declaration.\"}}}\n\nThe original implementation is in the MAI-UI/ subdirectory, with an Apache-2.0 LICENSE there. The current root README also promotes Qwen-UI-Agent, but that follow-up directory contains only documentation/display assets in the inspected tree. This entry does not claim the follow-up runtime, training system or weights are released.\n\nPrimary source: https://github.com/Tongyi-MAI/MAI-UI\nCatalog record: https://awesomecua.github.io/resources/github-tongyi-mai-mai-ui/"}
{"id": "github:tongyi-mai/mobileworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Tongyi-MAI/MobileWorld", "citation_url": "https://awesomecua.github.io/resources/github-tongyi-mai-mobileworld/", "details": {"availability": "Public task definitions, evaluator and container orchestration source.", "benchmark_protocol": {"actions": ["Android GUI operations", "MCP calls when enabled", "agent-user interaction when enabled"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "population": "README benchmark overview", "qualifier": "exact", "unit": "tasks", "value": 201}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "population": "README benchmark overview", "qualifier": "exact", "unit": "mobile applications", "value": 20}], "evaluation": "The publisher documents text-answer matching, backend database verification, local-storage inspection and application callbacks.", "evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md"], "focus": ["Android task completion", "agent-user interaction", "hybrid GUI and MCP use"], "notes": "Publisher-reported inventory. GUI-only, agent-user and MCP-augmented configurations have different dependencies and task eligibility; preserve enabled modes and step budget with results.", "observations": ["Android application interface", "user-agent interaction when enabled"], "setting": "interactive"}, "leaderboard_url": "https://tongyi-mai.github.io/MobileWorld/#leaderboard", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"], "reason": "Repository code license; self-hosted applications, model APIs and MCP services keep their own terms."}}, "repository_revision": "e7eee8981135ecbe39005a231edf63574defd79c"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:16Z", "role": "primary", "sha256": "fbc306b3414dfdb230171358efd6eb8d67573fbb94e11487a4a7e971e2abf0d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:12:15Z", "role": "primary", "sha256": "266e1fe5b5b1a4c1540ab0fd5eba206395084ee283f0079f424930a156cf8e18", "supports": ["identity"], "url": "https://api.github.com/repos/Tongyi-MAI/MobileWorld"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:08Z", "role": "primary", "sha256": "14500ed2dfa14bb5c34211090f0c8827e4a16c3233e35a2d695c0d7261f9358b", "supports": ["release_status"], "url": "https://api.github.com/repos/Tongyi-MAI/MobileWorld/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:09Z", "role": "primary", "sha256": "b09fcaf4fbc265223c2479c410beef2b98c5e687c78ecedf3452aa12ffc465f1", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T09:12:16Z", "licenses": {"code": {"basis": "Repository code license; self-hosted applications, model APIs and MCP services keep their own terms.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T09:14:09Z", "url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Tongyi-MAI/MobileWorld"}], "name": "MobileWorld (Tongyi-MAI)", "publication": null, "published_at": null, "record_sha256": "0f0cf7a288030baf2b426067cf234c556fdd55a3146829d877faa416a2267fc9", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md", "target_id": "arxiv:2512.19432", "type": "documented_by"}], "resource_id": "github:tongyi-mai/mobileworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Tongyi-MAI/MobileWorld", "topics": ["evaluation", "mobile-navigation", "human-in-the-loop", "tool-integration"], "updated_at": null}, "text": "MobileWorld (Tongyi-MAI)\nResource type: benchmark\nReview status: curated\nScope: core\n\nA containerized Android evaluation environment with GUI-only, agent-user interaction and MCP-augmented tasks, plus checks on application and backend state.\nPlatforms: android\nTopics: evaluation, mobile-navigation, human-in-the-loop, tool-integration\n\nRecorded specifications: {\"availability\": \"Public task definitions, evaluator and container orchestration source.\", \"benchmark_protocol\": {\"actions\": [\"Android GUI operations\", \"MCP calls when enabled\", \"agent-user interaction when enabled\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md\", \"population\": \"README benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 201}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md\", \"population\": \"README benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"mobile applications\", \"value\": 20}], \"evaluation\": \"The publisher documents text-answer matching, backend database verification, local-storage inspection and application callbacks.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/README.md\"], \"focus\": [\"Android task completion\", \"agent-user interaction\", \"hybrid GUI and MCP use\"], \"notes\": \"Publisher-reported inventory. GUI-only, agent-user and MCP-augmented configurations have different dependencies and task eligibility; preserve enabled modes and step budget with results.\", \"observations\": [\"Android application interface\", \"user-agent interaction when enabled\"], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://tongyi-mai.github.io/MobileWorld/#leaderboard\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/Tongyi-MAI/MobileWorld/HEAD/LICENSE\"], \"reason\": \"Repository code license; self-hosted applications, model APIs and MCP services keep their own terms.\"}}, \"repository_revision\": \"e7eee8981135ecbe39005a231edf63574defd79c\"}\n\nThis is Tongyi-MAI/MobileWorld, distinct from jacklishufan/MobileWorld (MobileWorldBench). Public source includes task definitions, verification logic and environment orchestration. The README reports 201 tasks across 20 applications; this is not an independent count of the evolving task tree. Linux/KVM or WSL2/KVM is required by the documented host setup, with macOS support in progress. MCP and simulated-user settings can require separate external services.\n\nPrimary source: https://github.com/Tongyi-MAI/MobileWorld\nCatalog record: https://awesomecua.github.io/resources/github-tongyi-mai-mobileworld/"}
{"id": "github:trycua/cua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/trycua/cua", "citation_url": "https://awesomecua.github.io/resources/github-trycua-cua/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "12830b2eeb3edea7af9953b27987f3467fa90803c641e786db09a33643c714d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "c0779290c1d4783169aa3dbfb55feb505e563ef8a004bbf55298ceffcfbda8d9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/LICENSE.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/trycua/cua/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/trycua/cua"}], "name": "Cua", "publication": null, "published_at": null, "record_sha256": "7ad468e7d38222dfc64f47c856d16db56d63a220bb157fc0bbd966c207138516", "relations": [], "resource_id": "github:trycua/cua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/trycua/cua", "topics": ["evaluation", "workflow"], "updated_at": null}, "text": "Cua\nResource type: framework\nReview status: curated\nScope: core\n\nMonorepo behind Cua Spaces, which gives agents full macOS or Linux desktops, plus Cua Driver for operating native apps, Lume for Apple Silicon VMs, Cua Bench and the CUA-S1 decision models.\nPlatforms: desktop\nTopics: evaluation, workflow\n\nPrimary source: https://github.com/trycua/cua\nCatalog record: https://awesomecua.github.io/resources/github-trycua-cua/"}
{"id": "github:trycua/lume", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/trycua/lume", "citation_url": "https://awesomecua.github.io/resources/github-trycua-lume/", "details": {}, "environments": ["desktop", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "12830b2eeb3edea7af9953b27987f3467fa90803c641e786db09a33643c714d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "c0779290c1d4783169aa3dbfb55feb505e563ef8a004bbf55298ceffcfbda8d9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/LICENSE.md"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/trycua/lume/HEAD/LICENSE.md"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/trycua/lume"}], "name": "Lume", "publication": null, "published_at": null, "record_sha256": "20d3096f0b8ef8432800987f0b834c2c03abb44134bebc21cdfb17dbd117f0f7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/trycua/lume/HEAD/README.md", "target_id": "github:trycua/cua", "type": "component_of"}], "resource_id": "github:trycua/lume", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/trycua/lume", "topics": ["sandbox"], "updated_at": null}, "text": "Lume\nResource type: infrastructure\nReview status: curated\nScope: core\n\nVirtual machine management for local computer-use environments.\nPlatforms: desktop, macos, linux\nTopics: sandbox\n\nLume is a component of the Cua repository. The legacy repository URL now serves the Cua monorepo README; counted as a component, not an independent project.\n\nPrimary source: https://github.com/trycua/lume\nCatalog record: https://awesomecua.github.io/resources/github-trycua-lume/"}
{"id": "github:tsinghua-fib-lab/fingertip-20k", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/tsinghua-fib-lab/FingerTip-20K", "citation_url": "https://awesomecua.github.io/resources/github-tsinghua-fib-lab-fingertip-20k/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "2c67c8b181e7196f3ca1e88888ffc5d6b32afca686a0e5c4353ad3f5af51ff9e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/tsinghua-fib-lab/FingerTip-20K/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/tsinghua-fib-lab/FingerTip-20K"}], "name": "FingerTip-20K", "publication": null, "published_at": null, "record_sha256": "9be6f0b6e1fbac3c5ff0bea96254cc64e63c4a3c865295fb2a34be402ffcc439", "relations": [], "resource_id": "github:tsinghua-fib-lab/fingertip-20k", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/tsinghua-fib-lab/FingerTip-20K", "topics": ["data-generation"], "updated_at": null}, "text": "FingerTip-20K\nResource type: dataset\nReview status: curated\nScope: core\n\nOver 20,000 human Android episodes across 500+ apps with user context and long-term usage patterns, plus tracks for proactive intent prediction and personalized execution.\nPlatforms: android\nTopics: data-generation\n\nPrimary source: https://github.com/tsinghua-fib-lab/FingerTip-20K\nCatalog record: https://awesomecua.github.io/resources/github-tsinghua-fib-lab-fingertip-20k/"}
{"id": "github:ubiquitouslearning/droidcall", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/UbiquitousLearning/DroidCall", "citation_url": "https://awesomecua.github.io/resources/github-ubiquitouslearning-droidcall/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "cfe9d8614b0b0558c37ba7c50371f6c47f119b914061ae815a9e3d796dd9ccfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:01:50Z", "role": "primary", "sha256": "7a4ca3daa6d011ff2d24d993191882d2b5545d8476c7eddf396581c0fef77496", "supports": ["identity"], "url": "https://api.github.com/repos/UbiquitousLearning/DroidCall"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "d0717d706359969376889eac73ea8764b659f41a2b4f3c1a409110267a912526", "supports": ["release_status"], "url": "https://api.github.com/repos/UbiquitousLearning/DroidCall/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:16Z", "role": "primary", "sha256": "2832cc5333f7166924a2823aa38a286d918b39b7fef2ca7af6521766c68a5303", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "08a6c5cfa655e1ecd774a41a9a815e845ad0afd47b289378bc1d6ee66c81dda9", "supports": ["relations"], "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:02:16Z", "role": "primary", "sha256": "2832cc5333f7166924a2823aa38a286d918b39b7fef2ca7af6521766c68a5303", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:02:06Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-17T02:02:16Z", "url": "https://raw.githubusercontent.com/UbiquitousLearning/DroidCall/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UbiquitousLearning/DroidCall"}], "name": "DroidCall generation and evaluation code", "publication": null, "published_at": null, "record_sha256": "6bc5216b69d2bd93c3639db5e30dbd422b3e88f148c0123cb1249d62cf498f1b", "relations": [{"evidence_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md", "target_id": "arxiv:2412.00402", "type": "documented_by"}], "resource_id": "github:ubiquitouslearning/droidcall", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://github.com/UbiquitousLearning/DroidCall", "topics": ["tool-integration", "data-generation", "evaluation"], "updated_at": null}, "text": "DroidCall generation and evaluation code\nResource type: framework\nReview status: curated\nScope: adjacent\n\nData-generation, fine-tuning and evaluation utilities for teaching small language models to invoke Android intents through predefined functions.\nPlatforms: android\nTopics: tool-integration, data-generation, evaluation\n\nSource utilities are published under MIT. This component operates through Android intent functions and is marked adjacent to visual GUI agents. Training data is separately released under its own card declaration; no checkpoint release is inferred from the fine-tuning instructions.\n\nPrimary source: https://github.com/UbiquitousLearning/DroidCall\nCatalog record: https://awesomecua.github.io/resources/github-ubiquitouslearning-droidcall/"}
{"id": "github:ucsb-ai/screen-point-and-read", "metadata": {"aliases": ["github:eric-ai-lab/screen-point-and-read", "eric-ai-lab/Screen-Point-and-Read", "https://github.com/eric-ai-lab/Screen-Point-and-Read"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/UCSB-AI/Screen-Point-and-Read", "citation_url": "https://awesomecua.github.io/resources/github-ucsb-ai-screen-point-and-read/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "7567287ceee0ad2d3d5605abb44ac50463c1008f9b6309393695763cf2ce89bb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "4748fb7d92f7a3a2a4a93353db96f6a2df547e5ea9ae87afa811c5423e96f363", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:01Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/eric-ai-lab/Screen-Point-and-Read"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-17T14:14:02Z", "role": "primary", "sha256": "868e18717ff76355c6552a5629b15de9fed1add1262fa5801ecabd7a7fb333d4", "supports": ["aliases"], "url": "https://api.github.com/repos/UCSB-AI/Screen-Point-and-Read"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/UCSB-AI/Screen-Point-and-Read/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UCSB-AI/Screen-Point-and-Read"}], "name": "Screen Point-and-Read / Tree-of-Lens", "publication": null, "published_at": null, "record_sha256": "5b4a1c8a60293e061734bb661c184eb319a7f41e43256c305f39c9ae5c82e202", "relations": [{"evidence_url": "https://raw.githubusercontent.com/eric-ai-lab/Screen-Point-and-Read/HEAD/README.md", "target_id": "arxiv:2406.19263", "type": "documented_by"}], "resource_id": "github:ucsb-ai/screen-point-and-read", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/UCSB-AI/Screen-Point-and-Read", "topics": ["screen-understanding", "grounding", "evaluation"], "updated_at": null}, "text": "Screen Point-and-Read / Tree-of-Lens\nResource type: benchmark\nReview status: curated\nScope: core\n\nScreen-reading benchmark and Tree-of-Lens implementation for interpreting user-indicated regions and their layout context.\nRecorded aliases: github:eric-ai-lab/screen-point-and-read, eric-ai-lab/Screen-Point-and-Read, https://github.com/eric-ai-lab/Screen-Point-and-Read\nPlatforms: cross-platform\nTopics: screen-understanding, grounding, evaluation\n\nThe canonical repository now belongs to UCSB-AI. Its README lists additional setup documentation and evaluation scripts as unfinished.\n\nPrimary source: https://github.com/UCSB-AI/Screen-Point-and-Read\nCatalog record: https://awesomecua.github.io/resources/github-ucsb-ai-screen-point-and-read/"}
{"id": "github:uitron-hub/treecua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/UITron-hub/TreeCUA", "citation_url": "https://awesomecua.github.io/resources/github-uitron-hub-treecua/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "91cffd585c667d169486506a785c6cbc3068fe28409efc97abf8911f9d0bd5a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UITron-hub/TreeCUA"}], "name": "TreeCUA", "publication": null, "published_at": null, "record_sha256": "177eeb6882a18813c242352166f46e78ed9e6f1973a9889bd5911e3385394221", "relations": [{"evidence_url": "https://raw.githubusercontent.com/UITron-hub/TreeCUA/HEAD/README.md", "target_id": "arxiv:2602.09662", "type": "documented_by"}], "resource_id": "github:uitron-hub/treecua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/UITron-hub/TreeCUA", "topics": ["data-generation"], "updated_at": null}, "text": "TreeCUA\nResource type: framework\nReview status: curated\nScope: core\n\nFramework for generating GUI trajectories through tree-structured exploration and verification.\nPlatforms: desktop\nTopics: data-generation\n\nPrimary source: https://github.com/UITron-hub/TreeCUA\nCatalog record: https://awesomecua.github.io/resources/github-uitron-hub-treecua/"}
{"id": "github:unipat-ai/saas-bench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/UniPat-AI/SaaS-Bench", "citation_url": "https://awesomecua.github.io/resources/github-unipat-ai-saas-bench/", "details": {"leaderboard_url": "https://unipat.ai/benchmarks/SaaS-Bench"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "8cfab6fca80090adb47c4ef634f89eedd49a6204dbdc2b5b8e0d6f04c32db289", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "971246559df3f118e367cb299d3544f94e7c3d30db39f05c0c164589af4f3a88", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/UniPat-AI/SaaS-Bench"}], "name": "SaaS-Bench", "publication": null, "published_at": null, "record_sha256": "7c91583ef53c52ef8d8065cdab3b8cba8c55b2ab2b371ff518b4021bc75b80db", "relations": [{"evidence_url": "https://raw.githubusercontent.com/UniPat-AI/SaaS-Bench/HEAD/README.md", "target_id": "arxiv:2605.15777", "type": "documented_by"}], "resource_id": "github:unipat-ai/saas-bench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/UniPat-AI/SaaS-Bench", "topics": ["evaluation", "workflow"], "updated_at": null}, "text": "SaaS-Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\n106 browser-driven business workflows across 23 self-hosted SaaS apps in six domains, scored by per-task verifiers that read each app's database and give partial credit.\nPlatforms: web\nTopics: evaluation, workflow\n\nRecorded specifications: {\"leaderboard_url\": \"https://unipat.ai/benchmarks/SaaS-Bench\"}\n\nPrimary source: https://github.com/UniPat-AI/SaaS-Bench\nCatalog record: https://awesomecua.github.io/resources/github-unipat-ai-saas-bench/"}
{"id": "github:vercel-labs/agent-browser", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/vercel-labs/agent-browser", "citation_url": "https://awesomecua.github.io/resources/github-vercel-labs-agent-browser/", "details": {"availability": "Published source; official release metadata inspected", "release_tag": "v0.39.1"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "4a9783ee8ebee10caaedfe2bdd2182539368e60e8eefadaee4cb7bab7fc75c12", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6b1120e122d5d110b0da568fbbf0560fd81a4cf8d50602c397a8382617343329", "supports": ["identity"], "url": "https://api.github.com/repos/vercel-labs/agent-browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:53Z", "role": "primary", "sha256": "6a90f0b79089f895fab5d35443fd113faa056f404c807966b23210c1c427e62e", "supports": ["release_status"], "url": "https://api.github.com/repos/vercel-labs/agent-browser/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:55Z", "role": "primary", "sha256": "1a4213f48240031c7837e9651bda8cebb84d69d0d97a65f012082a53a7a1f702", "supports": ["release_status"], "url": "https://api.github.com/repos/vercel-labs/agent-browser/releases/latest"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "014bb31e83d5c2e76aea1cc6e82217346ab41362f32cb355ad0f5c10aa0aeaff", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T04:00:56Z", "role": "primary", "sha256": "014bb31e83d5c2e76aea1cc6e82217346ab41362f32cb355ad0f5c10aa0aeaff", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-17T03:59:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T04:00:56Z", "url": "https://raw.githubusercontent.com/vercel-labs/agent-browser/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vercel-labs/agent-browser"}], "name": "agent-browser", "publication": null, "published_at": null, "record_sha256": "28fc9284820edbb0c6af337050bdf2571edda83fa3a140c79feadb2d642312ab", "relations": [], "resource_id": "github:vercel-labs/agent-browser", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/vercel-labs/agent-browser", "topics": ["web-navigation", "automation", "tool-integration"], "updated_at": null}, "text": "agent-browser\nResource type: tool\nReview status: curated\nScope: core\n\nA browser-automation command-line interface that exposes browser state and interaction commands for AI agents.\nPlatforms: web\nTopics: web-navigation, automation, tool-integration\n\nRecorded specifications: {\"availability\": \"Published source; official release metadata inspected\", \"release_tag\": \"v0.39.1\"}\n\nThe repository publishes a Rust-based CLI and platform-specific release binaries. The inspected release is v0.39.1. Apache-2.0 is declared for the code; Chrome and other external browser components retain separate terms. No package or browser was installed.\n\nPrimary source: https://github.com/vercel-labs/agent-browser\nCatalog record: https://awesomecua.github.io/resources/github-vercel-labs-agent-browser/"}
{"id": "github:verigui-team/verigui", "metadata": {"aliases": ["VeriGUI", "VeriWeb"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/VeriGUI-Team/VeriGUI", "citation_url": "https://awesomecua.github.io/resources/github-verigui-team-verigui/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "e9a0866c8d9fa1759dedebd5ea3a4176009e51eb669f5877540524e53ac1077d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/VeriGUI-Team/VeriGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VeriGUI-Team/VeriGUI"}], "name": "VeriWeb (formerly VeriGUI)", "publication": null, "published_at": null, "record_sha256": "e4452b65abda3930c32ed052305c824f2453c3fb22156e84ecf8e43e9c9b318f", "relations": [], "resource_id": "github:verigui-team/verigui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/VeriGUI-Team/VeriGUI", "topics": ["evaluation"], "updated_at": null}, "text": "VeriWeb (formerly VeriGUI)\nResource type: benchmark\nReview status: curated\nScope: core\n\nVerifiable web information-seeking benchmark renamed from VeriGUI; desktop evaluation is described as future work.\nRecorded aliases: VeriGUI, VeriWeb\nPlatforms: web\nTopics: evaluation\n\nPrimary source: https://github.com/VeriGUI-Team/VeriGUI\nCatalog record: https://awesomecua.github.io/resources/github-verigui-team-verigui/"}
{"id": "github:verisafeagent/verisafeagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/VeriSafeAgent/VeriSafeAgent", "citation_url": "https://awesomecua.github.io/resources/github-verisafeagent-verisafeagent/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "2403ef3ec300e257dc631e3c569fc9aabaab1d3f385ac37abe1e4021aaae6e21", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VeriSafeAgent/VeriSafeAgent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T13:06:20.028238+00:00", "role": "primary", "sha256": "9b04efe4136a39e3e5f21a1d3bb8a44075f1370bdbeec6a4307783f3836f8311", "supports": ["summary", "relations", "notes"], "url": "https://arxiv.org/html/2503.18492v2"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VeriSafeAgent/VeriSafeAgent"}], "name": "VeriSafe Agent", "publication": null, "published_at": null, "record_sha256": "a83ef097c35a6b371c27451e77d8f671cd8757b5fbcd5b67d182d0120384e44a", "relations": [{"evidence_url": "https://arxiv.org/html/2503.18492v2", "target_id": "arxiv:2503.18492", "type": "documented_by"}], "resource_id": "github:verisafeagent/verisafeagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/VeriSafeAgent/VeriSafeAgent", "topics": ["safety-security", "mobile-navigation", "state-tracking"], "updated_at": null}, "text": "VeriSafe Agent\nResource type: framework\nReview status: curated\nScope: core\n\nRuntime verification framework for Android GUI agents that checks proposed actions against logical rules derived from user instructions and developer-defined app states.\nPlatforms: android\nTopics: safety-security, mobile-navigation, state-tracking\n\nThe verification method requires developers to supply app-state predicates and transition hooks. The paper-to-repository relation is supported by the exact-version paper, not inferred from the repository name.\n\nPrimary source: https://github.com/VeriSafeAgent/VeriSafeAgent\nCatalog record: https://awesomecua.github.io/resources/github-verisafeagent-verisafeagent/"}
{"id": "github:visualwebbench/visualwebbench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/VisualWebBench/VisualWebBench", "citation_url": "https://awesomecua.github.io/resources/github-visualwebbench-visualwebbench/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "262947081ddd17dbf5bb3d072574a19bbf30510c42c3ff42c9851a748fd37546", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/VisualWebBench/VisualWebBench/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/VisualWebBench/VisualWebBench"}], "name": "VisualWebBench", "publication": null, "published_at": null, "record_sha256": "0958a077bf70693dc8ca802f67aceed3dc415e30e9efe728e997800ec3c4b33c", "relations": [{"evidence_url": "https://raw.githubusercontent.com/VisualWebBench/VisualWebBench/HEAD/README.md", "target_id": "arxiv:2404.05955", "type": "documented_by"}], "resource_id": "github:visualwebbench/visualwebbench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/VisualWebBench/VisualWebBench", "topics": ["evaluation", "screen-understanding"], "updated_at": null}, "text": "VisualWebBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nEvaluation of multimodal models on visual web understanding tasks.\nPlatforms: web\nTopics: evaluation, screen-understanding\n\nPrimary source: https://github.com/VisualWebBench/VisualWebBench\nCatalog record: https://awesomecua.github.io/resources/github-visualwebbench-visualwebbench/"}
{"id": "github:vivo-ai-lab/bluelm-gui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/vivo-ai-lab/BlueLM-GUI", "citation_url": "https://awesomecua.github.io/resources/github-vivo-ai-lab-bluelm-gui/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "5f460f9b2a18315d07b8c48d78dcc764892d09d9250e95f5e591b5966fe59328", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:41Z", "role": "primary", "sha256": "6036aa17e52be8051fc3c8ede37fc4d0b2bc2589518be7623566c75868612178", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:41Z", "url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vivo-ai-lab/BlueLM-GUI"}], "name": "BlueLM-GUI", "publication": null, "published_at": null, "record_sha256": "4dec3e05e7855a9605c233390fbc1214d883e67fbf910ae5bc0046398d067cc7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/vivo-ai-lab/BlueLM-GUI/HEAD/README.md", "target_id": "arxiv:2609.12394", "type": "documented_by"}], "resource_id": "github:vivo-ai-lab/bluelm-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/vivo-ai-lab/BlueLM-GUI", "topics": ["mobile-navigation", "reinforcement-learning"], "updated_at": null}, "text": "BlueLM-GUI\nResource type: model\nReview status: curated\nScope: core\n\nMobile GUI agent research release with technical report, benchmark data, and evaluation tooling from vivo AI Lab.\nPlatforms: android\nTopics: mobile-navigation, reinforcement-learning\n\nThe repository lists the agent as in-house. Public model weights were not established by this review; linked benchmark assets have their own access terms.\n\nPrimary source: https://github.com/vivo-ai-lab/BlueLM-GUI\nCatalog record: https://awesomecua.github.io/resources/github-vivo-ai-lab-bluelm-gui/"}
{"id": "github:vivo/dimo-gui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/vivo/DiMo-GUI", "citation_url": "https://awesomecua.github.io/resources/github-vivo-dimo-gui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4f578ecd5cbba12797586892a9190e480291bd267a180f0ce56847764eaac524", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:01Z", "role": "primary", "sha256": "d1a708ddcd4afc4e40a5d66f2d56835a12d815515ab6a61578e541ed2c0483bb", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:01Z", "url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/vivo/DiMo-GUI"}], "name": "DiMo-GUI", "publication": null, "published_at": null, "record_sha256": "937b300cf92998b04ba11141bad5202f0bdfc4e006b1e35558f9d243b18632f4", "relations": [{"evidence_url": "https://raw.githubusercontent.com/vivo/DiMo-GUI/HEAD/readme.md", "target_id": "arxiv:2507.00008", "type": "documented_by"}], "resource_id": "github:vivo/dimo-gui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/vivo/DiMo-GUI", "topics": ["grounding", "efficiency"], "updated_at": null}, "text": "DiMo-GUI\nResource type: model\nReview status: curated\nScope: core\n\nGUI grounding research using modality-aware visual reasoning and test-time scaling.\nPlatforms: cross-platform\nTopics: grounding, efficiency\n\nPrimary source: https://github.com/vivo/DiMo-GUI\nCatalog record: https://awesomecua.github.io/resources/github-vivo-dimo-gui/"}
{"id": "github:web-arena-x/visualwebarena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/web-arena-x/visualwebarena", "citation_url": "https://awesomecua.github.io/resources/github-web-arena-x-visualwebarena/", "details": {"benchmark_protocol": {"actions": ["browser element actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "population": "full VWA suite", "qualifier": "exact", "unit": "tasks", "value": 910}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "population": "one task from each template type", "qualifier": "exact", "unit": "human trajectories", "value": 233}], "evaluation": "Execution-based completion checks for visually grounded web tasks.", "evidence_urls": ["https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"], "focus": ["multimodal web task completion"], "notes": "The 233 demonstration trajectories are not the full 910-task evaluation set. Observation modes and captioning components differ across baseline agents.", "observations": ["screenshots with set-of-mark annotations", "caption-augmented accessibility tree"], "setting": "interactive"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "182529c2fce1e41f3a2162727205c54da8b3d2381b2273d726170ee9d8cf65cc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-arena-x/visualwebarena"}], "name": "VisualWebArena", "publication": null, "published_at": null, "record_sha256": "55b359a66b21e2ec07ba71a87a0ec1fb76ae183f61ae523921390f54596f38d9", "relations": [{"evidence_url": "https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md", "target_id": "arxiv:2401.13649", "type": "documented_by"}], "resource_id": "github:web-arena-x/visualwebarena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/web-arena-x/visualwebarena", "topics": ["evaluation", "screen-understanding"], "updated_at": null}, "text": "VisualWebArena\nResource type: benchmark\nReview status: curated\nScope: core\n\nVisually grounded web tasks extending the WebArena evaluation setting.\nPlatforms: web\nTopics: evaluation, screen-understanding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"browser element actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md\", \"population\": \"full VWA suite\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 910}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md\", \"population\": \"one task from each template type\", \"qualifier\": \"exact\", \"unit\": \"human trajectories\", \"value\": 233}], \"evaluation\": \"Execution-based completion checks for visually grounded web tasks.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/web-arena-x/visualwebarena/HEAD/README.md\"], \"focus\": [\"multimodal web task completion\"], \"notes\": \"The 233 demonstration trajectories are not the full 910-task evaluation set. Observation modes and captioning components differ across baseline agents.\", \"observations\": [\"screenshots with set-of-mark annotations\", \"caption-augmented accessibility tree\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/web-arena-x/visualwebarena\nCatalog record: https://awesomecua.github.io/resources/github-web-arena-x-visualwebarena/"}
{"id": "github:web-arena-x/webarena", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/web-arena-x/webarena", "citation_url": "https://awesomecua.github.io/resources/github-web-arena-x-webarena/", "details": {"benchmark_protocol": {"actions": ["browser element actions"], "counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "population": "canonical WebArena evaluation suite", "qualifier": "exact", "unit": "task examples", "value": 812}], "evaluation": "Functional correctness of completed web tasks.", "evidence_urls": ["https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"], "focus": ["web task completion"], "notes": "The canonical repository documents self-hosted websites and environment resets. Its README recommends BrowserGym/AgentLab for newer infrastructure; harness changes must be recorded.", "observations": ["HTML or accessibility tree"], "setting": "interactive"}, "leaderboard_url": "https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?usp=sharing"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "081c826497f9b7da4952771ee128c2215ec3532379db0fe9f7fcbd899d9c9287", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:20Z", "role": "primary", "sha256": "cbe8bfcbce4170a94a6cf42de01fbe0a448e7b1d29fa75a0de877105a98ec08e", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-arena-x/webarena"}], "name": "WebArena", "publication": null, "published_at": null, "record_sha256": "d0f6fe8e0898a8fc6145b3e136ef83ec80bcc6c900f2c9cea951eae5645e2a61", "relations": [{"evidence_url": "https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md", "target_id": "arxiv:2307.13854", "type": "documented_by"}], "resource_id": "github:web-arena-x/webarena", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/web-arena-x/webarena", "topics": ["evaluation", "web-navigation"], "updated_at": null}, "text": "WebArena\nResource type: benchmark\nReview status: curated\nScope: core\n\nSelf-hosted web environments and tasks for evaluating autonomous web agents.\nPlatforms: web\nTopics: evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"browser element actions\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md\", \"population\": \"canonical WebArena evaluation suite\", \"qualifier\": \"exact\", \"unit\": \"task examples\", \"value\": 812}], \"evaluation\": \"Functional correctness of completed web tasks.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/web-arena-x/webarena/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=1802.08802%2C1808.09132%2C2005.03776%2C2101.09465%2C2103.16057%2C2104.08560%2C2105.13231%2C2107.03374%2C2107.13731%2C2108.03353%2C2202.02312%2C2202.08137%2C2204.01691%2C2205.11029%2C2207.01206%2C2209.08199%2C2209.08655%2C2209.14927%2C2209.15099%2C2210.02663%2C2210.03629%2C2210.03945%2C2211.07615%2C2301.10165%2C2301.13280%2C2302.04761%2C2303.07280%2C2303.08774%2C2303.17491%2C2303.17580%2C2304.04370%2C2304.07061%2C2305.04091%2C2305.05662%2C2305.08144%2C2305.10601%2C2305.11854%2C2305.12487%2C2305.14257%2C2305.14325%2C2305.16504%2C2305.16653%2C2305.19308%2C2306.00245%2C2306.06070%2C2306.07863%2C2307.10088%2C2307.12856%2C2307.13854%2C2308.01990&max_results=50\"], \"focus\": [\"web task completion\"], \"notes\": \"The canonical repository documents self-hosted websites and environment resets. Its README recommends BrowserGym/AgentLab for newer infrastructure; harness changes must be recorded.\", \"observations\": [\"HTML or accessibility tree\"], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?usp=sharing\"}\n\nPrimary source: https://github.com/web-arena-x/webarena\nCatalog record: https://awesomecua.github.io/resources/github-web-arena-x-webarena/"}
{"id": "github:web-infra-dev/midscene", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/web-infra-dev/midscene", "citation_url": "https://awesomecua.github.io/resources/github-web-infra-dev-midscene/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "8cd83b12f4e1c1b22c27b6ae3ee5415564e61ff4072349ec512982a548ea6ec3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "9d26ef3ec511840e10eeaaf1e1b211e196c838823baf6bf9033761f27ab349ce", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/web-infra-dev/midscene/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/web-infra-dev/midscene"}], "name": "Midscene.js", "publication": null, "published_at": null, "record_sha256": "187068f8217d2937935d7288a855797da6555d03b17ef764fd16308ee0b9efe4", "relations": [], "resource_id": "github:web-infra-dev/midscene", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/web-infra-dev/midscene", "topics": ["automation", "grounding"], "updated_at": null}, "text": "Midscene.js\nResource type: framework\nReview status: curated\nScope: core\n\nVisual automation toolkit for web, mobile, and desktop interfaces.\nPlatforms: cross-platform\nTopics: automation, grounding\n\nPrimary source: https://github.com/web-infra-dev/midscene\nCatalog record: https://awesomecua.github.io/resources/github-web-infra-dev-midscene/"}
{"id": "github:webvln/webvln", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/WebVLN/WebVLN", "citation_url": "https://awesomecua.github.io/resources/github-webvln-webvln/", "details": {"availability": "Public WebVLN training/evaluation and environment-wrapper source; external data/models unverified.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"], "reason": "README MIT badge is not backed by a project license file or component-scoped statement in the inspected release."}}, "repository_revision": "59aba284f5888fb7730e1ff07a7e682a1bd1968a"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "cb98feab09648856ab30d8eb739a29e837673cc5b4a6642b7ba3a8943ad71053", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.code"], "url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "112e67cc9f865b9640da3827e32eaba5a50f2feb8d85c20018dccbcafa6fd3e5", "supports": ["identity"], "url": "https://api.github.com/repos/WebVLN/WebVLN"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:05Z", "role": "primary", "sha256": "966adf983bae6052c271f7153c30776a29f7272cf55d05e0175713f02d1900fb", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/WebVLN/WebVLN"}], "name": "WebVLN research implementation", "publication": null, "published_at": null, "record_sha256": "2355587b69accd6c5422ab2b2ea4201c981f242ab512800121f872cb5de28206", "relations": [{"evidence_url": "https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md", "target_id": "arxiv:2312.15820", "type": "documented_by"}], "resource_id": "github:webvln/webvln", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/WebVLN/WebVLN", "topics": ["web-navigation", "grounding", "evaluation"], "updated_at": null}, "text": "WebVLN research implementation\nResource type: framework\nReview status: curated\nScope: core\n\nTraining and evaluation code for question-guided website navigation that combines visual observations with web-specific HTML information.\nPlatforms: web\nTopics: web-navigation, grounding, evaluation\n\nRecorded specifications: {\"availability\": \"Public WebVLN training/evaluation and environment-wrapper source; external data/models unverified.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/WebVLN/WebVLN/HEAD/README.md\", \"https://api.github.com/repos/WebVLN/WebVLN/git/trees/main?recursive=1\"], \"reason\": \"README MIT badge is not backed by a project license file or component-scoped statement in the inspected release.\"}}, \"repository_revision\": \"59aba284f5888fb7730e1ff07a7e682a1bd1968a\"}\n\nThe released tree contains navigation models, an environment wrapper, scoring utilities and train/eval scripts. Full WebVLN-v1 data and pretrained files are linked through an external Drive folder; their complete contents and data/weight terms were not verified, so no separate benchmark download or checkpoint is approved here. The README MIT badge has no matching repository license file or explicit component scope; code permissions remain unestablished.\n\nPrimary source: https://github.com/WebVLN/WebVLN\nCatalog record: https://awesomecua.github.io/resources/github-webvln-webvln/"}
{"id": "github:weizhepei/webagent-r1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/weizhepei/WebAgent-R1", "citation_url": "https://awesomecua.github.io/resources/github-weizhepei-webagent-r1/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "813b7580cdae140917db573522b83a7986448025b31064442d34757292a6935c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/WebAgent-R1/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:19Z", "role": "primary", "sha256": "db1f71491c1d9482f403db30ea5599bd76e93d1db02c4ba8134d1a5cae5f0487", "supports": ["identity"], "url": "https://api.github.com/repos/weizhepei/WebAgent-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:39Z", "role": "primary", "sha256": "cc1a74037c4a2c4c094b11f6874ef67f88b24d14293ec09568ac6f378035c0b0", "supports": ["release_status"], "url": "https://api.github.com/repos/weizhepei/WebAgent-R1/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}, {"method": "license_header", "retrieved_at": "2026-10-17T02:05:22Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T02:05:22Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T02:05:22Z", "url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/weizhepei/WebAgent-R1"}], "name": "WebAgent-R1", "publication": null, "published_at": null, "record_sha256": "4a55f7d95cfd85633afc71a924913e10024d74b1ce73eeb01d4b8653de8f1e6d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/weizhepei/WebAgent-R1/HEAD/WebAgent-R1/README.md", "target_id": "arxiv:2505.16421", "type": "documented_by"}], "resource_id": "github:weizhepei/webagent-r1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/weizhepei/WebAgent-R1", "topics": ["reinforcement-learning", "web-navigation", "evaluation"], "updated_at": null}, "text": "WebAgent-R1\nResource type: framework\nReview status: curated\nScope: core\n\nEnvironment setup, multi-turn reinforcement-learning and evaluation tooling for browser agents interacting with WebArena.\nPlatforms: web\nTopics: reinforcement-learning, web-navigation, evaluation\n\nThe implementation is inside the WebAgent-R1 subdirectory and includes training, evaluation and WebArena setup components. Root source licensing is Apache-2.0. No downloadable author checkpoint was established by the reviewed README, so this entry represents code.\n\nPrimary source: https://github.com/weizhepei/WebAgent-R1\nCatalog record: https://awesomecua.github.io/resources/github-weizhepei-webagent-r1/"}
{"id": "github:westlake-agi-lab/appagentx", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Westlake-AGI-Lab/AppAgentX", "citation_url": "https://awesomecua.github.io/resources/github-westlake-agi-lab-appagentx/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "085f1a1810805bcade9555b1e5c374a7db1b8b1377afeeea285d927853b02f3d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Westlake-AGI-Lab/AppAgentX/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Westlake-AGI-Lab/AppAgentX"}], "name": "AppAgentX", "publication": null, "published_at": null, "record_sha256": "5ac57aeedf5174e23ea9819f44b8f3cf26af8b2dde4fc8ab88e0a9d4ac568151", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Westlake-AGI-Lab/AppAgentX/HEAD/README.md", "target_id": "arxiv:2503.02268", "type": "documented_by"}], "resource_id": "github:westlake-agi-lab/appagentx", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Westlake-AGI-Lab/AppAgentX", "topics": ["mobile-navigation", "memory"], "updated_at": null}, "text": "AppAgentX\nResource type: agent\nReview status: curated\nScope: core\n\nMobile agent research using accumulated interaction experience.\nPlatforms: android\nTopics: mobile-navigation, memory\n\nPrimary source: https://github.com/Westlake-AGI-Lab/AppAgentX\nCatalog record: https://awesomecua.github.io/resources/github-westlake-agi-lab-appagentx/"}
{"id": "github:wlanderson0/orbit", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/wlanderson0/orbit", "citation_url": "https://awesomecua.github.io/resources/github-wlanderson0-orbit/", "details": {"license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE", "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/NOTICE"], "reason": "Record only the publisher's top-level Apache-2.0 declaration for Orbit-original code. NOTICE explicitly preserves per-component terms; neither bundled or external data nor model weights inherit this declaration. Its account of BrowserART licensing is not an independent resolution of upstream conflicting declarations."}}}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:22.467183+00:00", "role": "primary", "sha256": "b6743e3e9e5256d7afa8bb411baf246b60ba71bcfb84eda5a816fba3ddd7d78b", "supports": ["name", "kind", "summary", "topics", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/README.md"}, {"method": "github_repository_identity", "retrieved_at": "2026-10-20T13:39:57.052202+00:00", "role": "primary", "sha256": "0436414b9bcc4307f3bcde476a544da6fe1dd5baa90939eed392942490fbfbf7", "supports": ["id", "url"], "url": "https://api.github.com/repos/wlanderson0/orbit"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:37.098639+00:00", "role": "primary", "sha256": "c47d046b129999da5e63b7723a3f3e2e0cc31c27ce74b45f71f8c6d0df1fc0d8", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/browser/browserart/tools.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:10:32.910533Z", "role": "primary", "sha256": "2cce76b22252a415d64c356b2062771e528b554dd83b6f88fca769f48fa8a826", "supports": ["summary", "environments", "scope", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/desktop/osworld/tools.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:09:35.060985Z", "role": "primary", "sha256": "66515d7ee87eab5a2f42c26132d0bf3e2a3b9c006156d2378181cf91966a2903", "supports": ["notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/docs/benchmark-qualification.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:10:29.541892Z", "role": "primary", "sha256": "94c781028dc80b2a748d759fae2456f6a986da7487651a6bcbb179e8ad4fe023", "supports": ["notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/orbit/scenarios/browser/browserart/scorer.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:29.036076+00:00", "role": "primary", "sha256": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T14:09:31.550310Z", "role": "primary", "sha256": "f68a64f775b1f329416f2fe294bb07b5e475893f77d1cdb5300bdb08acf95410", "supports": ["licenses.code", "licenses.data", "notes"], "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/NOTICE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:12:28.833839+00:00", "role": "primary", "sha256": "3ea667f4187ba601011e5ff00c4d63676b6d9903d12b80a0d7d5dd9920f00c65", "supports": ["relations"], "url": "https://arxiv.org/html/2609.33102v1"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-20T13:41:22.467183+00:00", "licenses": {"code": {"basis": "Complete root Apache License 2.0 text for Orbit-original code. NOTICE preserves separate third-party component terms, including MIT-vendored modules; this is not a uniform license for the entire dependency/data bundle.", "declared": "Apache-2.0", "retrieved_at": "2026-10-20T13:41:29.036076+00:00", "url": "https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/wlanderson0/orbit"}], "name": "ORBIT framework", "publication": null, "published_at": null, "record_sha256": "fb38dd97a3cf9d47fbbb6131b92410583c8852b7c6d6ead10e8a9c22bcd5b54f", "relations": [{"evidence_url": "https://arxiv.org/html/2609.33102v1", "target_id": "arxiv:2609.33102", "type": "documented_by"}], "resource_id": "github:wlanderson0/orbit", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/wlanderson0/orbit", "topics": ["evaluation", "safety-security", "web-navigation", "observability"], "updated_at": null}, "text": "ORBIT framework\nResource type: framework\nReview status: curated\nScope: core\n\nInspect-based multi-agent evaluation framework with role, scheduling and communication controls, BrowserART accessibility-tree actions, and desktop mouse, keyboard and screenshot tools.\nPlatforms: web, desktop\nTopics: evaluation, safety-security, web-navigation, observability\n\nRecorded specifications: {\"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/LICENSE\", \"https://raw.githubusercontent.com/wlanderson0/orbit/588b3035f92450ec4f0496bb8b629daa205e84c7/NOTICE\"], \"reason\": \"Record only the publisher's top-level Apache-2.0 declaration for Orbit-original code. NOTICE explicitly preserves per-component terms; neither bundled or external data nor model weights inherit this declaration. Its account of BrowserART licensing is not an independent resolution of upstream conflicting declarations.\"}}}\n\nReviewed version 1.0.3 source at commit 588b3035, not an asserted reproduction of the paper. BrowserART tools return accessibility trees, including the screenshot-named refresh. Qualification is scenario-specific: the project documents browser-judge and desktop-scoring limits; its opt-in exact-send oracle covers one derived local task, not SMTP delivery. Runtime profiles and scoring policies are separate. Apache-2.0 is the top-level Orbit code declaration; third-party components and external data/environments retain separate terms. No installation, release archive or live evaluation was tested.\n\nPrimary source: https://github.com/wlanderson0/orbit\nCatalog record: https://awesomecua.github.io/resources/github-wlanderson0-orbit/"}
{"id": "github:wuklab/osworld-human", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/WukLab/osworld-human", "citation_url": "https://awesomecua.github.io/resources/github-wuklab-osworld-human/", "details": {"benchmark_protocol": {"actions": ["efficiency scores"], "counts": [], "evaluation": "The scorer consumes OSWorld results and a specified maximum step budget; the README distinguishes single-action and grouped-action WES+ and WES-.", "evidence_urls": ["https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md"], "focus": ["action efficiency relative to humans"], "notes": "Reference solutions are published for evaluation and expressly discouraged for training. No task count is inferred from file count.", "observations": ["OSWorld result trajectories", "human reference actions"], "setting": "offline"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "35c2f4dbed9d9bfc57f4c5a3f2fd70d313d6b463403b482c2fd4066d27b02cc1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "21c86601da6e77a3326fbcd143e3c5d9a033cd59e6e13973250966dff6f51ae9", "supports": ["identity"], "url": "https://api.github.com/repos/WukLab/osworld-human"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:55Z", "role": "primary", "sha256": "c7b958d7a075adf5d282be08b68ad2b090568da342d4cb4b501ae808e54d0053", "supports": ["release_status"], "url": "https://api.github.com/repos/WukLab/osworld-human/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T03:59:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/WukLab/osworld-human"}], "name": "OSWorld-Human", "publication": null, "published_at": null, "record_sha256": "bac434f0929757e85440eca1a86920065545028a69322e277738f5df64158d51", "relations": [{"evidence_url": "https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md", "target_id": "arxiv:2506.16042", "type": "documented_by"}], "resource_id": "github:wuklab/osworld-human", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/WukLab/osworld-human", "topics": ["evaluation", "efficiency", "workflow"], "updated_at": null}, "text": "OSWorld-Human\nResource type: benchmark\nReview status: curated\nScope: core\n\nHuman reference trajectories and scoring code for measuring the action efficiency of computer-use agents on OSWorld tasks.\nPlatforms: desktop\nTopics: evaluation, efficiency, workflow\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"efficiency scores\"], \"counts\": [], \"evaluation\": \"The scorer consumes OSWorld results and a specified maximum step budget; the README distinguishes single-action and grouped-action WES+ and WES-.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/WukLab/osworld-human/HEAD/README.md\"], \"focus\": [\"action efficiency relative to humans\"], \"notes\": \"Reference solutions are published for evaluation and expressly discouraged for training. No task count is inferred from file count.\", \"observations\": [\"OSWorld result trajectories\", \"human reference actions\"], \"setting\": \"offline\"}}\n\nThe repository publishes per-application reference JSON files and score.py. It explicitly states that these evaluation solutions should not be used for training. No root code or data-license declaration was found; no license is inherited from OSWorld, and the score script was not executed.\n\nPrimary source: https://github.com/WukLab/osworld-human\nCatalog record: https://awesomecua.github.io/resources/github-wuklab-osworld-human/"}
{"id": "github:wuzheng02/gui-cider", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Wuzheng02/GUI-CIDER", "citation_url": "https://awesomecua.github.io/resources/github-wuzheng02-gui-cider/", "details": {"availability": "Public synthesis and evaluation source; model checkpoints not verified.", "repository_revision": "1ec25af942a46ce46f6c00d596332e308198342f"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "9a46741bbde1373ed4b87887cdadde98e05ea8795d34efe39fc25d69b4d50213", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "ae1d5687a72a39eacbf4fe281b5048914e98adf7cc1517481a5acc8a56a81446", "supports": ["identity"], "url": "https://api.github.com/repos/Wuzheng02/GUI-CIDER"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:59Z", "role": "primary", "sha256": "355128b62a281074eebc14a1e3b85a260f06c2bf3c1a6ed140cba3b9aefaa76c", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/Wuzheng02/GUI-CIDER/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:24:47Z", "role": "primary", "sha256": "bffca456e649a08239f6a1a3be10a7d7c21bcb427e783ba2c304e4b8410d385a", "supports": ["environments", "summary", "notes"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T11:18:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/GUI-CIDER"}], "name": "GUI-CIDER research code", "publication": null, "published_at": null, "record_sha256": "6038535773fcfa4a35f82ed4071c4e31d1548b4ece7307c96e9ff1ae3063071d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "arxiv:2605.28534", "type": "documented_by"}], "resource_id": "github:wuzheng02/gui-cider", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Wuzheng02/GUI-CIDER", "topics": ["data-generation", "imitation-learning", "planning"], "updated_at": null}, "text": "GUI-CIDER research code\nResource type: framework\nReview status: curated\nScope: core\n\nReleased data-synthesis and evaluation scripts that distill planning and causal GUI knowledge into text for mid-training mobile interface agents.\nPlatforms: android\nTopics: data-generation, imitation-learning, planning\n\nRecorded specifications: {\"availability\": \"Public synthesis and evaluation source; model checkpoints not verified.\", \"repository_revision\": \"1ec25af942a46ce46f6c00d596332e308198342f\"}\n\nThe source tree contains static/causal knowledge synthesis and action/knowledge evaluation scripts. Training procedures are delegated to LLaMA-Factory; no trained checkpoint is verified here. The README displays an MIT badge, but the inspected tree has no license file and no explicit component scope, so the code license remains unrecorded rather than assuming a blanket grant. Evaluation requires separately acquired original datasets and model paths.\n\nPrimary source: https://github.com/Wuzheng02/GUI-CIDER\nCatalog record: https://awesomecua.github.io/resources/github-wuzheng02-gui-cider/"}
{"id": "github:wuzheng02/os-kairos", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Wuzheng02/OS-Kairos", "citation_url": "https://awesomecua.github.io/resources/github-wuzheng02-os-kairos/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "15d40c50df8493575a94aa2aa4913c654fea7b7f821d99b7bf443a64f1cc50d3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Wuzheng02/OS-Kairos/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/OS-Kairos"}], "name": "OS-Kairos", "publication": null, "published_at": null, "record_sha256": "f111b79bdfd4290f168a1614497c48d82b85054e3a02bb9cb41be1f8e0706ace", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/OS-Kairos/HEAD/README.md", "target_id": "arxiv:2503.16465", "type": "documented_by"}], "resource_id": "github:wuzheng02/os-kairos", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Wuzheng02/OS-Kairos", "topics": ["human-in-the-loop", "safety-security", "mobile-navigation"], "updated_at": null}, "text": "OS-Kairos\nResource type: agent\nReview status: curated\nScope: core\n\nAndroid GUI-agent research implementation that uses confidence-guided interaction to choose between autonomous actions and human assistance.\nPlatforms: android\nTopics: human-in-the-loop, safety-security, mobile-navigation\n\nThe documented evaluation modes distinguish model-simulated human feedback from actual human interaction; preserve that distinction when interpreting results.\n\nPrimary source: https://github.com/Wuzheng02/OS-Kairos\nCatalog record: https://awesomecua.github.io/resources/github-wuzheng02-os-kairos/"}
{"id": "github:wuzheng02/verios", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Wuzheng02/VeriOS", "citation_url": "https://awesomecua.github.io/resources/github-wuzheng02-verios/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T10:59:55Z", "role": "primary", "sha256": "282883d02391a86c5a4d0cdafdc1d33f10a5073f31eba8c852095601d0aeb939", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "notes"], "url": "https://raw.githubusercontent.com/Wuzheng02/VeriOS/HEAD/README.md"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T10:59:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Wuzheng02/VeriOS"}], "name": "VeriOS", "publication": null, "published_at": null, "record_sha256": "99e7d35dd297afbd0bf2eafa21c749f9030e888b268c6fd1d89e465f7346ac3f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/VeriOS/HEAD/README.md", "target_id": "arxiv:2509.07553", "type": "documented_by"}], "resource_id": "github:wuzheng02/verios", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Wuzheng02/VeriOS", "topics": ["safety-security", "human-in-the-loop", "mobile-navigation"], "updated_at": null}, "text": "VeriOS\nResource type: framework\nReview status: curated\nScope: core\n\nResearch code for proactive human–GUI-agent verification, including offline interaction-loop and dual-agent evaluation scripts plus LLaMA-Factory fine-tuning configurations.\nPlatforms: unspecified\nTopics: safety-security, human-in-the-loop, mobile-navigation\n\nThe released scripts evaluate recorded task inputs; the retained README does not establish a live Android execution environment. No runtime behavior or reported results were independently reproduced.\n\nPrimary source: https://github.com/Wuzheng02/VeriOS\nCatalog record: https://awesomecua.github.io/resources/github-wuzheng02-verios/"}
{"id": "github:wwh0411/fedmabench", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/wwh0411/FedMABench", "citation_url": "https://awesomecua.github.io/resources/github-wwh0411-fedmabench/", "details": {"availability": "Public federated-training/evaluation code and ungated annotation partitions; original images are separate dependencies.", "benchmark_protocol": {"actions": ["model predictions on mobile-agent examples"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md", "population": "released FedMABench partition families", "qualifier": "exact", "unit": "dataset families", "value": 6}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "heterogeneity types", "value": 3}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "population": "publisher benchmark overview", "qualifier": "exact", "unit": "federated learning algorithms", "value": 8}], "evaluation": "The provided evaluation workflow performs model inference over selected partitions and invokes test_swift_fed.py for accuracy scoring.", "evidence_urls": ["https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md"], "focus": ["mobile-agent federated training", "data heterogeneity"], "notes": "The dataset-family count does not count tasks, apps, screenshots or client instances. Code and data licenses are unknown; external image sources are required.", "observations": ["mobile-agent examples with externally sourced screenshots"], "setting": "offline"}, "repository_revision": "417d8b0006830f423c7d07c213fc57277d75d966"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "b49e2b419918078ef9c68f3c9becc6c13d8fe530c165a36bdd61e370746eeea8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "3e1b2b643a263af129b3a4092a84efe47d3643f00509f7b9517da21a4aac0371", "supports": ["identity"], "url": "https://api.github.com/repos/wwh0411/FedMABench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:54Z", "role": "primary", "sha256": "4fb977a469cee3d5083da0b5bf2984fedc6b867146a90b20963985e08f9e9716", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/wwh0411/FedMABench/git/trees/master?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "df5379f4a9b350dad0c3d374d2f9a1d6eacd974b6d54750e2a14ecec1faec6e5", "supports": ["identity", "release_status"], "url": "https://huggingface.co/api/datasets/wwh0411/FedMABench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "2874269995d8c9f5dabe266433ecf724bf54c27af599d899f3f0090b8ee37cdc", "supports": ["summary", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:17Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/wwh0411/FedMABench"}], "name": "FedMABench", "publication": null, "published_at": null, "record_sha256": "f29643c5b43190e2cc28afec8ff2d602323a70ad92f958f6bff2d90f6780303d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md", "target_id": "arxiv:2503.05143", "type": "documented_by"}], "resource_id": "github:wwh0411/fedmabench", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/wwh0411/FedMABench", "topics": ["evaluation", "imitation-learning", "mobile-navigation"], "updated_at": null}, "text": "FedMABench\nResource type: benchmark\nReview status: curated\nScope: core\n\nA federated mobile-agent training and evaluation suite that varies user-data heterogeneity and application distribution, with released partition annotations and inference/scoring scripts.\nPlatforms: android\nTopics: evaluation, imitation-learning, mobile-navigation\n\nRecorded specifications: {\"availability\": \"Public federated-training/evaluation code and ungated annotation partitions; original images are separate dependencies.\", \"benchmark_protocol\": {\"actions\": [\"model predictions on mobile-agent examples\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md\", \"population\": \"released FedMABench partition families\", \"qualifier\": \"exact\", \"unit\": \"dataset families\", \"value\": 6}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md\", \"population\": \"publisher benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"heterogeneity types\", \"value\": 3}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md\", \"population\": \"publisher benchmark overview\", \"qualifier\": \"exact\", \"unit\": \"federated learning algorithms\", \"value\": 8}], \"evaluation\": \"The provided evaluation workflow performs model inference over selected partitions and invokes test_swift_fed.py for accuracy scoring.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/wwh0411/FedMABench/HEAD/README.md\", \"https://huggingface.co/datasets/wwh0411/FedMABench/raw/main/README.md\"], \"focus\": [\"mobile-agent federated training\", \"data heterogeneity\"], \"notes\": \"The dataset-family count does not count tasks, apps, screenshots or client instances. Code and data licenses are unknown; external image sources are required.\", \"observations\": [\"mobile-agent examples with externally sourced screenshots\"], \"setting\": \"offline\"}, \"repository_revision\": \"417d8b0006830f423c7d07c213fc57277d75d966\"}\n\nThe public Hub lists six dataset families of JSON/JSONL partition annotations. The repository explicitly says unpacked screenshots are not provided and must be reconstructed from original AndroidControl and AITW sources. This is offline training/evaluation infrastructure rather than a live Android task environment. No repository license file or dataset-license declaration was found in the inspected sources; neither is inferred from the underlying ms-swift project.\n\nPrimary source: https://github.com/wwh0411/FedMABench\nCatalog record: https://awesomecua.github.io/resources/github-wwh0411-fedmabench/"}
{"id": "github:x-lance/meta-gui-baseline", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/X-LANCE/META-GUI-baseline", "citation_url": "https://awesomecua.github.io/resources/github-x-lance-meta-gui-baseline/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "2cea1be73ea75acd46990ab1fb567fbd83ad58a2bc69ca858b207660df9fe8c1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:42Z", "role": "primary", "sha256": "01fb614f0536ea76c1c25809cc8cde66cde21a96ac6ca0ea81b2c093e2157d2a", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T10:48:42Z", "url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-LANCE/META-GUI-baseline"}], "name": "META-GUI baseline", "publication": null, "published_at": null, "record_sha256": "be3d92fd4c894fbbed70e2b51af4cc491f59f23a2524074a904cb46ede09e844", "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-LANCE/META-GUI-baseline/HEAD/README.md", "target_id": "arxiv:2205.11029", "type": "documented_by"}], "resource_id": "github:x-lance/meta-gui-baseline", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/X-LANCE/META-GUI-baseline", "topics": ["mobile-navigation", "human-in-the-loop"], "updated_at": null}, "text": "META-GUI baseline\nResource type: framework\nReview status: curated\nScope: core\n\nTraining and evaluation pipeline for conversational mobile GUI agents, with action prediction, response generation, and dataset processing.\nPlatforms: unspecified\nTopics: mobile-navigation, human-in-the-loop\n\nIts README describes mobile GUI agents without naming an operating system.\n\nPrimary source: https://github.com/X-LANCE/META-GUI-baseline\nCatalog record: https://awesomecua.github.io/resources/github-x-lance-meta-gui-baseline/"}
{"id": "github:x-lance/mobile-env", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/X-LANCE/Mobile-Env", "citation_url": "https://awesomecua.github.io/resources/github-x-lance-mobile-env/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "fa70fd96c57a163404b77431f89dd16864ed104d7160e347742ff87587229961", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:06Z", "role": "primary", "sha256": "58d1e17ffe5109a7ae296caafcadfdbe6a7d176f0bc4ab01e12a689b0499d8bd", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/LICENSE"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:06Z", "url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-LANCE/Mobile-Env"}], "name": "Mobile Env", "publication": null, "published_at": null, "record_sha256": "647687e7d70cccc3b8e6519533804054aadcc83dea095f4db19f87bda9686225", "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-LANCE/Mobile-Env/HEAD/README.md", "target_id": "arxiv:2305.08144", "type": "documented_by"}], "resource_id": "github:x-lance/mobile-env", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/X-LANCE/Mobile-Env", "topics": ["evaluation"], "updated_at": null}, "text": "Mobile Env\nResource type: infrastructure\nReview status: curated\nScope: core\n\nAndroidEnv-based platform for building GUI agent benchmarks and training agents, where task definition files emit instructions, rewards and episode ends from screen text, view hierarchy or log events.\nPlatforms: android\nTopics: evaluation\n\nPrimary source: https://github.com/X-LANCE/Mobile-Env\nCatalog record: https://awesomecua.github.io/resources/github-x-lance-mobile-env/"}
{"id": "github:x-plug/mobileagent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/X-PLUG/MobileAgent", "citation_url": "https://awesomecua.github.io/resources/github-x-plug-mobileagent/", "details": {}, "environments": ["android", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:02Z", "role": "primary", "sha256": "ca402ffd3b05f72a6d72f47de52c962f92e410749a77939044f3301d4330af16", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:03Z", "role": "primary", "sha256": "961819a89dacd1f4f161e1d2ebf616b538bf095a2eecfa3ef40b7ccec633fac4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:36:02Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:03Z", "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/X-PLUG/MobileAgent"}], "name": "Mobile-Agent", "publication": null, "published_at": null, "record_sha256": "9dfcfd132fadefd6a9bf282b339209a4f558f181b6dfa7e52193b45cb546bb09", "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2401.16158", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2406.01014", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2508.15144", "type": "documented_by"}, {"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/README.md", "target_id": "arxiv:2602.16855", "type": "documented_by"}], "resource_id": "github:x-plug/mobileagent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/X-PLUG/MobileAgent", "topics": ["mobile-navigation", "planning"], "updated_at": null}, "text": "Mobile-Agent\nResource type: agent\nReview status: curated\nScope: core\n\nMobile device agent family using visual interface interaction.\nPlatforms: android, cross-platform\nTopics: mobile-navigation, planning\n\nThis repository contains multiple Mobile-Agent generations. Platform coverage differs by version; use the corresponding paper and implementation instructions.\n\nPrimary source: https://github.com/X-PLUG/MobileAgent\nCatalog record: https://awesomecua.github.io/resources/github-x-plug-mobileagent/"}
{"id": "github:xbmxb/coco-agent", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xbmxb/CoCo-Agent", "citation_url": "https://awesomecua.github.io/resources/github-xbmxb-coco-agent/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:12Z", "role": "primary", "sha256": "55de0ed6627a5812d1f26d2a920e0af47afb17f8a99cae41b70a85ba30a23370", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xbmxb/CoCo-Agent/HEAD/README.md"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T10:42:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xbmxb/CoCo-Agent"}], "name": "CoCo-Agent", "publication": null, "published_at": null, "record_sha256": "ef0267ed315d90fee31c9f2284bba158623d9d1038694ab640664e632fca1c1b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xbmxb/CoCo-Agent/HEAD/README.md", "target_id": "arxiv:2402.11941", "type": "documented_by"}], "resource_id": "github:xbmxb/coco-agent", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xbmxb/CoCo-Agent", "topics": ["mobile-navigation", "screen-understanding"], "updated_at": null}, "text": "CoCo-Agent\nResource type: agent\nReview status: curated\nScope: core\n\nResearch implementation combining screen and layout perception with conditional action prediction for smartphone automation.\nPlatforms: android\nTopics: mobile-navigation, screen-understanding\n\nPrimary source: https://github.com/xbmxb/CoCo-Agent\nCatalog record: https://awesomecua.github.io/resources/github-xbmxb-coco-agent/"}
{"id": "github:xlang-ai/agenttrek", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/AgentTrek", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-agenttrek/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "17fa475b96bdd00270c656072d3f8fc55d8433f0e3a1c2762110160ec21a5c35", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/AgentTrek/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/AgentTrek"}], "name": "AgentTrek", "publication": null, "published_at": null, "record_sha256": "e46f6187a48e53aa46b2a7a8984282bc9f061294d3b96e96fa4f2aa24fdff304", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/AgentTrek/HEAD/README.md", "target_id": "arxiv:2412.09605", "type": "documented_by"}], "resource_id": "github:xlang-ai/agenttrek", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/AgentTrek", "topics": ["data-generation", "video"], "updated_at": null}, "text": "AgentTrek\nResource type: dataset\nReview status: curated\nScope: core\n\nSynthesizes browser-agent trajectories by replaying web tutorials as guidance, releasing the trajectory data with multimodal grounding and reasoning and an AgentTrek-32B agent.\nPlatforms: web\nTopics: data-generation, video\n\nPrimary source: https://github.com/xlang-ai/AgentTrek\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-agenttrek/"}
{"id": "github:xlang-ai/aguvis", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/aguvis", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-aguvis/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/aguvis"}], "name": "Aguvis", "publication": null, "published_at": null, "record_sha256": "31637abded4cfcbce1e815155e9ed37bd5d4836ce8ed7bf23c1d1ddcde39ccbf", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "resource_id": "github:xlang-ai/aguvis", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/aguvis", "topics": ["grounding", "planning"], "updated_at": null}, "text": "Aguvis\nResource type: model\nReview status: curated\nScope: core\n\nVision-based GUI agent models with unified action representations.\nPlatforms: cross-platform\nTopics: grounding, planning\n\nPrimary source: https://github.com/xlang-ai/aguvis\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-aguvis/"}
{"id": "github:xlang-ai/opencua", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/OpenCUA", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-opencua/", "details": {}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "91fed11b234775bde91bb766ee65e1015897ced6668de8e80418fcee09f8f75d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:00Z", "role": "primary", "sha256": "a53fdf6cdedc93c99879b2881d8626d61b09e9a03841b3dc6ebe4e267f4fb2bf", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T06:42:00Z", "url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OpenCUA"}], "name": "OpenCUA", "publication": null, "published_at": null, "record_sha256": "44dd9ffdef7d98dda0a5a76c67af7e6154273009296910b00e2b6d33bf6a1c09", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OpenCUA/HEAD/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "resource_id": "github:xlang-ai/opencua", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/OpenCUA", "topics": ["data-generation", "planning"], "updated_at": null}, "text": "OpenCUA\nResource type: model\nReview status: curated\nScope: core\n\nOpen computer-use research stack including models and training resources.\nPlatforms: desktop\nTopics: data-generation, planning\n\nPrimary source: https://github.com/xlang-ai/OpenCUA\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-opencua/"}
{"id": "github:xlang-ai/osworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/OSWorld", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-osworld/", "details": {"benchmark_protocol": {"actions": ["mouse and keyboard via PyAutoGUI"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json", "population": "test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 369}, {"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json", "population": "test_nogdrive.json", "qualifier": "exact", "unit": "task definitions", "value": 361}], "evaluation": "Task-specific execution-based evaluators inspect application or system state.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md", "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50", "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json", "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json"], "focus": ["task completion", "cross-application workflows"], "notes": "Counts were computed from unique task identifiers in the two inspected manifests. The subsets differ; record the manifest, repository revision, agent interface, and step budget with a score.", "observations": ["screenshots"], "setting": "interactive"}, "leaderboard_url": "https://os-world.github.io/"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "83b8684c2f4ac4f31996b8b4752195fcb6cc5429f89d92398177b9b721d787fc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol", "details.leaderboard_url"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:28Z", "role": "primary", "sha256": "be244bfae7f89ef9449b69d85aac44295eb7b5f7e5e1bbedd76df312696676a7", "supports": ["details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:34:59Z", "role": "primary", "sha256": "9ebc5187cbd727ef26c24626820076b102fff812863c640a67c467fea9542ab5", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:34:59Z", "role": "primary", "sha256": "fcb9497e93a8986407345d3012c872c9c2fed253420730fbea64ccfcace67dbb", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OSWorld"}], "name": "OSWorld", "publication": null, "published_at": null, "record_sha256": "97a98547ac077c3576d9a9ba698415c3c3b3fdb8e5a0ca5a7bb24ad92e9d6fa0", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md", "target_id": "arxiv:2404.07972", "type": "documented_by"}], "resource_id": "github:xlang-ai/osworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/OSWorld", "topics": ["evaluation", "cross-app"], "updated_at": null}, "text": "OSWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nExecution-based evaluation of agents performing tasks in real desktop environments.\nPlatforms: desktop, linux, windows, macos\nTopics: evaluation, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"mouse and keyboard via PyAutoGUI\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json\", \"population\": \"test_all.json\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 369}, {\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json\", \"population\": \"test_nogdrive.json\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 361}], \"evaluation\": \"Task-specific execution-based evaluators inspect application or system state.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/README.md\", \"https://export.arxiv.org/api/query?id_list=2403.03186%2C2403.03788%2C2403.07718%2C2403.10170%2C2403.11905%2C2403.16971%2C2403.17918%2C2403.20271%2C2404.01549%2C2404.01744%2C2404.02475%2C2404.03648%2C2404.05427%2C2404.05719%2C2404.05955%2C2404.06474%2C2404.07972%2C2404.08755%2C2404.08860%2C2404.08948%2C2404.09992%2C2404.10887%2C2404.11459%2C2404.16006%2C2404.16048%2C2404.16054%2C2404.16660%2C2404.18074%2C2404.19296%2C2405.00516%2C2405.11120%2C2405.12842%2C2405.14573%2C2405.14751%2C2405.15793%2C2405.16247%2C2405.20309%2C2406.01014%2C2406.01623%2C2406.03679%2C2406.06947%2C2406.07822%2C2406.08184%2C2406.08340%2C2406.08451%2C2406.10227%2C2406.10819%2C2406.11317%2C2406.11896%2C2406.12373&max_results=50\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_all.json\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld/HEAD/evaluation_examples/test_nogdrive.json\"], \"focus\": [\"task completion\", \"cross-application workflows\"], \"notes\": \"Counts were computed from unique task identifiers in the two inspected manifests. The subsets differ; record the manifest, repository revision, agent interface, and step budget with a score.\", \"observations\": [\"screenshots\"], \"setting\": \"interactive\"}, \"leaderboard_url\": \"https://os-world.github.io/\"}\n\nPrimary source: https://github.com/xlang-ai/OSWorld\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-osworld/"}
{"id": "github:xlang-ai/osworld-g", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/OSWorld-G", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-osworld-g/", "details": {"benchmark_protocol": {"actions": ["click point or box center", "refusal represented by negative coordinates in the evaluator"], "counts": [], "evaluation": "The inspected Jedi runner maps predictions from resized-image coordinates back to the original image, then checks the predicted center against rectangle/polygon targets or the refusal rule. It reports overall and classification-group accuracy. This is offline grounding, not OSWorld task-execution success.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["desktop GUI grounding", "infeasible-target refusal"], "notes": "The repository retains original OSWorld-G.json and refined OSWorld-G_refined.json instructions; treat them as alternative benchmark versions, not additional disjoint tasks. Refined instructions reduce requirements beyond grounding. The same repository also contains Jedi training/rendering material, which is excluded from benchmark counts. The standalone generic eval.py entry point is a stub; the named model runner supplies image_size to its geometry helper. No task count or successful execution is inferred from repository file totals. Apache-2.0 is recorded for code only.", "observations": ["screenshot", "instruction"], "setting": "offline"}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:57Z", "role": "primary", "sha256": "e814851bc3016a2abf59358f99b80afd9c2bd5894fc34baddfb7e4d0aa03bd45", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:14:05Z", "role": "primary", "sha256": "48308bba4f26b02cb7b8fbbc433bb31ec81c494dabd3af437ddbb10b8d9ce667", "supports": ["details.benchmark_protocol"], "url": "https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "64f64e0b847e0fa200aa6051e20b8035a436df00e2ef0d769688ab13fa4d3046", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:57Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:57Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:57Z", "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/OSWorld-G"}], "name": "OSWorld-G", "publication": null, "published_at": null, "record_sha256": "e3aa00833c7cb31f03ff95905d73bc591940859356e55e3562e01884e5be8686", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "resource_id": "github:xlang-ai/osworld-g", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/OSWorld-G", "topics": ["evaluation", "grounding"], "updated_at": null}, "text": "OSWorld-G\nResource type: benchmark\nReview status: curated\nScope: core\n\nComputer-use grounding benchmark with original and refined instructions, released with the Jedi synthetic data pipeline, a 4-million-example Jedi dataset and Jedi-3B and Jedi-7B models.\nPlatforms: desktop\nTopics: evaluation, grounding\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"click point or box center\", \"refusal represented by negative coordinates in the evaluator\"], \"counts\": [], \"evaluation\": \"The inspected Jedi runner maps predictions from resized-image coordinates back to the original image, then checks the predicted center against rectangle/polygon targets or the refusal rule. It reports overall and classification-group accuracy. This is offline grounding, not OSWorld task-execution success.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/README.md\", \"https://api.github.com/repos/xlang-ai/OSWorld-G/git/trees/main?recursive=1\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_osworld_g_jedi.py\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py\"], \"focus\": [\"desktop GUI grounding\", \"infeasible-target refusal\"], \"notes\": \"The repository retains original OSWorld-G.json and refined OSWorld-G_refined.json instructions; treat them as alternative benchmark versions, not additional disjoint tasks. Refined instructions reduce requirements beyond grounding. The same repository also contains Jedi training/rendering material, which is excluded from benchmark counts. The standalone generic eval.py entry point is a stub; the named model runner supplies image_size to its geometry helper. No task count or successful execution is inferred from repository file totals. Apache-2.0 is recorded for code only.\", \"observations\": [\"screenshot\", \"instruction\"], \"setting\": \"offline\"}}\n\nPrimary source: https://github.com/xlang-ai/OSWorld-G\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-osworld-g/"}
{"id": "github:xlang-ai/spider2-v", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/xlang-ai/Spider2-V", "citation_url": "https://awesomecua.github.io/resources/github-xlang-ai-spider2-v/", "details": {"benchmark_protocol": {"actions": ["PyAutoGUI mouse and keyboard actions"], "counts": [{"basis": "computed", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json", "population": "evaluation_examples/test_all.json", "qualifier": "exact", "unit": "task definitions", "value": 494}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "population": "README task-suite coverage", "qualifier": "exact", "unit": "applications", "value": 20}], "evaluation": "Task-specific evaluation functions are invoked after agent execution.", "evidence_urls": ["https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json"], "focus": ["data-science workflows", "multi-application task completion"], "notes": "Counted unique task IDs in the full manifest. Verbose versus abstract instructions and account versus non-account tasks form different subsets. The Ubuntu VM and external-account requirements are part of the protocol; no accounts or environments were provisioned.", "observations": ["screenshots", "accessibility trees", "set-of-marks views"], "setting": "interactive"}}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e3edf30ab45da16382b46d8c542607c1bbbfcf077b6df57a2decf938fc84ff0f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "e3edf30ab45da16382b46d8c542607c1bbbfcf077b6df57a2decf938fc84ff0f", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:06:30Z", "role": "primary", "sha256": "338dd38ba7047b0cf6dbc68f3f7b9374f962708fa2dadcc8d126c232dfc31746", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:41:59Z", "role": "primary", "sha256": "a8d1d88a0f9cce3cdbff73b03a80dc5e5d69e3ba2443946c268815a550c9e0ec", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/LICENSE"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:58Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:41:59Z", "url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/xlang-ai/Spider2-V"}], "name": "Spider2-V", "publication": null, "published_at": null, "record_sha256": "ebf904cd3b394dae13c4c1026dd95ff18f127b5c91edcad64a304bcef1c70ffa", "relations": [{"evidence_url": "https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md", "target_id": "arxiv:2407.10956", "type": "documented_by"}], "resource_id": "github:xlang-ai/spider2-v", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/xlang-ai/Spider2-V", "topics": ["evaluation", "workflow"], "updated_at": null}, "text": "Spider2-V\nResource type: benchmark\nReview status: curated\nScope: core\n\nVisual agent evaluation on data-science and engineering workflows.\nPlatforms: web, desktop\nTopics: evaluation, workflow\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"PyAutoGUI mouse and keyboard actions\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json\", \"population\": \"evaluation_examples/test_all.json\", \"qualifier\": \"exact\", \"unit\": \"task definitions\", \"value\": 494}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md\", \"population\": \"README task-suite coverage\", \"qualifier\": \"exact\", \"unit\": \"applications\", \"value\": 20}], \"evaluation\": \"Task-specific evaluation functions are invoked after agent execution.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/README.md\", \"https://raw.githubusercontent.com/xlang-ai/Spider2-V/HEAD/evaluation_examples/test_all.json\"], \"focus\": [\"data-science workflows\", \"multi-application task completion\"], \"notes\": \"Counted unique task IDs in the full manifest. Verbose versus abstract instructions and account versus non-account tasks form different subsets. The Ubuntu VM and external-account requirements are part of the protocol; no accounts or environments were provisioned.\", \"observations\": [\"screenshots\", \"accessibility trees\", \"set-of-marks views\"], \"setting\": \"interactive\"}}\n\nPrimary source: https://github.com/xlang-ai/Spider2-V\nCatalog record: https://awesomecua.github.io/resources/github-xlang-ai-spider2-v/"}
{"id": "github:yiye3/guicourse", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/yiye3/GUICourse", "citation_url": "https://awesomecua.github.io/resources/github-yiye3-guicourse/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/yiye3/GUICourse"}], "name": "GUICourse", "publication": null, "published_at": null, "record_sha256": "c7d691c9c9fa73cfd573a1f8b84817776096eed26d75f966ad3020f91e914d19", "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "resource_id": "github:yiye3/guicourse", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/yiye3/GUICourse", "topics": ["data-generation"], "updated_at": null}, "text": "GUICourse\nResource type: dataset\nReview status: curated\nScope: core\n\nThree datasets for turning general VLMs into GUI agents: GUIEnv for OCR and grounding, GUIAct with 67k single-step and 15k multi-step web and Android actions, and GUIChat dialogues.\nPlatforms: cross-platform\nTopics: data-generation\n\nPrimary source: https://github.com/yiye3/GUICourse\nCatalog record: https://awesomecua.github.io/resources/github-yiye3-guicourse/"}
{"id": "github:yjyddq/riosworld", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/yjyddq/RiOSWorld", "citation_url": "https://awesomecua.github.io/resources/github-yjyddq-riosworld/", "details": {}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:59Z", "role": "primary", "sha256": "c984f114df33af953a97d22e14e14246f8a98e0866edcaf283900b309b71d8d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/yjyddq/RiOSWorld/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:35:59Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/yjyddq/RiOSWorld"}], "name": "RiOSWorld", "publication": null, "published_at": null, "record_sha256": "74c6c9ee7e8baf9c0081d58a4d4b6162daa5caad40865467c0efa1a15076c972", "relations": [{"evidence_url": "https://raw.githubusercontent.com/yjyddq/RiOSWorld/HEAD/README.md", "target_id": "arxiv:2506.00618", "type": "documented_by"}], "resource_id": "github:yjyddq/riosworld", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/yjyddq/RiOSWorld", "topics": ["evaluation"], "updated_at": null}, "text": "RiOSWorld\nResource type: benchmark\nReview status: curated\nScope: core\n\nMeasures risky behavior of multimodal computer-use agents in an Ubuntu virtual machine, with risk scenarios such as phishing emails and released evaluation trajectories.\nPlatforms: desktop, linux\nTopics: evaluation\n\nPrimary source: https://github.com/yjyddq/RiOSWorld\nCatalog record: https://awesomecua.github.io/resources/github-yjyddq-riosworld/"}
{"id": "github:yuqi-zhou/chop", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Yuqi-Zhou/CHOP", "citation_url": "https://awesomecua.github.io/resources/github-yuqi-zhou-chop/", "details": {"platform_constraints": {"harmonyos": "The README limits device debugging support to HarmonyOS 4 or earlier."}}, "environments": ["android", "harmonyos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:14:29Z", "role": "primary", "sha256": "02e231c68ffc37c829789eabc508a220f3d99a33ee8bf676e53c999a0ff54021", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "1a9f9af82764fdfc07dd8c2ec02e66cbb5c7c2cfc899b0e7b53b26e2a774ddf9", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T13:14:29Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Yuqi-Zhou/CHOP"}], "name": "CHOP", "publication": null, "published_at": null, "record_sha256": "5ac4eb0b8071a9a4e88a1f35bfbb290cc7796bb95490d8f8e1085599b8df6d3d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Yuqi-Zhou/CHOP/HEAD/README.md", "target_id": "arxiv:2503.03743", "type": "documented_by"}], "resource_id": "github:yuqi-zhou/chop", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Yuqi-Zhou/CHOP", "topics": ["mobile-navigation", "planning", "workflow"], "updated_at": null}, "text": "CHOP\nResource type: framework\nReview status: curated\nScope: core\n\nMobile automation framework that uses reusable basis subtasks to constrain and refine planning between user goals and executable actions.\nPlatforms: android, harmonyos\nTopics: mobile-navigation, planning, workflow\n\nRecorded specifications: {\"platform_constraints\": {\"harmonyos\": \"The README limits device debugging support to HarmonyOS 4 or earlier.\"}}\n\nThe README says the complete CHOP-ZH dataset remains unreleased. Android and HarmonyOS compatibility is publisher-documented and has not been runtime-tested here.\n\nPrimary source: https://github.com/Yuqi-Zhou/CHOP\nCatalog record: https://awesomecua.github.io/resources/github-yuqi-zhou-chop/"}
{"id": "github:yuqi-zhou/gui-g1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/Yuqi-Zhou/GUI-G1", "citation_url": "https://awesomecua.github.io/resources/github-yuqi-zhou-gui-g1/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "2d64349b514c43fb9d0260247300412d33cadc2a029d8279a00cb2a12f26a25a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T10:48:43Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T10:43:11Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T10:48:43Z", "url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/Yuqi-Zhou/GUI-G1"}], "name": "GUI-G1", "publication": null, "published_at": null, "record_sha256": "fff072c447e8ef64872f674809e342ae0d267354a44d24058b8b8655bf76cebb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Yuqi-Zhou/GUI-G1/HEAD/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "resource_id": "github:yuqi-zhou/gui-g1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/Yuqi-Zhou/GUI-G1", "topics": ["grounding", "reinforcement-learning"], "updated_at": null}, "text": "GUI-G1\nResource type: model\nReview status: curated\nScope: core\n\nGUI-grounding training code that modifies reasoning templates, reward functions, and the GRPO objective, with a linked model release.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\n\nThe README retains an older release-pending notice alongside a later checkpoint-release announcement. Check the linked model card for current weight availability and terms.\n\nPrimary source: https://github.com/Yuqi-Zhou/GUI-G1\nCatalog record: https://awesomecua.github.io/resources/github-yuqi-zhou-gui-g1/"}
{"id": "github:yuxiangchai/amex-codebase", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/YuxiangChai/AMEX-codebase", "citation_url": "https://awesomecua.github.io/resources/github-yuxiangchai-amex-codebase/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "population": "AMEX collection described in the paper", "qualifier": "more-than", "reference_url": "https://arxiv.org/abs/2407.17490", "unit": "screenshots", "value": 104000}], "evidence_urls": ["https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/README.md", "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"], "formats": [], "modalities": ["screenshots", "element annotations", "functionality descriptions", "action chains"], "notes": "The main repository contains companion training code. The scale and annotation levels here come from the paper, not an independent download or completeness audit.", "splits": [], "supervision": ["element grounding", "UI functionality", "interaction trajectories"]}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "3184bfd612543621e57226e9079132dc8d5695339795a60366d3f6d445e13551", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["details.dataset_profile", "relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:11Z", "role": "primary", "sha256": "4aab61e6a95ab6b6598fb64b8e55d787cd49b5c4f242d15372aed2dd312f97aa", "supports": ["relations"], "url": "https://yxchai.com/AMEX/"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:02Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/LICENSE"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:02Z", "url": "https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/YuxiangChai/AMEX-codebase"}], "name": "AMEX", "publication": null, "published_at": null, "record_sha256": "85efd8e08c9bd7a9db8ed859fad13fa857448f4db0c4209112c8f5e346660a7a", "relations": [{"evidence_url": "https://yxchai.com/AMEX/", "target_id": "arxiv:2407.17490", "type": "documented_by"}], "resource_id": "github:yuxiangchai/amex-codebase", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/YuxiangChai/AMEX-codebase", "topics": ["data-generation"], "updated_at": null}, "text": "AMEX\nResource type: dataset\nReview status: curated\nScope: core\n\nCode for AMEX Android GUI data that converts its three annotation levels into question-answer pairs and fine-tunes a LLaMA2-Accessory model, with a released checkpoint.\nPlatforms: android\nTopics: data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending\", \"population\": \"AMEX collection described in the paper\", \"qualifier\": \"more-than\", \"reference_url\": \"https://arxiv.org/abs/2407.17490\", \"unit\": \"screenshots\", \"value\": 104000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/YuxiangChai/AMEX-codebase/HEAD/README.md\", \"https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending\"], \"formats\": [], \"modalities\": [\"screenshots\", \"element annotations\", \"functionality descriptions\", \"action chains\"], \"notes\": \"The main repository contains companion training code. The scale and annotation levels here come from the paper, not an independent download or completeness audit.\", \"splits\": [], \"supervision\": [\"element grounding\", \"UI functionality\", \"interaction trajectories\"]}}\n\nPrimary source: https://github.com/YuxiangChai/AMEX-codebase\nCatalog record: https://awesomecua.github.io/resources/github-yuxiangchai-amex-codebase/"}
{"id": "github:zai-org/open-autoglm", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/zai-org/Open-AutoGLM", "citation_url": "https://awesomecua.github.io/resources/github-zai-org-open-autoglm/", "details": {}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:53Z", "role": "primary", "sha256": "de9c289c418218d40cdf1731775158af771d010266d7ea2aa398f2dc91f22acc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:07Z", "role": "primary", "sha256": "af69329b4667040c498cc87dfcc165db1f5f9452fe2ad3024ad7eef63d467666", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/LICENSE"}], "identifiers": {}, "kind": "agent", "last_verified_at": "2026-10-16T06:40:53Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:07Z", "url": "https://raw.githubusercontent.com/zai-org/Open-AutoGLM/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/zai-org/Open-AutoGLM"}], "name": "Open-AutoGLM", "publication": null, "published_at": null, "record_sha256": "b6048f1d5455570ffd22449ac1c571d5469ea0188a734e65115a030100cfa3ff", "relations": [], "resource_id": "github:zai-org/open-autoglm", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/zai-org/Open-AutoGLM", "topics": ["mobile-navigation"], "updated_at": null}, "text": "Open-AutoGLM\nResource type: agent\nReview status: curated\nScope: core\n\nPhone agent model and framework with device interaction tooling.\nPlatforms: android, ios\nTopics: mobile-navigation\n\nPrimary source: https://github.com/zai-org/Open-AutoGLM\nCatalog record: https://awesomecua.github.io/resources/github-zai-org-open-autoglm/"}
{"id": "github:zju-aces-ise/uidefectmocker", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ZJU-ACES-ISE/UIDefectMocker", "citation_url": "https://awesomecua.github.io/resources/github-zju-aces-ise-uidefectmocker/", "details": {}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T13:17:23Z", "role": "primary", "sha256": "fc3009b355e46549c5128c933bb333aa37b1db5ac8ea8f2d0f300848d8a62200", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T13:29:25Z", "role": "primary", "sha256": "2b75b184f28b204d8e2c7b45a341e4583202cf4c9da8964f108770929d1193ba", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/LICENSE"}], "identifiers": {}, "kind": "tool", "last_verified_at": "2026-10-16T13:17:23Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "MIT", "retrieved_at": "2026-10-16T13:29:25Z", "url": "https://raw.githubusercontent.com/ZJU-ACES-ISE/UIDefectMocker/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-ACES-ISE/UIDefectMocker"}], "name": "UI Defect Mocker", "publication": null, "published_at": null, "record_sha256": "1bdefc5270ae4d73a789b81f04e158bbe92d1b60d47e0d145f7d243f7e9f598c", "relations": [], "resource_id": "github:zju-aces-ise/uidefectmocker", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ZJU-ACES-ISE/UIDefectMocker", "topics": ["data-generation", "evaluation", "screen-understanding"], "updated_at": null}, "text": "UI Defect Mocker\nResource type: tool\nReview status: curated\nScope: core\n\nSynthesizes visual UI defects from mobile screenshots and Android layout metadata to support GUI-testing experiments.\nPlatforms: android\nTopics: data-generation, evaluation, screen-understanding\n\nThis entry describes the released screenshot-defect tool, separately from the under-construction ChatUITest project hub.\n\nPrimary source: https://github.com/ZJU-ACES-ISE/UIDefectMocker\nCatalog record: https://awesomecua.github.io/resources/github-zju-aces-ise-uidefectmocker/"}
{"id": "github:zju-real/clawgui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ZJU-REAL/ClawGUI", "citation_url": "https://awesomecua.github.io/resources/github-zju-real-clawgui/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:03Z", "role": "primary", "sha256": "3f0aaefce32210dfba428fcf9e609e04e75d4d1520d24a3a6377611a9b83fa6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/README.md"}, {"method": "license_header", "retrieved_at": "2026-10-16T06:42:05Z", "role": "primary", "sha256": "af69329b4667040c498cc87dfcc165db1f5f9452fe2ad3024ad7eef63d467666", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-16T06:36:03Z", "licenses": {"code": {"basis": "repository LICENSE header; not a weights/data license", "declared": "Apache-2.0", "retrieved_at": "2026-10-16T06:42:05Z", "url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/ClawGUI"}], "name": "ClawGUI", "publication": null, "published_at": null, "record_sha256": "a5690ff1db7e778a075a9458d8a176ed114337a0c5ab2e0f537bdee856a17c2b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZJU-REAL/ClawGUI/HEAD/README.md", "target_id": "arxiv:2604.11784", "type": "documented_by"}], "resource_id": "github:zju-real/clawgui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ZJU-REAL/ClawGUI", "topics": ["reinforcement-learning", "evaluation"], "updated_at": null}, "text": "ClawGUI\nResource type: framework\nReview status: curated\nScope: core\n\nFramework for GUI agent training, evaluation, and deployment.\nPlatforms: cross-platform\nTopics: reinforcement-learning, evaluation\n\nPrimary source: https://github.com/ZJU-REAL/ClawGUI\nCatalog record: https://awesomecua.github.io/resources/github-zju-real-clawgui/"}
{"id": "github:zju-real/gui-g2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ZJU-REAL/GUI-G2", "citation_url": "https://awesomecua.github.io/resources/github-zju-real-gui-g2/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f9ecdf208ed4b7845ea4dbd1a7c936c016d11f7ec28f17e71338ab3c2f5b5c0d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:00Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-G2"}], "name": "GUI-G2", "publication": null, "published_at": null, "record_sha256": "c85cf07498a28d56aba46eb9a73ddbaa584631a2269abee9d850659029fdc140", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-G2/HEAD/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "resource_id": "github:zju-real/gui-g2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ZJU-REAL/GUI-G2", "topics": ["grounding"], "updated_at": null}, "text": "GUI-G2\nResource type: model\nReview status: curated\nScope: core\n\nGUI grounding models at 3B and 7B trained with Gaussian rewards for click proximity to element centers and region coverage, with variance scaled to element size.\nPlatforms: cross-platform\nTopics: grounding\n\nPrimary source: https://github.com/ZJU-REAL/GUI-G2\nCatalog record: https://awesomecua.github.io/resources/github-zju-real-gui-g2/"}
{"id": "github:zju-real/gui-rcpo", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ZJU-REAL/GUI-RCPO", "citation_url": "https://awesomecua.github.io/resources/github-zju-real-gui-rcpo/", "details": {"availability": "Public region-voting evaluation and policy-optimization source; model/data dependencies external.", "license_components_note": "Apache-2.0 notices in VLM-R1 subdirectories do not establish a root-project license.", "license_review": {"code": {"evidence_urls": ["https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE", "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"], "reason": "Only subdirectory Apache-2.0 notices were established; no blanket project-source declaration is inferred."}}, "repository_revision": "af15ed5fe8b89b0fe5043a3e94f2984c7b126a4b"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "41c235305d28f6ba77ba1c5d1d9298fd6b1bd68ae3c9ea97eb34a8d22d0ba1dc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:55Z", "role": "primary", "sha256": "f5823de05f4828a8ab367ef13ec3a8338e020f0876e153ac2e944e4eb0e242c0", "supports": ["identity"], "url": "https://api.github.com/repos/zju-real/gui-rcpo"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "98ed8f0fa50699b8a3d0e8a71cd008113009a3072b971b8fadf2b92f68877e1e", "supports": ["release_status", "details.repository_revision", "licenses.code"], "url": "https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.code"], "url": "https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:55Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZJU-REAL/GUI-RCPO"}], "name": "GUI-RC / GUI-RCPO", "publication": null, "published_at": null, "record_sha256": "8a26d0096d938edcef443cf30f7d0b1156a744a8908c849b5a1ce3ab3ea7175f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/zju-real/gui-rcpo/HEAD/README.md", "target_id": "arxiv:2508.05615", "type": "documented_by"}], "resource_id": "github:zju-real/gui-rcpo", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ZJU-REAL/GUI-RCPO", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null}, "text": "GUI-RC / GUI-RCPO\nResource type: framework\nReview status: curated\nScope: core\n\nGUI-grounding research code for spatial voting across sampled predictions and test-time policy optimization using region-consistency rewards.\nPlatforms: cross-platform, desktop, web\nTopics: grounding, reinforcement-learning, efficiency\n\nRecorded specifications: {\"availability\": \"Public region-voting evaluation and policy-optimization source; model/data dependencies external.\", \"license_components_note\": \"Apache-2.0 notices in VLM-R1 subdirectories do not establish a root-project license.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://api.github.com/repos/ZJU-REAL/GUI-RCPO/git/trees/main?recursive=1\", \"https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/LICENSE\", \"https://raw.githubusercontent.com/ZJU-REAL/GUI-RCPO/HEAD/VLM-R1/src/open-r1-multimodal/LICENSE\"], \"reason\": \"Only subdirectory Apache-2.0 notices were established; no blanket project-source declaration is inferred.\"}}, \"repository_revision\": \"af15ed5fe8b89b0fe5043a3e94f2984c7b126a4b\"}\n\nThe tree contains GUI-RC evaluation code and VLM-R1-based training scripts, with example JSONL-format paths. It requires separately supplied models and benchmark images; no released trained checkpoint is established here. Apache-2.0 files exist inside VLM-R1 subdirectories, but no repository-wide license is present for all newly added code, so the code slot remains unknown and the component notices are retained.\n\nPrimary source: https://github.com/ZJU-REAL/GUI-RCPO\nCatalog record: https://awesomecua.github.io/resources/github-zju-real-gui-rcpo/"}
{"id": "github:zorazrw/agent-workflow-memory", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/zorazrw/agent-workflow-memory", "citation_url": "https://awesomecua.github.io/resources/github-zorazrw-agent-workflow-memory/", "details": {"availability": "Public WebArena/Mind2Web pipeline, workflow induction and inference source.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"], "reason": "Project source LICENSE, without transferring rights to external benchmark data or model services."}}, "repository_revision": "8c0ff8cd11d648c8fceb99e4e42f37e3b75381b1"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "dd2cc4823b2a0b913f44fbaecf9f2cea98257871bdf974de6979459e49cb7b63", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:56Z", "role": "primary", "sha256": "f5c44d559932a22add3e94727b9b413d7c286f835db1396ae0fa1f05e20e773e", "supports": ["identity"], "url": "https://api.github.com/repos/zorazrw/agent-workflow-memory"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:06Z", "role": "primary", "sha256": "602e2238829ad29ef15c75ac7dc61f8bd3dee6a1a74adf0d8ce023067fde6da3", "supports": ["release_status", "details.repository_revision"], "url": "https://api.github.com/repos/zorazrw/agent-workflow-memory/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1e504b3bd326f7858b39e7f2bedf5ff55d69ffb126e11af364b4184fb980882e", "supports": ["notes", "release_status"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/webarena/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:11Z", "role": "primary", "sha256": "e047de93e1c73a39ab6bbf310f02e6023d15c2b06a2b1e80ad4a0875181a9533", "supports": ["notes", "release_status"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/mind2web/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-17T12:30:56Z", "licenses": {"code": {"basis": "Project source LICENSE, without transferring rights to external benchmark data or model services.", "declared": "Apache-2.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/zorazrw/agent-workflow-memory"}], "name": "Agent Workflow Memory implementation", "publication": null, "published_at": null, "record_sha256": "74cf70a7501d26caf166b50f3e755da529822c38fe781a1ed7a204d66100678f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/README.md", "target_id": "arxiv:2409.07429", "type": "documented_by"}], "resource_id": "github:zorazrw/agent-workflow-memory", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/zorazrw/agent-workflow-memory", "topics": ["workflow", "memory", "web-navigation"], "updated_at": null}, "text": "Agent Workflow Memory implementation\nResource type: framework\nReview status: curated\nScope: core\n\nResearch code that extracts reusable browser-task routines from demonstrations or prior executions, stores them as workflow memory and reuses them during subsequent navigation.\nPlatforms: web\nTopics: workflow, memory, web-navigation\n\nRecorded specifications: {\"availability\": \"Public WebArena/Mind2Web pipeline, workflow induction and inference source.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/zorazrw/agent-workflow-memory/HEAD/LICENSE\"], \"reason\": \"Project source LICENSE, without transferring rights to external benchmark data or model services.\"}}, \"repository_revision\": \"8c0ff8cd11d648c8fceb99e4e42f37e3b75381b1\"}\n\nThe repository releases separate WebArena and Mind2Web pipelines for online or offline workflow induction and use. WebArena execution requires separately configured browser environments, while Mind2Web workflows require the original dataset and rankings. Publisher performance claims were not reproduced. Apache-2.0 is recorded for this implementation; external benchmark, model and service terms remain separate.\n\nPrimary source: https://github.com/zorazrw/agent-workflow-memory\nCatalog record: https://awesomecua.github.io/resources/github-zorazrw-agent-workflow-memory/"}
{"id": "github:zrw00/guipivot", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://github.com/ZrW00/GUIPivot", "citation_url": "https://awesomecua.github.io/resources/github-zrw00-guipivot/", "details": {"availability": "Public preprocessing/refinement code, offline action-matching evaluation, and multipart annotation archives; original images and a fine-tuned model are separate.", "license_review": {"code": {"evidence_urls": ["https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"], "reason": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred."}}, "repository_revision": "b02e8efdbe02ab82169a2c54d724e33b4037330e"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:55.267942+00:00", "role": "primary", "sha256": "77384ff1dfceac74f2e60af56b016a815963481f9e41a29915698777f55e828e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T11:25:27+00:00", "role": "primary", "sha256": "3d57dc43db78620832e41416e4bfb7ae445afe782e5538d85a9784a8f1e5e4eb", "supports": ["identity"], "url": "https://api.github.com/repos/ZrW00/GUIPivot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:54.310430+00:00", "role": "primary", "sha256": "a4d3ddbb2b2f4a936c9bc1a1c388b9870dbc5d2c61ff6841a1082b884f5c54e9", "supports": ["release_status", "details.repository_revision", "details.availability", "notes"], "url": "https://api.github.com/repos/ZrW00/GUIPivot/git/trees/b02e8efdbe02ab82169a2c54d724e33b4037330e?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "role": "primary", "sha256": "3972dc9744f6499f0f9b2dbf76696f2ae7ad8af9b23dde66d6af86c9dfb36986", "supports": ["licenses.code"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:36.501726+00:00", "role": "primary", "sha256": "25e5282e7d85bca133fc77e294cd95d90e4ae65b0d8796cd7c49f53a633b4893", "supports": ["summary", "environments", "notes"], "url": "https://arxiv.org/html/2503.00401v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:03:30.911159+00:00", "role": "primary", "sha256": "9b5b196e065be2ebc05169a68710b8a6e1bd3b6eb663c68bb449234dbb7a45af", "supports": ["notes", "details.availability"], "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/data/dataset_info.json"}], "identifiers": {}, "kind": "framework", "last_verified_at": "2026-10-18T12:02:55.267942+00:00", "licenses": {"code": {"basis": "Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred.", "declared": "GPL-3.0", "retrieved_at": "2026-10-18T12:03:28.611747+00:00", "url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE"}, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://github.com/ZrW00/GUIPivot"}], "name": "GUIPivot", "publication": null, "published_at": null, "record_sha256": "fd08578cd0286ed547daf9d490570816038f9ee3b45bb1280761b5d0598edf29", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/README.md", "target_id": "arxiv:2503.00401", "type": "documented_by"}], "resource_id": "github:zrw00/guipivot", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://github.com/ZrW00/GUIPivot", "topics": ["grounding", "planning", "imitation-learning"], "updated_at": null}, "text": "GUIPivot\nResource type: framework\nReview status: curated\nScope: core\n\nQuery-inference training tools connecting screenshot grounding with GUI action reasoning, including preprocessing, query refinement and offline single-step evaluation on AndroidControl and AITZ.\nPlatforms: android\nTopics: grounding, planning, imitation-learning\n\nRecorded specifications: {\"availability\": \"Public preprocessing/refinement code, offline action-matching evaluation, and multipart annotation archives; original images and a fine-tuned model are separate.\", \"license_review\": {\"code\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/ZrW00/GUIPivot/b02e8efdbe02ab82169a2c54d724e33b4037330e/LICENSE\"], \"reason\": \"Repository GNU GPL version 3 license for code; no data or pretrained-weight license inferred.\"}}, \"repository_revision\": \"b02e8efdbe02ab82169a2c54d724e33b4037330e\"}\n\nThe repository releases preprocessing/refinement/evaluation code and multipart ShareGPT-format JSON annotations. Evaluation matches recorded single-step action predictions on AndroidControl and AITZ; it does not run live device tasks. The README links a UIBERT training-JSON path that is not a standalone file in this revision; annotations are packaged in data/GUI.z01, data/GUI.z02 and data/GUI.zip. Original UIBERT/OS-Atlas, AITZ and AndroidControl images must be obtained separately. This entry does not claim a released trained checkpoint or transfer the GPL-3.0 code license to third-party datasets and images.\n\nPrimary source: https://github.com/ZrW00/GUIPivot\nCatalog record: https://awesomecua.github.io/resources/github-zrw00-guipivot/"}
{"id": "organization:anthropic", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://claude.com/chrome", "citation_url": "https://awesomecua.github.io/resources/organization-anthropic/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "bd3e99c1e12588bc3136dfa27de2c0d5790b6f6f25f2696988eb182c6b1ef670", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://claude.com/chrome"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://claude.com/chrome"}], "name": "Anthropic", "publication": null, "published_at": null, "record_sha256": "306bb8dd42a2c52f89b23613bc73e190d9efbf29c4848d5c140146bbb70eb499", "relations": [], "resource_id": "organization:anthropic", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://claude.com/chrome", "topics": ["industry"], "updated_at": null}, "text": "Anthropic\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Anthropic computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://claude.com/chrome\nCatalog record: https://awesomecua.github.io/resources/organization-anthropic/"}
{"id": "organization:askui", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.askui.com/", "citation_url": "https://awesomecua.github.io/resources/organization-askui/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "df8615d0687cba27e5558ea3f7ecb1823871e3804522603dd62b52dcf127ac89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.askui.com/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.askui.com/"}], "name": "AskUI", "publication": null, "published_at": null, "record_sha256": "8d2597b7831b4d37e4dc29db35b4338c41e6a26e6167409598c550a05934a302", "relations": [{"evidence_url": "https://www.askui.com/", "target_id": "organization:askui", "type": "published_by"}], "resource_id": "organization:askui", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.askui.com/", "topics": ["industry"], "updated_at": null}, "text": "AskUI\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with AskUI computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.askui.com/\nCatalog record: https://awesomecua.github.io/resources/organization-askui/"}
{"id": "organization:browser-use", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://browser-use.com/", "citation_url": "https://awesomecua.github.io/resources/organization-browser-use/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "69e029d01dda3806f88abfaee01da08c671adcb3555f1314e975ab8942a686ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://browser-use.com/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://browser-use.com/"}], "name": "Browser Use", "publication": null, "published_at": null, "record_sha256": "b092c8978e3dadd1810290489093e978008df4be11ef742549c14378ff4f57f6", "relations": [{"evidence_url": "https://browser-use.com/", "target_id": "organization:browser-use", "type": "published_by"}], "resource_id": "organization:browser-use", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://browser-use.com/", "topics": ["industry"], "updated_at": null}, "text": "Browser Use\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Browser Use computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://browser-use.com/\nCatalog record: https://awesomecua.github.io/resources/organization-browser-use/"}
{"id": "organization:browseract", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browseract.com/", "citation_url": "https://awesomecua.github.io/resources/organization-browseract/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "463db02b607e7215a4421440c4d274ee3546bd9eefc50ed9e10f64d27d7f8f38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browseract.com/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browseract.com/"}], "name": "BrowserAct", "publication": null, "published_at": null, "record_sha256": "0b11fe4567ef00dd82b548b1cd14b0d4782420fdd74fd0bb84025a4613d5d5eb", "relations": [{"evidence_url": "https://www.browseract.com/", "target_id": "organization:browseract", "type": "published_by"}], "resource_id": "organization:browseract", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browseract.com/", "topics": ["industry"], "updated_at": null}, "text": "BrowserAct\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with BrowserAct computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.browseract.com/\nCatalog record: https://awesomecua.github.io/resources/organization-browseract/"}
{"id": "organization:browserbase", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browserbase.com/", "citation_url": "https://awesomecua.github.io/resources/organization-browserbase/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "d86315819750e4a5fde12bce85a03c61cf68e704cb54146e6180027a37471b72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browserbase.com/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserbase.com/"}], "name": "Browserbase", "publication": null, "published_at": null, "record_sha256": "7515cd233726a806f6b17867c0b9a62987819f2903523f67df4cd06f72d77583", "relations": [], "resource_id": "organization:browserbase", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browserbase.com/", "topics": ["industry"], "updated_at": null}, "text": "Browserbase\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Browserbase computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.browserbase.com/\nCatalog record: https://awesomecua.github.io/resources/organization-browserbase/"}
{"id": "organization:browserless", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browserless.io/", "citation_url": "https://awesomecua.github.io/resources/organization-browserless/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "e044e75ec1c2a74f2bee8c02b6efae05c399ec3537e34dd861dcf4a0a25dc182", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.browserless.io/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserless.io/"}], "name": "Browserless", "publication": null, "published_at": null, "record_sha256": "29c8d4b3ee5c3b20df60f24d475976e628fc7eba24083d5eade21c3e36c4faa0", "relations": [{"evidence_url": "https://www.browserless.io/", "target_id": "organization:browserless", "type": "published_by"}], "resource_id": "organization:browserless", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browserless.io/", "topics": ["industry"], "updated_at": null}, "text": "Browserless\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Browserless computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.browserless.io/\nCatalog record: https://awesomecua.github.io/resources/organization-browserless/"}
{"id": "organization:e2b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://e2b.dev/", "citation_url": "https://awesomecua.github.io/resources/organization-e2b/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "6232364481bbae45a3fcec1d1aca858493baff0693ececa98b09b610183e6796", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://e2b.dev/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://e2b.dev/"}], "name": "E2B", "publication": null, "published_at": null, "record_sha256": "06377215e50c1401cd4dc807c4880d33288509e457fb7036e61f171cad81451c", "relations": [{"evidence_url": "https://e2b.dev/", "target_id": "organization:e2b", "type": "published_by"}], "resource_id": "organization:e2b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://e2b.dev/", "topics": ["industry"], "updated_at": null}, "text": "E2B\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with E2B computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://e2b.dev/\nCatalog record: https://awesomecua.github.io/resources/organization-e2b/"}
{"id": "organization:google", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://ai.google.dev/gemini-api/docs/computer-use", "citation_url": "https://awesomecua.github.io/resources/organization-google/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e7062cfb4c0948d3a2274843f9f493f34cb40dfd634371408086df6141a0e036", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "name": "Google", "publication": null, "published_at": null, "record_sha256": "446d7fa1d563a10bbd9aed4e5c669ea01ffe23b92d0cf0beb243bc265f1e9943", "relations": [], "resource_id": "organization:google", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://ai.google.dev/gemini-api/docs/computer-use", "topics": ["industry"], "updated_at": null}, "text": "Google\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Google computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://ai.google.dev/gemini-api/docs/computer-use\nCatalog record: https://awesomecua.github.io/resources/organization-google/"}
{"id": "organization:h", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://hcompany.ai/research/models", "citation_url": "https://awesomecua.github.io/resources/organization-h/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "84ce0fd682367a6dac5cc5abc39d5c877556024bbf6ffdad30f76fc2ec08523f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://hcompany.ai/research/models"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/research/models"}], "name": "H", "publication": null, "published_at": null, "record_sha256": "cd150004c67b62acadbdef0924625cd503acf5d092d6976a2e9582b582ec948c", "relations": [], "resource_id": "organization:h", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://hcompany.ai/research/models", "topics": ["industry"], "updated_at": null}, "text": "H\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with H computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://hcompany.ai/research/models\nCatalog record: https://awesomecua.github.io/resources/organization-h/"}
{"id": "organization:hyperbrowser", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.hyperbrowser.ai/", "citation_url": "https://awesomecua.github.io/resources/organization-hyperbrowser/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "47f31d41c806e97143fa046c4e1f5e916096d158bee0bff43efd15a7d78053b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.hyperbrowser.ai/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.hyperbrowser.ai/"}], "name": "Hyperbrowser", "publication": null, "published_at": null, "record_sha256": "a3dcd177e7d056041256ef181073de0ddd09144b4859299be0e18b89811f00f8", "relations": [{"evidence_url": "https://www.hyperbrowser.ai/", "target_id": "organization:hyperbrowser", "type": "published_by"}], "resource_id": "organization:hyperbrowser", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.hyperbrowser.ai/", "topics": ["industry"], "updated_at": null}, "text": "Hyperbrowser\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Hyperbrowser computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.hyperbrowser.ai/\nCatalog record: https://awesomecua.github.io/resources/organization-hyperbrowser/"}
{"id": "organization:manus", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://manus.im/", "citation_url": "https://awesomecua.github.io/resources/organization-manus/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "832b84c366eaadc69e328f207f2659cf07a0901c47662244831c0f1695d02efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://manus.im/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://manus.im/"}], "name": "Manus", "publication": null, "published_at": null, "record_sha256": "648ec9e5b326bd1093844593ff40d429873350240c2bf30574c3f9c9c11f5dd9", "relations": [{"evidence_url": "https://manus.im/", "target_id": "organization:manus", "type": "published_by"}], "resource_id": "organization:manus", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://manus.im/", "topics": ["industry"], "updated_at": null}, "text": "Manus\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Manus computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://manus.im/\nCatalog record: https://awesomecua.github.io/resources/organization-manus/"}
{"id": "organization:openai", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://openai.com/index/computer-using-agent/", "citation_url": "https://awesomecua.github.io/resources/organization-openai/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "14f573af0c4ea3e5922b8831b7f03724936963e293d7c0aaa3fccbb71dbb0cfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://openai.com/index/computer-using-agent/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/computer-using-agent/"}], "name": "OpenAI", "publication": null, "published_at": null, "record_sha256": "20bc7111ba1c73298f6d18d9ed091019b9dfde3d2a8fee0d08c463300c2f99cb", "relations": [], "resource_id": "organization:openai", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openai.com/index/computer-using-agent/", "topics": ["industry"], "updated_at": null}, "text": "OpenAI\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with OpenAI computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://openai.com/index/computer-using-agent/\nCatalog record: https://awesomecua.github.io/resources/organization-openai/"}
{"id": "organization:simular", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.simular.ai/", "citation_url": "https://awesomecua.github.io/resources/organization-simular/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "4cb358ab96e57a66d2b268a17963f92be249325fb0687a78985a6d9773ff5a81", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.simular.ai/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.simular.ai/"}], "name": "Simular", "publication": null, "published_at": null, "record_sha256": "14d47d80e8d2fa8787bfcee231ad6baf284081ccd8a8dc1536e0cf00da0b425f", "relations": [{"evidence_url": "https://www.simular.ai/", "target_id": "organization:simular", "type": "published_by"}], "resource_id": "organization:simular", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.simular.ai/", "topics": ["industry"], "updated_at": null}, "text": "Simular\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Simular computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.simular.ai/\nCatalog record: https://awesomecua.github.io/resources/organization-simular/"}
{"id": "organization:skyvern", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.skyvern.com/", "citation_url": "https://awesomecua.github.io/resources/organization-skyvern/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7a8813c174043d1a27cb3766f77c90beb1a08eeedee935264d6237d6a57672a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.skyvern.com/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.skyvern.com/"}], "name": "Skyvern", "publication": null, "published_at": null, "record_sha256": "ace8bb147284a11e336a951f66f7d8d6dc817548b6b2cab292f8f08cc5169d3e", "relations": [{"evidence_url": "https://www.skyvern.com/", "target_id": "organization:skyvern", "type": "published_by"}], "resource_id": "organization:skyvern", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.skyvern.com/", "topics": ["industry"], "updated_at": null}, "text": "Skyvern\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Skyvern computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.skyvern.com/\nCatalog record: https://awesomecua.github.io/resources/organization-skyvern/"}
{"id": "organization:steel", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://steel.dev/", "citation_url": "https://awesomecua.github.io/resources/organization-steel/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "5eacfa15c116ddeb02f6fd731fbb2e9856a933d743870fe41d947ef3b01aeb6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://steel.dev/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://steel.dev/"}], "name": "Steel", "publication": null, "published_at": null, "record_sha256": "4c8d97f8a9de72c0d9943cb6b9cca89e23f2d012026e511ec56fbbd7de34381c", "relations": [], "resource_id": "organization:steel", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://steel.dev/", "topics": ["industry"], "updated_at": null}, "text": "Steel\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with Steel computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://steel.dev/\nCatalog record: https://awesomecua.github.io/resources/organization-steel/"}
{"id": "organization:tinyfish", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.tinyfish.ai/", "citation_url": "https://awesomecua.github.io/resources/organization-tinyfish/", "details": {}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "461de0060fae1ded024dfa724624b001665b463b14f454afafebd0c9a2dce71d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://www.tinyfish.ai/"}], "identifiers": {}, "kind": "company", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.tinyfish.ai/"}], "name": "TinyFish", "publication": null, "published_at": null, "record_sha256": "d336e95cfbe79f4c5e0262d656592904830782a804c777b703d37afb019839a0", "relations": [{"evidence_url": "https://www.tinyfish.ai/", "target_id": "organization:tinyfish", "type": "published_by"}], "resource_id": "organization:tinyfish", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.tinyfish.ai/", "topics": ["industry"], "updated_at": null}, "text": "TinyFish\nResource type: company\nReview status: curated\nScope: core\n\nPublisher or product brand associated with TinyFish computer-use or browser automation offerings.\nPlatforms: unspecified\nTopics: industry\n\nPublisher/brand identity only. Corporate ownership, jurisdiction, financing, and commercial traction are not inferred.\n\nPrimary source: https://www.tinyfish.ai/\nCatalog record: https://awesomecua.github.io/resources/organization-tinyfish/"}
{"id": "url:00b1a3abf9bcb424b264", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B", "citation_url": "https://awesomecua.github.io/resources/url-00b1a3abf9bcb424b264/", "details": {"disabled": false, "download_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/tree/main", "gated": false, "hub_id": "Tongyi-MAI/MAI-UI-8B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "349e6d5753fa94a47feb4b2fc53ecd6b9188b645a0cffabef92fd1aa0976ecad", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Tongyi-MAI/MAI-UI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "a1087d2a0dde6e7ba6d79db546ee6360a5ffd95ffbc3add1478eedec603a4187", "supports": ["licenses.weights"], "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B"}], "name": "MAI-UI 8B", "publication": null, "published_at": null, "record_sha256": "2dbd86086f525f9c83533f8de7f0173f612a7a62ad1ece355b6a64ae95c7ece9", "relations": [{"evidence_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B/raw/main/README.md", "target_id": "arxiv:2512.22047", "type": "documented_by"}], "resource_id": "url:00b1a3abf9bcb424b264", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Tongyi-MAI/MAI-UI-8B", "topics": ["grounding", "mobile-navigation", "tool-integration"], "updated_at": null}, "text": "MAI-UI 8B\nResource type: model\nReview status: curated\nScope: core\n\nA released MAI-UI checkpoint combining mobile GUI actions, agent-user interaction and tool-use-oriented navigation.\nPlatforms: android\nTopics: grounding, mobile-navigation, tool-integration\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Tongyi-MAI/MAI-UI-8B/tree/main\", \"gated\": false, \"hub_id\": \"Tongyi-MAI/MAI-UI-8B\"}\n\nFour public safetensors shards are present and Apache-2.0 is declared. The card presents family-wide results and device-cloud architecture; those claims are not all assigned to the 8B checkpoint. Other parameter sizes are not duplicated in this proposal batch.\n\nPrimary source: https://huggingface.co/Tongyi-MAI/MAI-UI-8B\nCatalog record: https://awesomecua.github.io/resources/url-00b1a3abf9bcb424b264/"}
{"id": "url:04aa7569d183379d615c", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/henryhe0123/PC-Agent-E", "citation_url": "https://awesomecua.github.io/resources/url-04aa7569d183379d615c/", "details": {"disabled": false, "download_url": "https://huggingface.co/henryhe0123/PC-Agent-E/tree/568ab45861d5dab041e158d47211c6f53ba343f2", "gated": false, "hub_id": "henryhe0123/PC-Agent-E", "hub_revision": "568ab45861d5dab041e158d47211c6f53ba343f2", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:19Z", "role": "primary", "sha256": "feb41002efff5ef30fe06f9891efbe954e0d046c0379f7c8ca72b5d8a8a304be", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "f81a6cea4234e65f8bab390a18487e65dd0b0cb62f0a9458c658ff07e8a7b5f7", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/henryhe0123/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["environments", "details.repository_url"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:19Z", "url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/henryhe0123/PC-Agent-E"}], "name": "PC-Agent-E checkpoint", "publication": null, "published_at": null, "record_sha256": "29d22dfba7d8f1a073c9fd477e02b2d9cf03ed38e1a50acfb69196ddaeadbf90", "relations": [{"evidence_url": "https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "resource_id": "url:04aa7569d183379d615c", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/henryhe0123/PC-Agent-E", "topics": ["action-prediction", "imitation-learning"], "updated_at": null}, "text": "PC-Agent-E checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nComputer-use checkpoint fine-tuned on PC-Agent-E trajectories; the model-card prose identifies Qwen2.5-VL-72B-Instruct as its starting model.\nPlatforms: desktop, windows, linux\nTopics: action-prediction, imitation-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/henryhe0123/PC-Agent-E/tree/568ab45861d5dab041e158d47211c6f53ba343f2\", \"gated\": false, \"hub_id\": \"henryhe0123/PC-Agent-E\", \"hub_revision\": \"568ab45861d5dab041e158d47211c6f53ba343f2\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/henryhe0123/PC-Agent-E/raw/568ab45861d5dab041e158d47211c6f53ba343f2/README.md\"], \"reason\": \"Publisher model-card declaration for the checkpoint; code and training-data terms are separate.\"}}, \"repository_url\": \"https://github.com/GAIR-NLP/PC-Agent-E\"}\n\nPublic manifest lists 31 Safetensors shards and their index. The card prose names Qwen2.5-VL-72B-Instruct while its YAML base_model field self-references PC-Agent-E; no base-model relationship is imported from that inconsistent metadata. The model-card MIT declaration is recorded for weights only.\n\nPrimary source: https://huggingface.co/henryhe0123/PC-Agent-E\nCatalog record: https://awesomecua.github.io/resources/url-04aa7569d183379d615c/"}
{"id": "url:053ebfe0afac66c86222", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.hyperbrowser.ai/", "citation_url": "https://awesomecua.github.io/resources/url-053ebfe0afac66c86222/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "47f31d41c806e97143fa046c4e1f5e916096d158bee0bff43efd15a7d78053b1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.hyperbrowser.ai/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.hyperbrowser.ai/"}], "name": "Hyperbrowser", "publication": null, "published_at": null, "record_sha256": "123c1ea1ad2dcbf757c3e736c658b50a7d512a5e22467ba3e37c6829d3d75d2c", "relations": [{"evidence_url": "https://www.hyperbrowser.ai/", "target_id": "organization:hyperbrowser", "type": "published_by"}], "resource_id": "url:053ebfe0afac66c86222", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.hyperbrowser.ai/", "topics": ["sandbox"], "updated_at": null}, "text": "Hyperbrowser\nResource type: product\nReview status: curated\nScope: core\n\nBrowser infrastructure and automation services for AI agents.\nPlatforms: web\nTopics: sandbox\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.hyperbrowser.ai/\nCatalog record: https://awesomecua.github.io/resources/url-053ebfe0afac66c86222/"}
{"id": "url:06dd5ac7917c45d48509", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1", "citation_url": "https://awesomecua.github.io/resources/url-06dd5ac7917c45d48509/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "https://huggingface.co/api/datasets/xlangai/aguvis-stage1", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart TAR image packages"], "modalities": ["GUI screenshots", "grounding annotations"], "notes": "Stage 1 grounding release, Hub revision 5d11f8f5. JSON annotations and ZIP/multipart TAR image packages are publicly listed. The official training configuration groups web, desktop and mobile inputs and selects all rows of its configured inputs; this is a training recipe, not a release sample count or held-out split. The config names omniact.json while the Hub manifest lists omniact_fix.json, so local paths need reconciliation. No sample total is inferred from names such as webui350k. Apache-2.0 is the release card declaration; upstream source rights remain separate.", "splits": ["training"], "supervision": ["GUI grounding supervised fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/tree/main", "gated": false, "hub_id": "xlangai/aguvis-stage1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "cb3b8ec88d811d946df3cbf38ef555db5a10d1124668765710c380e9b1988b06", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xlangai/aguvis-stage1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["environments", "summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "f6a9e55c5d48469e2c8f7799d917d25e7a758036f71c7c045fa92c35057d20da", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "38bda302cf741b9d2b2394b53d157b3c19d8a73e3704c01ed2f9a27c374feb77", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage1"}], "name": "Aguvis stage 1 grounding data", "publication": null, "published_at": null, "record_sha256": "0b6e093a2cfb2a6a592c1a466722e2337730ed676e301a0c8c23eaa8804403e8", "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "resource_id": "url:06dd5ac7917c45d48509", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xlangai/aguvis-stage1", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "Aguvis stage 1 grounding data\nResource type: dataset\nReview status: curated\nScope: core\n\nThe first-stage Aguvis training collection, packaging GUI element-grounding annotations and image archives from multiple interface-data sources.\nPlatforms: cross-platform\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/xlangai/aguvis-stage1/raw/main/README.md\", \"https://huggingface.co/api/datasets/xlangai/aguvis-stage1\", \"https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md\", \"https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage1.yaml\"], \"formats\": [\"JSON annotations\", \"ZIP image packages\", \"multipart TAR image packages\"], \"modalities\": [\"GUI screenshots\", \"grounding annotations\"], \"notes\": \"Stage 1 grounding release, Hub revision 5d11f8f5. JSON annotations and ZIP/multipart TAR image packages are publicly listed. The official training configuration groups web, desktop and mobile inputs and selects all rows of its configured inputs; this is a training recipe, not a release sample count or held-out split. The config names omniact.json while the Hub manifest lists omniact_fix.json, so local paths need reconciliation. No sample total is inferred from names such as webui350k. Apache-2.0 is the release card declaration; upstream source rights remain separate.\", \"splits\": [\"training\"], \"supervision\": [\"GUI grounding supervised fine-tuning\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xlangai/aguvis-stage1/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/aguvis-stage1\"}\n\nThe manifest contains grounding JSON annotations and packaged images. The card declares Apache-2.0, which is recorded for this release without overriding source-dataset rights. This first-stage grounding collection is distinct from the project’s second-stage trajectory collection. No archive or training sample was downloaded.\n\nPrimary source: https://huggingface.co/datasets/xlangai/aguvis-stage1\nCatalog record: https://awesomecua.github.io/resources/url-06dd5ac7917c45d48509/"}
{"id": "url:07b65dc937756ddff4a1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data", "citation_url": "https://awesomecua.github.io/resources/url-07b65dc937756ddff4a1/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "population": "context-aware episode training release", "qualifier": "approximately", "unit": "instruction-output pairs", "value": 992000}], "evidence_urls": ["https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"], "formats": [], "modalities": ["screen observations", "instructions", "interaction history"], "notes": "The Hub card contains only licensing metadata. The project README supports the episode-training description and rounded count; archive-level quality and exact row counts remain unverified.", "splits": [], "supervision": ["context-aware grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI_Context-aware_Data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["name", "summary", "environments", "topics", "scope", "details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "79f91aab71bb8a0ec3ad9859cbc42d70227e01db729ec80dc5690b466b259cad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data"}], "name": "Aria-UI Context-aware Data", "publication": null, "published_at": null, "record_sha256": "437593296254e5a6c3085544d96ba55fbbc70168226a44ef5071bc3d2a0499db", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "github:ariaui/aria-ui", "type": "component_of"}], "resource_id": "url:07b65dc937756ddff4a1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data", "topics": ["grounding", "memory"], "updated_at": null}, "text": "Aria-UI Context-aware Data\nResource type: dataset\nReview status: curated\nScope: core\n\nEpisode-based mobile GUI grounding data released with Aria-UI for incorporating interaction history into target localization.\nPlatforms: android\nTopics: grounding, memory\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md\", \"population\": \"context-aware episode training release\", \"qualifier\": \"approximately\", \"unit\": \"instruction-output pairs\", \"value\": 992000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md\"], \"formats\": [], \"modalities\": [\"screen observations\", \"instructions\", \"interaction history\"], \"notes\": \"The Hub card contains only licensing metadata. The project README supports the episode-training description and rounded count; archive-level quality and exact row counts remain unverified.\", \"splits\": [], \"supervision\": [\"context-aware grounding\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data/tree/main\", \"gated\": false, \"hub_id\": \"Aria-UI/Aria-UI_Context-aware_Data\"}\n\nThe publisher reports approximately 992K instruction-output pairs in its repository. The Hub card contains only a license declaration; schema quality and exact sample counts were not independently verified.\n\nPrimary source: https://huggingface.co/datasets/Aria-UI/Aria-UI_Context-aware_Data\nCatalog record: https://awesomecua.github.io/resources/url-07b65dc937756ddff4a1/"}
{"id": "url:07b6bbce7ce26a847f79", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://claude.com/chrome", "citation_url": "https://awesomecua.github.io/resources/url-07b6bbce7ce26a847f79/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "bd3e99c1e12588bc3136dfa27de2c0d5790b6f6f25f2696988eb182c6b1ef670", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://claude.com/chrome"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://claude.com/chrome"}], "name": "Claude for Chrome", "publication": null, "published_at": null, "record_sha256": "f1ce0810e6b2e08d742dfbdf297c1fab9ee1db1fa1454a161125e6397bc8e6f9", "relations": [{"evidence_url": "https://claude.com/chrome", "target_id": "organization:anthropic", "type": "published_by"}], "resource_id": "url:07b6bbce7ce26a847f79", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://claude.com/chrome", "topics": ["web-navigation"], "updated_at": null}, "text": "Claude for Chrome\nResource type: product\nReview status: curated\nScope: core\n\nClaude browser extension for interacting with websites.\nPlatforms: web\nTopics: web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://claude.com/chrome\nCatalog record: https://awesomecua.github.io/resources/url-07b6bbce7ce26a847f79/"}
{"id": "url:085b6811ddd6f797475d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware", "citation_url": "https://awesomecua.github.io/resources/url-085b6811ddd6f797475d/", "details": {"disabled": false, "download_url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware/tree/main", "gated": false, "hub_id": "Aria-UI/Aria-UI-context-aware", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["name", "kind", "summary", "environments", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a2a3c5ad115e4e2109b56686c9e0b073bcfec6d9f052a58bc92683af44933480", "supports": ["summary", "environments", "relations", "release_status", "topics"], "url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "183a8de8928b3b71b171a35149621de078aaf80b4ebde377a5dc3b5cd4a04a27", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/Aria-UI/Aria-UI-context-aware"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware"}], "name": "Aria-UI context-aware checkpoint", "publication": null, "published_at": null, "record_sha256": "9ad510042521bd72805d6225cd57344c9a8126df8c381d1a8edd65586338a6cf", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AriaUI/Aria-UI/HEAD/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "resource_id": "url:085b6811ddd6f797475d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Aria-UI/Aria-UI-context-aware", "topics": ["grounding", "interaction-history", "action-prediction"], "updated_at": null}, "text": "Aria-UI context-aware checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nThe context-aware Aria-UI release that uses interaction context to support grounding in dynamic Android and desktop agent tasks.\nPlatforms: android, desktop\nTopics: grounding, interaction-history, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Aria-UI/Aria-UI-context-aware/tree/main\", \"gated\": false, \"hub_id\": \"Aria-UI/Aria-UI-context-aware\"}\n\nThe Hub manifest exposes twelve safetensors shards and model code/configuration, but no README card or model-license declaration. The official repository explicitly announces this context-aware release separately from the base model. Code or base-model licenses are not inherited.\n\nPrimary source: https://huggingface.co/Aria-UI/Aria-UI-context-aware\nCatalog record: https://awesomecua.github.io/resources/url-085b6811ddd6f797475d/"}
{"id": "url:0af07e00658a430cee6a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/yiye2023/GUIEnv", "citation_url": "https://awesomecua.github.io/resources/url-0af07e00658a430cee6a/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "GUIEnv-local SFT data", "qualifier": "approximately", "unit": "region-text QA pairs", "value": 700000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["OCR", "region-text grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIEnv/tree/main", "gated": false, "hub_id": "yiye2023/GUIEnv", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "15813be05dc76aed5f667875188bf9612e74ff2f4699d270dd0d1aeda86767f8", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIEnv"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIEnv"}], "name": "GUIEnv-local", "publication": null, "published_at": null, "record_sha256": "8c9e1f63bc16c8bbb8f2488dee50fea8758263a7e3e5af6365bd098d916dbb78", "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "resource_id": "url:0af07e00658a430cee6a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/yiye2023/GUIEnv", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "GUIEnv-local\nResource type: dataset\nReview status: curated\nScope: core\n\nThe released local-supervision portion of GUIEnv, pairing webpage regions and text for OCR and GUI grounding.\nPlatforms: web\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"population\": \"GUIEnv-local SFT data\", \"qualifier\": \"approximately\", \"unit\": \"region-text QA pairs\", \"value\": 700000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\"], \"formats\": [\"JSON\", \"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"bounding boxes\"], \"notes\": \"Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.\", \"splits\": [\"train\", \"test\"], \"supervision\": [\"OCR\", \"region-text grounding\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/yiye2023/GUIEnv/tree/main\", \"gated\": false, \"hub_id\": \"yiye2023/GUIEnv\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"https://huggingface.co/datasets/yiye2023/GUIEnv/raw/main/README.md\"], \"reason\": \"Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict.\"}}}\n\nPublic JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate. Only GUIEnv-local is marked released; the advertised 10M-example GUIEnv-global pretraining collection is still unchecked and is not counted here.\n\nPrimary source: https://huggingface.co/datasets/yiye2023/GUIEnv\nCatalog record: https://awesomecua.github.io/resources/url-0af07e00658a430cee6a/"}
{"id": "url:0b134b082a7bd507fc6a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.interactionmining.org/archive/rico", "citation_url": "https://awesomecua.github.io/resources/url-0b134b082a7bd507fc6a/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "original Rico downloads described on the project page", "qualifier": "more-than", "unit": "unique UI screens", "value": 66000}, {"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "original Rico collection described on the project page", "qualifier": "more-than", "unit": "Android apps", "value": 9300}, {"basis": "reported", "evidence_url": "https://www.interactionmining.org/archive/rico", "population": "crowd-worker collection before automated exploration", "qualifier": "exact", "unit": "human interaction traces", "value": 10811}], "evidence_urls": ["https://www.interactionmining.org/archive/rico", "https://danafergan.com/publications/deka2017rico.pdf"], "formats": ["PNG", "JSON", "CSV", "NumPy NPY", "GIF"], "modalities": ["screenshots", "Android view hierarchies", "gesture traces", "transition animations", "app/UI metadata"], "notes": "Project-page scope: screenshots paired with augmented view hierarchies, gesture coordinates, layout vectors and metadata. These observations are not action/task totals. The later semantic-annotation download is a related augmentation of screens, not an additional disjoint corpus. The retained original paper reports more than 72k screens from more than 9.7k apps while this project page says more than 66k screens and more than 9.3k apps; the difference is unresolved. Profile counts follow the project page and do not merge these populations. No train/validation/test partition or data license is established here; archives were not downloaded.", "splits": [], "supervision": ["UI structure and element properties", "recorded human interactions", "learned UI layout embeddings"]}, "modalities": ["screenshot", "view-hierarchy", "interaction-trace"], "reported_scale": {"comparison": ">", "unit": "unique UI screens", "value": 66000}, "sample_count": null, "sample_unit": "UI screens"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:29Z", "role": "primary", "sha256": "daa370b9aae5bce9f5171ea81fa4574c6155e04177af1db138529efd76377f89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "notes", "details.dataset_profile"], "url": "https://www.interactionmining.org/archive/rico"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["notes", "details.dataset_profile"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.interactionmining.org/archive/rico"}], "name": "Rico", "publication": null, "published_at": null, "record_sha256": "caf1b383312551fc61b1698721d1409eda13867b4b1f1d57ad711152ee23dfd7", "relations": [{"evidence_url": "https://www.interactionmining.org/archive/rico", "target_id": "url:b560b5d6abaab98483f6", "type": "documented_by"}], "resource_id": "url:0b134b082a7bd507fc6a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.interactionmining.org/archive/rico", "topics": ["screen-understanding"], "updated_at": null}, "text": "Rico\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile app screenshots, view hierarchies, and interaction traces for UI research.\nPlatforms: android\nTopics: screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://www.interactionmining.org/archive/rico\", \"population\": \"original Rico downloads described on the project page\", \"qualifier\": \"more-than\", \"unit\": \"unique UI screens\", \"value\": 66000}, {\"basis\": \"reported\", \"evidence_url\": \"https://www.interactionmining.org/archive/rico\", \"population\": \"original Rico collection described on the project page\", \"qualifier\": \"more-than\", \"unit\": \"Android apps\", \"value\": 9300}, {\"basis\": \"reported\", \"evidence_url\": \"https://www.interactionmining.org/archive/rico\", \"population\": \"crowd-worker collection before automated exploration\", \"qualifier\": \"exact\", \"unit\": \"human interaction traces\", \"value\": 10811}], \"evidence_urls\": [\"https://www.interactionmining.org/archive/rico\", \"https://danafergan.com/publications/deka2017rico.pdf\"], \"formats\": [\"PNG\", \"JSON\", \"CSV\", \"NumPy NPY\", \"GIF\"], \"modalities\": [\"screenshots\", \"Android view hierarchies\", \"gesture traces\", \"transition animations\", \"app/UI metadata\"], \"notes\": \"Project-page scope: screenshots paired with augmented view hierarchies, gesture coordinates, layout vectors and metadata. These observations are not action/task totals. The later semantic-annotation download is a related augmentation of screens, not an additional disjoint corpus. The retained original paper reports more than 72k screens from more than 9.7k apps while this project page says more than 66k screens and more than 9.3k apps; the difference is unresolved. Profile counts follow the project page and do not merge these populations. No train/validation/test partition or data license is established here; archives were not downloaded.\", \"splits\": [], \"supervision\": [\"UI structure and element properties\", \"recorded human interactions\", \"learned UI layout embeddings\"]}, \"modalities\": [\"screenshot\", \"view-hierarchy\", \"interaction-trace\"], \"reported_scale\": {\"comparison\": \">\", \"unit\": \"unique UI screens\", \"value\": 66000}, \"sample_unit\": \"UI screens\"}\n\nThe official project page cites the 2017 Rico paper. The project page reports more than 66,000 screens, while the original publication reports more than 72,000. These source-specific counts are kept separate; the reason for the difference is not established by these sources. App counts also differ: the official project page reports more than 9.3k Android apps and the retained published abstract more than 9.7k. Profile counts use the project-page scope; no reconciliation is inferred.\n\nPrimary source: https://www.interactionmining.org/archive/rico\nCatalog record: https://awesomecua.github.io/resources/url-0b134b082a7bd507fc6a/"}
{"id": "url:0b590d03a3d5103a79e4", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/vyokky/GUI-360", "citation_url": "https://awesomecua.github.io/resources/url-0b590d03a3d5103a79e4/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/vyokky/GUI-360/tree/main", "gated": false, "hub_id": "vyokky/GUI-360", "languages": [], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["desktop", "windows"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "e51af2845c2f31c7af112579b42c3ff9bf5a2c2356e80932e829bc2be1a341bd", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/vyokky/GUI-360"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/vyokky/GUI-360"}], "name": "GUI-360", "publication": null, "published_at": null, "record_sha256": "d45c01e263f84bc7498c761317410a4c93947dc93f53b471b8abd1433dd2bd57", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/vyokky/GUI-360", "target_id": "arxiv:2511.04307", "type": "documented_by"}], "resource_id": "url:0b590d03a3d5103a79e4", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/vyokky/GUI-360", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "GUI-360\nResource type: dataset\nReview status: curated\nScope: core\n\nDesktop GUI trajectory data supporting multiple interface understanding tasks.\nPlatforms: desktop, windows\nTopics: grounding, data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/vyokky/GUI-360/tree/main\", \"gated\": false, \"hub_id\": \"vyokky/GUI-360\", \"size_category_reported\": [\"1M<n<10M\"]}\n\nPrimary source: https://huggingface.co/datasets/vyokky/GUI-360\nCatalog record: https://awesomecua.github.io/resources/url-0b590d03a3d5103a79e4/"}
{"id": "url:0bdd6b5ebf2e1f248828", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.simular.ai/", "citation_url": "https://awesomecua.github.io/resources/url-0bdd6b5ebf2e1f248828/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "4cb358ab96e57a66d2b268a17963f92be249325fb0687a78985a6d9773ff5a81", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.simular.ai/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.simular.ai/"}], "name": "Simular", "publication": null, "published_at": null, "record_sha256": "2cea1cf162328e1c1c7076cc4ec64a86cf37f86d7b470f1c2186bebffad304b8", "relations": [{"evidence_url": "https://www.simular.ai/", "target_id": "organization:simular", "type": "published_by"}], "resource_id": "url:0bdd6b5ebf2e1f248828", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.simular.ai/", "topics": ["workflow"], "updated_at": null}, "text": "Simular\nResource type: product\nReview status: curated\nScope: core\n\nComputer-use products and research from Simular.\nPlatforms: desktop\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.simular.ai/\nCatalog record: https://awesomecua.github.io/resources/url-0bdd6b5ebf2e1f248828/"}
{"id": "url:0c5659ca9e59d942aea3", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1", "citation_url": "https://awesomecua.github.io/resources/url-0c5659ca9e59d942aea3/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/tree/main", "gated": false, "hub_id": "SenseLLM/GUI-Lasagne-L1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "224d6487de47cc63251c3b5c5c9f6e75e692b5ef0e12f0a7f71000da0b866e6f", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/SenseLLM/GUI-Lasagne-L1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["environments", "summary"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "270b3af0904a2a50d040948100a327458839dffa5083fc74bb030298f5ab3985", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1"}], "name": "GUI-Lasagne L1", "publication": null, "published_at": null, "record_sha256": "1c7763074254c27cbccaf4f848367a42a1f6b89a129dec1875583ebeceff0115", "relations": [{"evidence_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "resource_id": "url:0c5659ca9e59d942aea3", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1", "topics": ["grounding", "screen-understanding", "data-generation"], "updated_at": null}, "text": "GUI-Lasagne L1\nResource type: dataset\nReview status: curated\nScope: core\n\nThe published GUI-Lasagne training-data collection used for SpiritSight, distributed as dated archive shards for visual interface learning.\nPlatforms: unspecified\nTopics: grounding, screen-understanding, data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1/tree/main\", \"gated\": false, \"hub_id\": \"SenseLLM/GUI-Lasagne-L1\"}\n\nPublic archive shards are present. The dataset card declares Apache-2.0; it does not resolve rights in third-party interface content. The short card supplies no reliable sample count or explicit train/test partition, so these remain unspecified. The dataset card and archive inventory do not establish its platform distribution; the associated model's platform coverage is not inherited. An associated SpiritSight model card still describes dataset release as forthcoming; the independent dataset card and public archive manifest establish this release's availability.\n\nPrimary source: https://huggingface.co/datasets/SenseLLM/GUI-Lasagne-L1\nCatalog record: https://awesomecua.github.io/resources/url-0c5659ca9e59d942aea3/"}
{"id": "url:0dbb55f127da04d95222", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/agentsea/wave-ui", "citation_url": "https://awesomecua.github.io/resources/url-0dbb55f127da04d95222/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 63530}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "validation split", "qualifier": "exact", "unit": "rows", "value": 7944}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "rows", "value": 7938}], "evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md", "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "formats": ["Parquet"], "modalities": ["images", "text", "bounding boxes", "OCR metadata"], "notes": "Card-reported row counts are not unique-screen counts. The source field is important for the per-image license terms; the compilation does not relicense original images.", "splits": ["train", "validation", "test"], "supervision": ["instruction-to-target boxes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/agentsea/wave-ui/tree/main", "gated": false, "hub_id": "agentsea/wave-ui", "languages": [], "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"], "reason": "The compilation expressly does not relicense its images and lists differing source terms. No blanket permissive data license is appropriate."}}, "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["name", "kind", "identity", "release_status", "details", "summary"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "27ec13aa77dc55113e4d1a19c2269070e2d69d2fe8df8e2313721a1e3fc66bd9", "supports": ["details"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "63c6b3ce6a3d9c21a505865ceed698df2894bcb292590c59686e9477df867942", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["summary", "topics", "scope"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "76e5e458eb8788f2c2cfddd06cb944327f363ec70120e122ce93d1350cf82e19", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/agentsea/wave-ui"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": {"basis": "Publisher compilation license; individual images keep their source-dataset terms and receive no additional rights from this compilation.", "declared": "Wave UI Dataset License; per-image source terms retained", "retrieved_at": "2026-10-17T04:06:57Z", "url": "https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/agentsea/wave-ui"}], "name": "Wave UI", "publication": null, "published_at": null, "record_sha256": "c88ef38ed7c28fb8e9b476d4d98033eda8431a04f173d2f28eba8c84a8e9d954", "relations": [], "resource_id": "url:0dbb55f127da04d95222", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/agentsea/wave-ui", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "Wave UI\nResource type: dataset\nReview status: curated\nScope: core\n\nScreenshot-instruction records with target boxes and UI metadata such as OCR, source, element type and platform.\nPlatforms: unspecified\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md\", \"population\": \"train split\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 63530}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md\", \"population\": \"validation split\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 7944}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md\", \"population\": \"test split\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 7938}], \"evidence_urls\": [\"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/README.md\", \"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"images\", \"text\", \"bounding boxes\", \"OCR metadata\"], \"notes\": \"Card-reported row counts are not unique-screen counts. The source field is important for the per-image license terms; the compilation does not relicense original images.\", \"splits\": [\"train\", \"validation\", \"test\"], \"supervision\": [\"instruction-to-target boxes\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/agentsea/wave-ui/tree/main\", \"gated\": false, \"hub_id\": \"agentsea/wave-ui\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/agentsea/wave-ui/raw/main/LICENSE.md\"], \"reason\": \"The compilation expressly does not relicense its images and lists differing source terms. No blanket permissive data license is appropriate.\"}}, \"modalities\": [\"image\", \"text\"]}\n\nTrain, validation and test Parquet files are publicly listed. The compilation license explicitly preserves each image’s original source terms and grants no extra rights to those images. The card exposes a platform column without listing its values, so operating-system coverage remains unspecified.\n\nPrimary source: https://huggingface.co/datasets/agentsea/wave-ui\nCatalog record: https://awesomecua.github.io/resources/url-0dbb55f127da04d95222/"}
{"id": "url:0fdeb1cf5f49f84820e6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data", "citation_url": "https://awesomecua.github.io/resources/url-0fdeb1cf5f49f84820e6/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "screenshots", "value": 270000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "SeeClick-derived web subset", "qualifier": "more-than", "unit": "UI elements", "value": 3000000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md", "population": "FineWeb-derived subset", "qualifier": "at-least", "unit": "images", "value": 1600000}], "evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "referring expressions", "bounding boxes"], "notes": "These are subset-specific counts, not a total for the full OS-Atlas release. Raw boxes use 0–1 coordinates; the documented training wrapper scales them to 0–1000.", "splits": [], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:40:46Z", "role": "primary", "sha256": "19f28b4bc0798e28180b968010c0f7c8dfed2dbfb24b122b958476b871d2e4fd", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "c2c5f7b3f07d92349ffd8256792599c856ed77c89108c45dbaef48a4699abe20", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data"}], "name": "OS-Atlas Data", "publication": null, "published_at": null, "record_sha256": "9bfc1d536d9adbcf8cbce1f1fcf5ae407e108467344d0fa2fbcf5fe203878d5c", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Atlas-data", "target_id": "arxiv:2410.23218", "type": "documented_by"}], "resource_id": "url:0fdeb1cf5f49f84820e6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "OS-Atlas Data\nResource type: dataset\nReview status: curated\nScope: core\n\nGUI grounding pre-training data for OS-Atlas in desktop (Linux, macOS, Windows), mobile and web domains, with element boxes stored as relative left, top, right and bottom ratios.\nPlatforms: cross-platform\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md\", \"population\": \"SeeClick-derived web subset\", \"qualifier\": \"more-than\", \"unit\": \"screenshots\", \"value\": 270000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md\", \"population\": \"SeeClick-derived web subset\", \"qualifier\": \"more-than\", \"unit\": \"UI elements\", \"value\": 3000000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md\", \"population\": \"FineWeb-derived subset\", \"qualifier\": \"at-least\", \"unit\": \"images\", \"value\": 1600000}], \"evidence_urls\": [\"https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/raw/main/README.md\"], \"formats\": [\"JSON annotations\", \"image archives\"], \"modalities\": [\"screenshots\", \"referring expressions\", \"bounding boxes\"], \"notes\": \"These are subset-specific counts, not a total for the full OS-Atlas release. Raw boxes use 0–1 coordinates; the documented training wrapper scales them to 0–1000.\", \"splits\": [], \"supervision\": [\"element grounding\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Atlas-data\"}\n\nPrimary source: https://huggingface.co/datasets/OS-Copilot/OS-Atlas-data\nCatalog record: https://awesomecua.github.io/resources/url-0fdeb1cf5f49f84820e6/"}
{"id": "url:13b9442bcb6a8b984c4f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/biglab/webui-all", "citation_url": "https://awesomecua.github.io/resources/url-13b9442bcb6a8b984c4f/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/biglab/webui-all/tree/main", "gated": false, "hub_id": "biglab/webui-all", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "70928db908760c1561bfc8e0cd7d925e0103ffddef4676fbaa7c1756a804b99a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/biglab/webui-all"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:53Z", "role": "primary", "sha256": "93bd04914735f748d687b71597970a79d95ae8636e7ed75d267ebd6954d11a8f", "supports": ["notes"], "url": "https://huggingface.co/datasets/biglab/webui-all/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "70928db908760c1561bfc8e0cd7d925e0103ffddef4676fbaa7c1756a804b99a", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/biglab/webui-all"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "other", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/biglab/webui-all"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/biglab/webui-all"}], "name": "WebUI", "publication": null, "published_at": null, "record_sha256": "24fcdea6c592ff9c63009e2e3a1c92300e872235a2d8747c4cf082e42a57b76b", "relations": [], "resource_id": "url:13b9442bcb6a8b984c4f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/biglab/webui-all", "topics": ["screen-understanding"], "updated_at": null}, "text": "WebUI\nResource type: dataset\nReview status: curated\nScope: core\n\nWeb screenshot and interface data for UI research.\nPlatforms: web\nTopics: screen-understanding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/biglab/webui-all/tree/main\", \"gated\": false, \"hub_id\": \"biglab/webui-all\"}\n\nThe card links the WebUI project's ACM publication (DOI 10.1145/3544548.3581158) and website, not an arXiv record or a catalogued repository, so no paper or family relation is recorded.\n\nPrimary source: https://huggingface.co/datasets/biglab/webui-all\nCatalog record: https://awesomecua.github.io/resources/url-13b9442bcb6a8b984c4f/"}
{"id": "url:14449fe321e0a1ca34ea", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E", "citation_url": "https://awesomecua.github.io/resources/url-14449fe321e0a1ca34ea/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "https://huggingface.co/api/datasets/henryhe0123/PC-Agent-E"], "formats": ["ZIP package"], "modalities": ["GUI screenshots", "action trajectories", "generated reasoning"], "notes": "Public manifest exposes data.zip; the internal file schema and number of examples remain unverified. Supervision describes the publisher’s refinement/thought-completion/trajectory-boost pipeline, not an independently sampled data-quality audit.", "splits": ["training"], "supervision": ["human demonstration refinement", "generated action reasoning", "trajectory augmentation"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/tree/4f64c7b055154bed9ffea677aa86aeb51a01cb73", "gated": false, "hub_id": "henryhe0123/PC-Agent-E", "hub_revision": "4f64c7b055154bed9ffea677aa86aeb51a01cb73", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/GAIR-NLP/PC-Agent-E"}, "environments": ["desktop", "windows", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:13Z", "role": "primary", "sha256": "3ffb8c843c6acb8c2d5bbba11ba49d6b36cd566553a89974b6252a2bfd289d5d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "66f70ee2704cdf2e8847cfe04be05d9e14b3a3c1bee953e744f8209d6c9c2b85", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/henryhe0123/PC-Agent-E"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "4dfd1f3ec40f7ff3c94896955216cb60a25114b320caff29e909d296d375d6ef", "supports": ["relations", "licenses.data"], "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:27:13Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E"}], "name": "PC-Agent-E training data", "publication": null, "published_at": null, "record_sha256": "f8498127f3c2f753899711b7ca2cff447800efc1f0df711c2da431727c93656f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md", "target_id": "arxiv:2505.13909", "type": "documented_by"}], "resource_id": "url:14449fe321e0a1ca34ea", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/henryhe0123/PC-Agent-E", "topics": ["data-generation", "imitation-learning"], "updated_at": null}, "text": "PC-Agent-E training data\nResource type: dataset\nReview status: curated\nScope: core\n\nReleased dataset used to train PC-Agent-E, linked by the official pipeline for refining human trajectories and synthesizing action reasoning.\nPlatforms: desktop, windows, linux\nTopics: data-generation, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://raw.githubusercontent.com/GAIR-NLP/PC-Agent-E/HEAD/README.md\", \"https://huggingface.co/api/datasets/henryhe0123/PC-Agent-E\"], \"formats\": [\"ZIP package\"], \"modalities\": [\"GUI screenshots\", \"action trajectories\", \"generated reasoning\"], \"notes\": \"Public manifest exposes data.zip; the internal file schema and number of examples remain unverified. Supervision describes the publisher’s refinement/thought-completion/trajectory-boost pipeline, not an independently sampled data-quality audit.\", \"splits\": [\"training\"], \"supervision\": [\"human demonstration refinement\", \"generated action reasoning\", \"trajectory augmentation\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/henryhe0123/PC-Agent-E/tree/4f64c7b055154bed9ffea677aa86aeb51a01cb73\", \"gated\": false, \"hub_id\": \"henryhe0123/PC-Agent-E\", \"hub_revision\": \"4f64c7b055154bed9ffea677aa86aeb51a01cb73\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/henryhe0123/PC-Agent-E/raw/4f64c7b055154bed9ffea677aa86aeb51a01cb73/README.md\"], \"reason\": \"Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.\"}}, \"repository_url\": \"https://github.com/GAIR-NLP/PC-Agent-E\"}\n\nHub release contains data.zip. The archive was not downloaded and its internal schema/example total were not independently inspected. Training-purpose and platform descriptions come from the official pipeline README; the brief card establishes the paper association and MIT data declaration.\n\nPrimary source: https://huggingface.co/datasets/henryhe0123/PC-Agent-E\nCatalog record: https://awesomecua.github.io/resources/url-14449fe321e0a1ca34ea/"}
{"id": "url:15ed5fd3cd950d837f98", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.skyvern.com/", "citation_url": "https://awesomecua.github.io/resources/url-15ed5fd3cd950d837f98/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7a8813c174043d1a27cb3766f77c90beb1a08eeedee935264d6237d6a57672a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.skyvern.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.skyvern.com/"}], "name": "Skyvern Cloud", "publication": null, "published_at": null, "record_sha256": "e6213b972f77a22d9a5c1c95b988742b9fc4a9e775544548831f4e6f3f15d7c0", "relations": [{"evidence_url": "https://www.skyvern.com/", "target_id": "organization:skyvern", "type": "published_by"}], "resource_id": "url:15ed5fd3cd950d837f98", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.skyvern.com/", "topics": ["workflow"], "updated_at": null}, "text": "Skyvern Cloud\nResource type: product\nReview status: curated\nScope: core\n\nHosted Skyvern service for AI browser workflows such as invoice downloads and form filling, with built-in CAPTCHA and 2FA handling, proxy geotargeting, run explanations and JSON or CSV output.\nPlatforms: web\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.skyvern.com/\nCatalog record: https://awesomecua.github.io/resources/url-15ed5fd3cd950d837f98/"}
{"id": "url:16083a0225a60882c704", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Hcompany/Holo4-35B-A3B", "citation_url": "https://awesomecua.github.io/resources/url-16083a0225a60882c704/", "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo4-35B-A3B"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "7a41bdbfb4a300302a7d3633f12962505d56ba874a37292afccbd48b1bec8287", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo4-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:11:45Z", "role": "primary", "sha256": "4f1aa5006aa9adaeb5ac549e5a5f8ea164352714a00528b1afb8602812d45b44", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:11:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:11:45Z", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo4-35B-A3B"}], "name": "Holo4 35B-A3B BF16", "publication": null, "published_at": null, "record_sha256": "51b8972cbe756e9b232a04d0003935c168625a5cb50f755ca3aab3aefcc51619", "relations": [], "resource_id": "url:16083a0225a60882c704", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Hcompany/Holo4-35B-A3B", "topics": ["grounding", "action-prediction", "tool-integration"], "updated_at": null}, "text": "Holo4 35B-A3B BF16\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen3.6-based computer-use checkpoint designed to request GUI actions, code execution and tool calls through an external agent harness.\nPlatforms: desktop, web\nTopics: grounding, action-prediction, tool-integration\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Hcompany/Holo4-35B-A3B/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/Holo4-35B-A3B\"}\n\nThe public manifest contains 26 safetensors shards for the BF16 checkpoint and the card declares Apache-2.0. Quantized FP8/NVFP4/GGUF variants are not split into duplicate entries here. The card also contains demos and scores for Holo4-27B; those are not assigned to this 35B-A3B checkpoint. No training corpus or runtime was downloaded.\n\nPrimary source: https://huggingface.co/Hcompany/Holo4-35B-A3B\nCatalog record: https://awesomecua.github.io/resources/url-16083a0225a60882c704/"}
{"id": "url:1611de0e6d606165d12d", "metadata": {"aliases": ["url:1ab38d1d6a4c46a8eb3e", "OpenMobile-2/OpenMobile-27B", "https://huggingface.co/OpenMobile-2/OpenMobile-27B"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B", "citation_url": "https://awesomecua.github.io/resources/url-1611de0e6d606165d12d/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.6-27B", "gated": false, "hub_revision": "3d844db8c20421524a5012837b151e444ad35ae6", "listed_weight_files": ["model-00001-of-00012.safetensors", "model-00002-of-00012.safetensors", "model-00003-of-00012.safetensors", "model-00004-of-00012.safetensors", "model-00005-of-00012.safetensors", "model-00006-of-00012.safetensors", "model-00007-of-00012.safetensors", "model-00008-of-00012.safetensors", "model-00009-of-00012.safetensors", "model-00010-of-00012.safetensors", "model-00011-of-00012.safetensors", "model-00012-of-00012.safetensors"]}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "ba760ab75e997941dabf0845b34e5fb83daeb83a5cfa16b13929477cfc61a31c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "c7e388efe1ecfd415b7b5af7f296e62646a0a1380037f1e57ad1a43c26463b9a", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-27B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "ba760ab75e997941dabf0845b34e5fb83daeb83a5cfa16b13929477cfc61a31c", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:11Z", "role": "primary", "sha256": "f669181b7f1b92bbbcb8ab71f0d15d8c3cf93b3665d73d20e9b9f0dc3a36e158", "supports": ["aliases"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-27B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:12Z", "role": "primary", "sha256": "f669181b7f1b92bbbcb8ab71f0d15d8c3cf93b3665d73d20e9b9f0dc3a36e158", "supports": ["aliases"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-27B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B"}], "name": "OpenMobile-2 27B", "publication": null, "published_at": null, "record_sha256": "0086ce1cd137b8101b167d00e05eedd970cbb5d8740de848e9285bcf092c3df8", "relations": [{"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "github:openmobile2/openmobile2-code", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "url:546e6c95825d1318c98d", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B/raw/main/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "resource_id": "url:1611de0e6d606165d12d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B", "topics": ["mobile-navigation", "tool-integration"], "updated_at": null}, "text": "OpenMobile-2 27B\nResource type: model\nReview status: curated\nScope: core\n\nMobile-agent checkpoint based on Qwen3.6-27B for GUI and app-native tool interaction.\nRecorded aliases: url:1ab38d1d6a4c46a8eb3e, OpenMobile-2/OpenMobile-27B, https://huggingface.co/OpenMobile-2/OpenMobile-27B\nPlatforms: android, web\nTopics: mobile-navigation, tool-integration\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.6-27B\", \"gated\": false, \"hub_revision\": \"3d844db8c20421524a5012837b151e444ad35ae6\", \"listed_weight_files\": [\"model-00001-of-00012.safetensors\", \"model-00002-of-00012.safetensors\", \"model-00003-of-00012.safetensors\", \"model-00004-of-00012.safetensors\", \"model-00005-of-00012.safetensors\", \"model-00006-of-00012.safetensors\", \"model-00007-of-00012.safetensors\", \"model-00008-of-00012.safetensors\", \"model-00009-of-00012.safetensors\", \"model-00010-of-00012.safetensors\", \"model-00011-of-00012.safetensors\", \"model-00012-of-00012.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.\n\nPrimary source: https://huggingface.co/OpenMobile-2/OpenMobile-2-27B\nCatalog record: https://awesomecua.github.io/resources/url-1611de0e6d606165d12d/"}
{"id": "url:16184830c0f4bdb146fa", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset", "citation_url": "https://awesomecua.github.io/resources/url-16184830c0f4bdb146fa/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md", "population": "public example subset", "qualifier": "exact", "unit": "training examples", "value": 1000}], "evidence_urls": ["https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md", "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"], "formats": ["JSONL", "Parquet"], "modalities": ["images", "text", "actions"], "notes": "The count is explicitly reported by the publisher, not inferred from the filename. Full training data remains outside this release.", "splits": ["training example subset"], "supervision": ["GUI actions"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/tree/main", "gated": false, "hub_id": "mPLUG/UI_S1_dataset"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "d826463e2aa159f083dc74869bfabb8abb5b8ea4f2cdf4aeb7676172b18aaaa5", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "cdfbcbf23a3b1b4a609e1943e80dcf144ca1224a74ebf2308fde4fcb4326ab3c", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mPLUG/UI_S1_dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "f668c890a5a87953a243715d0d47367c23f0c99b524d80f48bbdde8163f3e942", "supports": ["summary", "environments", "details.dataset_profile", "topics", "scope"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "d826463e2aa159f083dc74869bfabb8abb5b8ea4f2cdf4aeb7676172b18aaaa5", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset"}], "name": "UI-S1 released training examples", "publication": null, "published_at": null, "record_sha256": "f8e8657c2d21297426886d0662536e46274f9c2666ce575237e4f3f1dd4a230d", "relations": [{"evidence_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md", "target_id": "arxiv:2509.11543", "type": "documented_by"}], "resource_id": "url:16184830c0f4bdb146fa", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/mPLUG/UI_S1_dataset", "topics": ["reinforcement-learning", "data-generation", "mobile-navigation"], "updated_at": null}, "text": "UI-S1 released training examples\nResource type: dataset\nReview status: curated\nScope: core\n\nA public example subset of the UI-S1 mobile training data, packaged as JSONL annotations and Parquet images.\nPlatforms: android\nTopics: reinforcement-learning, data-generation, mobile-navigation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md\", \"population\": \"public example subset\", \"qualifier\": \"exact\", \"unit\": \"training examples\", \"value\": 1000}], \"evidence_urls\": [\"https://huggingface.co/datasets/mPLUG/UI_S1_dataset/raw/main/README.md\", \"https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md\"], \"formats\": [\"JSONL\", \"Parquet\"], \"modalities\": [\"images\", \"text\", \"actions\"], \"notes\": \"The count is explicitly reported by the publisher, not inferred from the filename. Full training data remains outside this release.\", \"splits\": [\"training example subset\"], \"supervision\": [\"GUI actions\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/mPLUG/UI_S1_dataset/tree/main\", \"gated\": false, \"hub_id\": \"mPLUG/UI_S1_dataset\"}\n\nThe release contains train_1000.jsonl and images.parquet. The official README explicitly describes 1,000 released examples and calls the release only part of the training dataset. Apache-2.0 is declared by the dataset card; this is not the complete training corpus.\n\nPrimary source: https://huggingface.co/datasets/mPLUG/UI_S1_dataset\nCatalog record: https://awesomecua.github.io/resources/url-16184830c0f4bdb146fa/"}
{"id": "url:16a51d9e725ad5a5c629", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Hcompany/Holo2-4B", "citation_url": "https://awesomecua.github.io/resources/url-16a51d9e725ad5a5c629/", "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo2-4B/tree/main", "gated": false, "hub_id": "Hcompany/Holo2-4B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bb870c78f3075002ae42a7d1f392fda048fa31a58c865290b2a9b4a1137c6dd3", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo2-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "32232445026dc93113beb6b23d3e43da44d6288a67501f8f0023a78a43efb100", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo2-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo2-4B"}], "name": "Holo2 4B", "publication": null, "published_at": null, "record_sha256": "0581a5f0907443e47abc56fbc0c12f071a69f3c8ea55f04929e0c00d16ef7f13", "relations": [], "resource_id": "url:16a51d9e725ad5a5c629", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Hcompany/Holo2-4B", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null}, "text": "Holo2 4B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen3-VL-based checkpoint for navigation and multi-step GUI actions across web, desktop and mobile interfaces.\nPlatforms: web, desktop, cross-platform\nTopics: grounding, planning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Hcompany/Holo2-4B/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/Holo2-4B\"}\n\nTwo public safetensors shards are present and this 4B model card declares Apache-2.0. The same card assigns different terms to the 30B family member, so the 4B declaration is not generalized to the family. Scores and proprietary training data are not treated as independently verified or publicly released.\n\nPrimary source: https://huggingface.co/Hcompany/Holo2-4B\nCatalog record: https://awesomecua.github.io/resources/url-16a51d9e725ad5a5c629/"}
{"id": "url:17729e1c532ba7043404", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants", "citation_url": "https://awesomecua.github.io/resources/url-17729e1c532ba7043404/", "details": {"benchmark_protocol": {"actions": ["target localization", "identification of impossible requests"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "population": "instruction-variant design", "qualifier": "exact", "unit": "instruction types", "value": 4}], "evaluation": "Publisher supplies original, action, description and negative instruction types; no common end-to-end success metric is asserted.", "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md", "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"], "focus": ["instruction-form robustness", "negative grounding requests"], "notes": "Reuses ScreenSpot-v2 imagery. Image-file count is not treated as a count of independent screenshots or tasks.", "observations": ["screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/tree/main", "gated": false, "hub_id": "likaixin/ScreenSpot-v2-variants"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "7fb2bc9499d8a2b37b663c116af49d11bba493e6767d76a5305fad93257bebf2", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "dea98efa75e063eee2d6bfb5d46da94b41df3d779e2fb0bdaea7866534b5c469", "supports": ["identity", "release_status", "details", "environments"], "url": "https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants"}], "name": "ScreenSpot-v2 instruction variants", "publication": null, "published_at": null, "record_sha256": "d4e8341bdf7ec9562a9370d337a70c3c29d49d65cddbb4750a0a5f3125d8b434", "relations": [], "resource_id": "url:17729e1c532ba7043404", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants", "topics": ["grounding", "evaluation", "robustness"], "updated_at": null}, "text": "ScreenSpot-v2 instruction variants\nResource type: benchmark\nReview status: curated\nScope: core\n\nInstruction variants over ScreenSpot-v2 screenshots that distinguish original commands, explicit actions, target descriptions and impossible operations.\nPlatforms: cross-platform\nTopics: grounding, evaluation, robustness\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"target localization\", \"identification of impossible requests\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md\", \"population\": \"instruction-variant design\", \"qualifier\": \"exact\", \"unit\": \"instruction types\", \"value\": 4}], \"evaluation\": \"Publisher supplies original, action, description and negative instruction types; no common end-to-end success metric is asserted.\", \"evidence_urls\": [\"https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/raw/main/README.md\", \"https://huggingface.co/api/datasets/likaixin/ScreenSpot-v2-variants\"], \"focus\": [\"instruction-form robustness\", \"negative grounding requests\"], \"notes\": \"Reuses ScreenSpot-v2 imagery. Image-file count is not treated as a count of independent screenshots or tasks.\", \"observations\": [\"screenshots\", \"instructions\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/tree/main\", \"gated\": false, \"hub_id\": \"likaixin/ScreenSpot-v2-variants\"}\n\nPublic image files and desktop/mobile/web annotation JSON files are listed. This is a changed instruction set over an existing screenshot collection, not an independent new screenshot corpus. No data license or precise example count is declared in the card; no paper relationship is inferred from its generic project/PDF links.\n\nPrimary source: https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants\nCatalog record: https://awesomecua.github.io/resources/url-17729e1c532ba7043404/"}
{"id": "url:1c4ee99a9a55ae71a7ac", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/mPLUG/UI-S1-7B", "citation_url": "https://awesomecua.github.io/resources/url-1c4ee99a9a55ae71a7ac/", "details": {"disabled": false, "download_url": "https://huggingface.co/mPLUG/UI-S1-7B/tree/main", "gated": false, "hub_id": "mPLUG/UI-S1-7B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "341abd3833131080682b27fe0b79e310d7cf9465602b77b8e399e45200552d61", "supports": ["name", "kind", "relations", "licenses.weights"], "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "d0c25fa2fad8372ffc8b51b63701c7e5548cca08a612c0081c4c317f9532894d", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/mPLUG/UI-S1-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "f668c890a5a87953a243715d0d47367c23f0c99b524d80f48bbdde8163f3e942", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/UI-S1/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "341abd3833131080682b27fe0b79e310d7cf9465602b77b8e399e45200552d61", "supports": ["licenses.weights"], "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/mPLUG/UI-S1-7B"}], "name": "UI-S1 7B", "publication": null, "published_at": null, "record_sha256": "c5da8cd438a7c5a49dcf38ba987e307d071c8051ba42b6f4a2efa51fb6a061dc", "relations": [{"evidence_url": "https://huggingface.co/mPLUG/UI-S1-7B/raw/main/README.md", "target_id": "arxiv:2509.11543", "type": "documented_by"}], "resource_id": "url:1c4ee99a9a55ae71a7ac", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/mPLUG/UI-S1-7B", "topics": ["reinforcement-learning", "mobile-navigation", "action-prediction"], "updated_at": null}, "text": "UI-S1 7B\nResource type: model\nReview status: curated\nScope: core\n\nA mobile GUI checkpoint trained with semi-online reinforcement learning that uses offline trajectories to improve multi-turn action prediction.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/mPLUG/UI-S1-7B/tree/main\", \"gated\": false, \"hub_id\": \"mPLUG/UI-S1-7B\"}\n\nFour public safetensors shards are present. The card declares Apache-2.0 for the weights. The project publishes only a portion of its training data; no complete training corpus or independent reproduction is inferred.\n\nPrimary source: https://huggingface.co/mPLUG/UI-S1-7B\nCatalog record: https://awesomecua.github.io/resources/url-1c4ee99a9a55ae71a7ac/"}
{"id": "url:1d5b85321b98765141fc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xwk123/MobileBench-v2", "citation_url": "https://awesomecua.github.io/resources/url-1d5b85321b98765141fc/", "details": {"benchmark_protocol": {"actions": ["mobile UI actions", "clarification responses"], "counts": [], "evaluation": "Paper defines multi-path and noise/ambiguity evaluations; the public card identifies which files are full common-task sets and which are samples.", "evidence_urls": ["https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md", "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"], "focus": ["offline multi-path evaluation", "noisy interfaces", "ambiguous instructions"], "notes": "Release completeness differs by split. No full-benchmark sample total is inferred from the archive manifest.", "observations": ["mobile screenshots", "XML UI state", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/xwk123/MobileBench-v2/tree/main", "gated": false, "hub_id": "xwk123/MobileBench-v2", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md"], "reason": "The same first-party card has an Apache-2.0 metadata header but explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to the dataset in its License section. Preserve both declarations and leave a single data license unresolved instead of displaying the permissive header as the data permission."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "13424be795c7b75b8c2195469c1252a7b3ec633e8bb31f1f6d7e44476611cbe9", "supports": ["name", "kind", "summary", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "247b1766b8cbdb2943c4602d2667f9f06dec8b84f0d24457080d499177475f22", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/xwk123/MobileBench-v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:42Z", "role": "primary", "sha256": "afe4e75ffb8a5c44f4c088a90ffa5f53553036f514aa768f5aee0c5aa38d731e", "supports": ["relations", "summary"], "url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["environments", "notes"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xwk123/MobileBench-v2"}], "name": "Mobile-Bench-v2 released data", "publication": null, "published_at": null, "record_sha256": "8c4fae2b9c4ee19766e6258157ce91e44742cdd19bd32bee2117913e325fa20c", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50", "target_id": "arxiv:2505.11891", "type": "documented_by"}], "resource_id": "url:1d5b85321b98765141fc", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xwk123/MobileBench-v2", "topics": ["evaluation", "mobile-navigation", "robustness"], "updated_at": null}, "text": "Mobile-Bench-v2 released data\nResource type: benchmark\nReview status: curated\nScope: core\n\nOffline mobile-agent task data spanning common tasks and released noisy, contaminated and ambiguous-instruction subsets.\nPlatforms: android\nTopics: evaluation, mobile-navigation, robustness\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"mobile UI actions\", \"clarification responses\"], \"counts\": [], \"evaluation\": \"Paper defines multi-path and noise/ambiguity evaluations; the public card identifies which files are full common-task sets and which are samples.\", \"evidence_urls\": [\"https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md\", \"https://export.arxiv.org/api/query?id_list=2503.16788%2C2503.17709%2C2503.18492%2C2503.19537%2C2503.20978%2C2503.21620%2C2503.23434%2C2504.01382%2C2504.07981%2C2504.10445%2C2504.10458%2C2504.11257%2C2504.11281%2C2504.12679%2C2504.13805%2C2504.13865%2C2504.14239%2C2504.18575%2C2504.19838%2C2504.20464%2C2505.10924%2C2505.11891%2C2505.12299%2C2505.12632%2C2505.13227%2C2505.13887%2C2505.13909%2C2505.14289%2C2505.15810%2C2505.16282%2C2505.16827%2C2505.18829%2C2505.19897%2C2505.21496%2C2505.21499%2C2505.21936%2C2505.22648%2C2505.23762%2C2505.24838%2C2506.00618%2C2506.01391%2C2506.01968%2C2506.02456%2C2506.03143%2C2506.03569%2C2506.04135%2C2506.04614%2C2506.07672%2C2506.08012%2C2506.08972&max_results=50\"], \"focus\": [\"offline multi-path evaluation\", \"noisy interfaces\", \"ambiguous instructions\"], \"notes\": \"Release completeness differs by split. No full-benchmark sample total is inferred from the archive manifest.\", \"observations\": [\"mobile screenshots\", \"XML UI state\", \"instructions\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xwk123/MobileBench-v2/tree/main\", \"gated\": false, \"hub_id\": \"xwk123/MobileBench-v2\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/xwk123/MobileBench-v2/raw/main/README.md\"], \"reason\": \"The same first-party card has an Apache-2.0 metadata header but explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to the dataset in its License section. Preserve both declarations and leave a single data license unresolved instead of displaying the permissive header as the data permission.\"}}}\n\nThe card distinguishes complete common-simple/common-complex JSON files from sampled noisy and ambiguous subsets; common tasks also require the separately cataloged Mobile3M dataset. Card metadata says Apache-2.0, while the prose explicitly assigns Apache-2.0 to code and CC-BY-NC-SA-4.0 to data. Do not apply the permissive header to dataset payloads, and do not describe every announced split as fully released.\n\nPrimary source: https://huggingface.co/datasets/xwk123/MobileBench-v2\nCatalog record: https://awesomecua.github.io/resources/url-1d5b85321b98765141fc/"}
{"id": "url:1e6b8682e0f1ddc3d36f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browserless.io/", "citation_url": "https://awesomecua.github.io/resources/url-1e6b8682e0f1ddc3d36f/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "e044e75ec1c2a74f2bee8c02b6efae05c399ec3537e34dd861dcf4a0a25dc182", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browserless.io/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserless.io/"}], "name": "Browserless", "publication": null, "published_at": null, "record_sha256": "b56b3acb304f305a9ad89d1a03d5b1a82101772f0a5752e78c2469fb7c6ef444", "relations": [{"evidence_url": "https://www.browserless.io/", "target_id": "organization:browserless", "type": "published_by"}], "resource_id": "url:1e6b8682e0f1ddc3d36f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browserless.io/", "topics": ["sandbox"], "updated_at": null}, "text": "Browserless\nResource type: product\nReview status: curated\nScope: core\n\nHosted browser infrastructure for automation applications.\nPlatforms: web\nTopics: sandbox\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.browserless.io/\nCatalog record: https://awesomecua.github.io/resources/url-1e6b8682e0f1ddc3d36f/"}
{"id": "url:23ad240757185c6f7eac", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B", "citation_url": "https://awesomecua.github.io/resources/url-23ad240757185c6f7eac/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen2.5-VL-3B-Instruct", "gated": false, "hub_revision": "514c49e398fed1203e5c4192be9860131681299e", "license_name": "qwen-research", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "c580cb5bd811285944337c4ebb461fa498bba09e6082b70e120ceb83e302fcce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "399895cd711d703ff4e016417ca14d13b76988c5625ff120e7083991eb0bd8f2", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B"}], "name": "Seg2Click Qwen2.5-VL 3B", "publication": null, "published_at": null, "record_sha256": "c50ee02866804cd5b69d4e84565e123418ba5a4a108b6a2ff5a3c387e6862028", "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "url:23ad240757185c6f7eac", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B", "topics": ["grounding"], "updated_at": null}, "text": "Seg2Click Qwen2.5-VL 3B\nResource type: model\nReview status: curated\nScope: core\n\nSeg2Click checkpoint based on Qwen2.5-VL-3B with a custom action-mask decoder.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen2.5-VL-3B-Instruct\", \"gated\": false, \"hub_revision\": \"514c49e398fed1203e5c4192be9860131681299e\", \"license_name\": \"qwen-research\", \"listed_weight_files\": [\"model.safetensors\", \"phase1_mask_decoder.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference. The card names the Qwen Research license, unlike the Apache-2.0 declarations on the Qwen3.5 variants.\n\nPrimary source: https://huggingface.co/jcwang0602/Seg2Click-Qwen2.5-VL-3B\nCatalog record: https://awesomecua.github.io/resources/url-23ad240757185c6f7eac/"}
{"id": "url:240a05c02f85d4da60dc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://openai.com/index/introducing-chatgpt-atlas/", "citation_url": "https://awesomecua.github.io/resources/url-240a05c02f85d4da60dc/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:40:54Z", "role": "primary", "sha256": "fa71ad6addd026ecd017059219c2457baf6b99c5c94f67350c0f9637c459c852", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-chatgpt-atlas/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:40:54Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-chatgpt-atlas/"}], "name": "ChatGPT Atlas", "publication": null, "published_at": null, "record_sha256": "a43c8127e401d1166bf11d01c69c327f0e56412e9a6b5a5c0c3579b78bef9df2", "relations": [{"evidence_url": "https://openai.com/index/introducing-chatgpt-atlas/", "target_id": "organization:openai", "type": "published_by"}], "resource_id": "url:240a05c02f85d4da60dc", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openai.com/index/introducing-chatgpt-atlas/", "topics": ["web-navigation"], "updated_at": null}, "text": "ChatGPT Atlas\nResource type: product\nReview status: curated\nScope: core\n\nOpenAI browser product integrating ChatGPT and agent capabilities.\nPlatforms: web\nTopics: web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://openai.com/index/introducing-chatgpt-atlas/\nCatalog record: https://awesomecua.github.io/resources/url-240a05c02f85d4da60dc/"}
{"id": "url:243ab00712ba19cc2089", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350", "citation_url": "https://awesomecua.github.io/resources/url-243ab00712ba19cc2089/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "packaged tasks/videos", "qualifier": "exact", "unit": "demonstrations", "value": 350}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "final packaged images", "qualifier": "exact", "unit": "evidence images", "value": 8388}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "population": "packaged action traces", "qualifier": "exact", "unit": "action records", "value": 11951}], "evidence_urls": ["https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"], "formats": ["MP4", "PNG", "JSON", "JSONL", "commented key logs"], "modalities": ["screen video", "screenshots", "input events", "task text"], "notes": "The OSWorld subset was selected post hoc using model outcomes; no train/test split is supplied. Coordinate calibration, task success, and complete anonymization are not established by package-format checks.", "splits": [], "supervision": ["human workflow demonstrations", "skill induction"]}, "download_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/tree/main", "modalities": ["video", "image", "text", "input-events"], "sample_count": 350, "sample_unit": "task demonstrations / videos"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:47:49Z", "role": "primary", "sha256": "ae14cfe91c8c9a79edf1f306273a2ee9d1790b1dc5be0cd2d89e93df09db42f0", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:47:49Z", "role": "primary", "sha256": "ae14cfe91c8c9a79edf1f306273a2ee9d1790b1dc5be0cd2d89e93df09db42f0", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:47:49Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T10:47:49Z", "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350"}], "name": "WorkDemo-350", "publication": null, "published_at": null, "record_sha256": "793cc54896ad3ade43951ad0843722553695879667f70eade1e9615e278d88f0", "relations": [{"evidence_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md", "target_id": "arxiv:2610.12984", "type": "documented_by"}], "resource_id": "url:243ab00712ba19cc2089", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/yuqi777/WorkDemo-350", "topics": ["recording", "workflow", "video"], "updated_at": null}, "text": "WorkDemo-350\nResource type: dataset\nReview status: curated\nScope: core\n\nDesktop demonstrations with task descriptions, videos, input logs, and action-linked screenshot evidence for studying reusable agent skills.\nPlatforms: desktop\nTopics: recording, workflow, video\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md\", \"population\": \"packaged tasks/videos\", \"qualifier\": \"exact\", \"unit\": \"demonstrations\", \"value\": 350}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md\", \"population\": \"final packaged images\", \"qualifier\": \"exact\", \"unit\": \"evidence images\", \"value\": 8388}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md\", \"population\": \"packaged action traces\", \"qualifier\": \"exact\", \"unit\": \"action records\", \"value\": 11951}], \"evidence_urls\": [\"https://huggingface.co/datasets/yuqi777/WorkDemo-350/raw/main/README.md\"], \"formats\": [\"MP4\", \"PNG\", \"JSON\", \"JSONL\", \"commented key logs\"], \"modalities\": [\"screen video\", \"screenshots\", \"input events\", \"task text\"], \"notes\": \"The OSWorld subset was selected post hoc using model outcomes; no train/test split is supplied. Coordinate calibration, task success, and complete anonymization are not established by package-format checks.\", \"splits\": [], \"supervision\": [\"human workflow demonstrations\", \"skill induction\"]}, \"download_url\": \"https://huggingface.co/datasets/yuqi777/WorkDemo-350/tree/main\", \"modalities\": [\"video\", \"image\", \"text\", \"input-events\"], \"sample_count\": 350, \"sample_unit\": \"task demonstrations / videos\"}\n\nThe card documents post hoc selection of the 286-task OSWorld subset based on model outcomes, uncalibrated coordinate annotations, and missing task-success acceptance records. It provides no train/test split or complete evaluator and does not claim complete anonymization. Review these limits before training or evaluation; the collection is an offline demonstration asset.\n\nPrimary source: https://huggingface.co/datasets/yuqi777/WorkDemo-350\nCatalog record: https://awesomecua.github.io/resources/url-243ab00712ba19cc2089/"}
{"id": "url:25a4d0ab29b60250863a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa", "citation_url": "https://awesomecua.github.io/resources/url-25a4d0ab29b60250863a/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa/tree/main", "gated": false, "hub_id": "stanfordnlp/nnetnav-wa", "languages": [], "modalities": ["text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "3fe40a45aa2bf9d6f0cefffc65ea6382849c3543b44017200a4a8d43ab61e9eb", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-wa"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "3fe40a45aa2bf9d6f0cefffc65ea6382849c3543b44017200a4a8d43ab61e9eb", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-wa"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa"}], "name": "NNetNav WebArena data", "publication": null, "published_at": null, "record_sha256": "4b148ce5f7377e4e08b6446d08e690438a6c19340d13e33b95197236efb47530", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "github:murtyshikhar/nnetnav", "type": "component_of"}], "resource_id": "url:25a4d0ab29b60250863a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-wa", "topics": ["data-generation", "web-navigation"], "updated_at": null}, "text": "NNetNav WebArena data\nResource type: dataset\nReview status: curated\nScope: core\n\nNNetNav's WebArena-oriented browser-interaction data release, with separate training and test files.\nPlatforms: web\nTopics: data-generation, web-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/stanfordnlp/nnetnav-wa/tree/main\", \"gated\": false, \"hub_id\": \"stanfordnlp/nnetnav-wa\", \"modalities\": [\"text\"]}\n\nThe Hub card supplies a license but no detailed schema. Dataset purpose is supported by the project's README; exact sample counts and annotation quality remain unverified.\n\nPrimary source: https://huggingface.co/datasets/stanfordnlp/nnetnav-wa\nCatalog record: https://awesomecua.github.io/resources/url-25a4d0ab29b60250863a/"}
{"id": "url:29588aa5651b5839c045", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B", "citation_url": "https://awesomecua.github.io/resources/url-29588aa5651b5839c045/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-4B", "gated": false, "hub_revision": "318dd331a8a2f60327dfd01e0ffc27998b460a3c", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8c94e2115879f474e122cffafe4d14a4e898d832fe094d7f0fffcdba4a14f9d9", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-4B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "98409c81aeba5c9261075ab720204842465041f7df3aee833b28a8e5d4347c56", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B"}], "name": "Seg2Click Qwen3.5 4B", "publication": null, "published_at": null, "record_sha256": "7d89c77d5ad8e30c880c0b22e18d3308a9ecb8a07d479ec8ede866e8901375ea", "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "url:29588aa5651b5839c045", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B", "topics": ["grounding"], "updated_at": null}, "text": "Seg2Click Qwen3.5 4B\nResource type: model\nReview status: curated\nScope: core\n\nSeg2Click checkpoint pairing a 4B Qwen3.5 backbone with pixel-level action-mask prediction.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.5-4B\", \"gated\": false, \"hub_revision\": \"318dd331a8a2f60327dfd01e0ffc27998b460a3c\", \"listed_weight_files\": [\"model.safetensors\", \"phase1_mask_decoder.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.\n\nPrimary source: https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-4B\nCatalog record: https://awesomecua.github.io/resources/url-29588aa5651b5839c045/"}
{"id": "url:2a2682f86dcb8ebc9a80", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/meituan/EvoCUA-8B-20260105", "citation_url": "https://awesomecua.github.io/resources/url-2a2682f86dcb8ebc9a80/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/meituan/EvoCUA-8B-20260105/tree/main", "gated": false, "hub_id": "meituan/EvoCUA-8B-20260105", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "3fd1e50729e97746fec57c7d5da6f093d76fb9174e74c3f2c630fed5486fe995", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "3fd1e50729e97746fec57c7d5da6f093d76fb9174e74c3f2c630fed5486fe995", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/meituan/EvoCUA-8B-20260105"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/meituan/EvoCUA-8B-20260105"}], "name": "EvoCUA 8B 20260105", "publication": null, "published_at": null, "record_sha256": "69d9d5f469ad1de6dcddf5d338ad8dd93802221f844c70f4c514e165e933a17a", "relations": [{"evidence_url": "https://huggingface.co/api/models/meituan/EvoCUA-8B-20260105", "target_id": "github:meituan/evocua", "type": "variant_of"}], "resource_id": "url:2a2682f86dcb8ebc9a80", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/meituan/EvoCUA-8B-20260105", "topics": ["reinforcement-learning", "planning"], "updated_at": null}, "text": "EvoCUA 8B 20260105\nResource type: model\nReview status: curated\nScope: core\n\nEvoCUA computer-use agent checkpoint (20260105 release) with about 8.8B BF16 parameters in the Qwen3-VL architecture, tagged for OSWorld and for English and Chinese.\nPlatforms: desktop\nTopics: reinforcement-learning, planning\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/meituan/EvoCUA-8B-20260105/tree/main\", \"gated\": false, \"hub_id\": \"meituan/EvoCUA-8B-20260105\", \"languages\": [\"en\", \"zh\"]}\n\nPrimary source: https://huggingface.co/meituan/EvoCUA-8B-20260105\nCatalog record: https://awesomecua.github.io/resources/url-2a2682f86dcb8ebc9a80/"}
{"id": "url:2b5dead3172c0b70da7e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK", "citation_url": "https://awesomecua.github.io/resources/url-2b5dead3172c0b70da7e/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "population": "VideoCAD published collection", "qualifier": "more-than", "unit": "video recordings", "value": 41000}], "evidence_urls": ["https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK", "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"], "formats": ["ZIP", "MP4", "JSON", "PNG"], "modalities": ["video", "UI action logs", "target CAD images"], "notes": "The reported video total is distinct from the Dataverse package-file count. The exact-version paper identifies the recorded interface as browser-based Onshape; operating-system compatibility is not established.", "splits": [], "supervision": ["timestamped UI actions", "target model images"]}, "dataverse": {"file_count": 261, "license": {"iconUri": "https://licensebuttons.net/p/zero/1.0/88x31.png", "languageCode": "en", "name": "CC0 1.0", "rightsIdentifier": "CC0-1.0", "rightsIdentifierScheme": "SPDX", "schemeUri": "https://spdx.org/licenses/", "uri": "http://creativecommons.org/publicdomain/zero/1.0"}, "persistent_url": "https://doi.org/10.7910/DVN/WX8PCK", "publication_date": "2025-05-15", "publisher": "Harvard Dataverse", "released_at": "2026-03-11T01:35:02Z", "version_minor_number": 0, "version_number": 5, "version_state": "RELEASED"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:53.536297+00:00", "role": "primary", "sha256": "cbb08803922b4b7736979ca0a5fe552b111fcd428fa9317401bc090b14bb4e54", "supports": ["summary", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T12:02:34.311949+00:00", "role": "primary", "sha256": "a303d95a811331cd6ee23b0340f1b9ef2dd996835bfbc66656a871ff94eead50", "supports": ["environments", "notes"], "url": "https://arxiv.org/html/2505.24838v2"}, {"method": "dataverse_api", "retrieved_at": "2026-10-17T02:03:20Z", "role": "primary", "sha256": "f67df0a4422a0484aaa2f93cd185b7000070d2a88d919d9ef995587f5e724007", "supports": ["details.dataverse", "licenses.data"], "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:20Z", "licenses": {"code": null, "data": {"basis": "publisher Dataverse dataset-version license metadata; original declaration and URI retained in details.dataverse.license", "declared": "CC0 1.0", "retrieved_at": "2026-10-17T02:03:20Z", "url": "https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK"}], "name": "VideoCAD data", "publication": null, "published_at": null, "record_sha256": "fcf7385f6a7868f6668bd2b4e97b8d2e5f2f033ea1d469e1967b1c983cda8d6e", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md", "target_id": "arxiv:2505.24838", "type": "documented_by"}], "resource_id": "url:2b5dead3172c0b70da7e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK", "topics": ["grounding", "imitation-learning", "video"], "updated_at": null}, "text": "VideoCAD data\nResource type: dataset\nReview status: curated\nScope: core\n\nCAD screen recordings paired with action logs and target model images for studying precise interface control and long-horizon visual imitation.\nPlatforms: web\nTopics: grounding, imitation-learning, video\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK\", \"population\": \"VideoCAD published collection\", \"qualifier\": \"more-than\", \"unit\": \"video recordings\", \"value\": 41000}], \"evidence_urls\": [\"https://dataverse.harvard.edu/api/datasets/:persistentId/?persistentId=doi:10.7910/DVN/WX8PCK\", \"https://raw.githubusercontent.com/ghadinehme/VideoCAD/2d69eeaf0ac44210c76dcab2d08b4d4610be236c/README.md\"], \"formats\": [\"ZIP\", \"MP4\", \"JSON\", \"PNG\"], \"modalities\": [\"video\", \"UI action logs\", \"target CAD images\"], \"notes\": \"The reported video total is distinct from the Dataverse package-file count. The exact-version paper identifies the recorded interface as browser-based Onshape; operating-system compatibility is not established.\", \"splits\": [], \"supervision\": [\"timestamped UI actions\", \"target model images\"]}, \"dataverse\": {\"file_count\": 261, \"license\": {\"iconUri\": \"https://licensebuttons.net/p/zero/1.0/88x31.png\", \"languageCode\": \"en\", \"name\": \"CC0 1.0\", \"rightsIdentifier\": \"CC0-1.0\", \"rightsIdentifierScheme\": \"SPDX\", \"schemeUri\": \"https://spdx.org/licenses/\", \"uri\": \"http://creativecommons.org/publicdomain/zero/1.0\"}, \"persistent_url\": \"https://doi.org/10.7910/DVN/WX8PCK\", \"publication_date\": \"2025-05-15\", \"publisher\": \"Harvard Dataverse\", \"released_at\": \"2026-03-11T01:35:02Z\", \"version_minor_number\": 0, \"version_number\": 5, \"version_state\": \"RELEASED\"}}\n\nDataverse reports a RELEASED version 5.0 with 261 unrestricted files and a CC0-1.0 dataset declaration. Archive count is not a video or trajectory count. The author repository documents the action and video formats, and Section 3 of the v2 paper identifies the recorded Onshape browser interface. No archives or training samples were downloaded; operating-system coverage is unspecified.\n\nPrimary source: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/WX8PCK\nCatalog record: https://awesomecua.github.io/resources/url-2b5dead3172c0b70da7e/"}
{"id": "url:2c2004baae4669f8169a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/ServiceNow/ui-vision", "citation_url": "https://awesomecua.github.io/resources/url-2c2004baae4669f8169a/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md", "population": "paper-described benchmark coverage", "qualifier": "exact", "unit": "software applications", "value": 83}], "evidence_urls": ["https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"], "formats": ["JSON annotations", "image files"], "modalities": ["screenshots", "bounding boxes", "UI labels", "action annotations"], "notes": "This is an offline evaluation dataset. Application coverage is not a sample count. The inspected file-layout documentation describes grounding annotations and images; it does not establish a common split or complete action-trajectory inventory.", "splits": [], "supervision": ["element grounding", "layout grounding", "action prediction"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/ServiceNow/ui-vision/tree/main", "gated": false, "hub_id": "ServiceNow/ui-vision", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:37Z", "role": "primary", "sha256": "295cf7f7500a18cea5c80eab1cdf4d7264f2f0e959b75efc5ba59083d93c9097", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "34f4ad5c9e7160b5be5aa8ae77e6e9f83a277aada41a272b82a2628d21159a8c", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ServiceNow/ui-vision"}], "name": "UI-Vision", "publication": null, "published_at": null, "record_sha256": "f5434cff68d0880d6426b3d3d732fa92e99800ee6e87cc40e6a2526f9a287246", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/ServiceNow/ui-vision", "target_id": "arxiv:2503.15661", "type": "documented_by"}], "resource_id": "url:2c2004baae4669f8169a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/ServiceNow/ui-vision", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "UI-Vision\nResource type: dataset\nReview status: curated\nScope: core\n\nDesktop GUI data for interface understanding and interaction research.\nPlatforms: desktop\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md\", \"population\": \"paper-described benchmark coverage\", \"qualifier\": \"exact\", \"unit\": \"software applications\", \"value\": 83}], \"evidence_urls\": [\"https://huggingface.co/datasets/ServiceNow/ui-vision/raw/main/README.md\"], \"formats\": [\"JSON annotations\", \"image files\"], \"modalities\": [\"screenshots\", \"bounding boxes\", \"UI labels\", \"action annotations\"], \"notes\": \"This is an offline evaluation dataset. Application coverage is not a sample count. The inspected file-layout documentation describes grounding annotations and images; it does not establish a common split or complete action-trajectory inventory.\", \"splits\": [], \"supervision\": [\"element grounding\", \"layout grounding\", \"action prediction\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/ServiceNow/ui-vision/tree/main\", \"gated\": false, \"hub_id\": \"ServiceNow/ui-vision\", \"modalities\": [\"image\"]}\n\nPrimary source: https://huggingface.co/datasets/ServiceNow/ui-vision\nCatalog record: https://awesomecua.github.io/resources/url-2c2004baae4669f8169a/"}
{"id": "url:2e1927450da2183a5abe", "metadata": {"aliases": ["url:1ff6b78a764bdcc6dd03", "HanXiao1999/UI-Genie-Agent-5k", "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-5k"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k", "citation_url": "https://awesomecua.github.io/resources/url-2e1927450da2183a5abe/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md", "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"], "formats": ["JSONL", "PNG"], "modalities": ["screenshots", "text", "actions"], "notes": "Public JSONL annotations and screenshots are present. MIT is declared in the dataset card. The release name, model-card links and annotation filename use different size labels; this review does not infer a reliable trajectory count from them.", "splits": [], "supervision": ["GUI action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-Agent-16k"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "a748fc7dbb53d5d751c5ee109e437c377e214b1d837762f8d09995349d86cec9", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "548df402288350920a26bff2099544bf21835200ee58ff2c657c331fa6200317", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-16k"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "a748fc7dbb53d5d751c5ee109e437c377e214b1d837762f8d09995349d86cec9", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:09Z", "role": "primary", "sha256": "66006a586e03be3eb1dfc29dc9c951241fd7e4204430e7db258da1037e17ac6e", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-5k"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:10Z", "role": "primary", "sha256": "66006a586e03be3eb1dfc29dc9c951241fd7e4204430e7db258da1037e17ac6e", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-Agent-16k"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k"}], "name": "UI-Genie trajectory data", "publication": null, "published_at": null, "record_sha256": "12d8eecebaa53e7b06ca69679fc4d2df5a5338b5b64ea811d51f0dfff079d8b4", "relations": [{"evidence_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "resource_id": "url:2e1927450da2183a5abe", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k", "topics": ["data-generation", "mobile-navigation", "imitation-learning"], "updated_at": null}, "text": "UI-Genie trajectory data\nResource type: dataset\nReview status: curated\nScope: core\n\nSynthetic mobile GUI trajectory supervision and screenshots released with the UI-Genie agent-training pipeline.\nRecorded aliases: url:1ff6b78a764bdcc6dd03, HanXiao1999/UI-Genie-Agent-5k, https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-5k\nPlatforms: android\nTopics: data-generation, mobile-navigation, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/raw/main/README.md\", \"https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md\"], \"formats\": [\"JSONL\", \"PNG\"], \"modalities\": [\"screenshots\", \"text\", \"actions\"], \"notes\": \"Public JSONL annotations and screenshots are present. MIT is declared in the dataset card. The release name, model-card links and annotation filename use different size labels; this review does not infer a reliable trajectory count from them.\", \"splits\": [], \"supervision\": [\"GUI action trajectories\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k/tree/main\", \"gated\": false, \"hub_id\": \"HanXiao1999/UI-Genie-Agent-16k\"}\n\nPublic JSONL annotations and screenshots are present. MIT is declared in the dataset card. The release name, model-card links and annotation filename use different size labels; this review does not infer a reliable trajectory count from them.\n\nPrimary source: https://huggingface.co/datasets/HanXiao1999/UI-Genie-Agent-16k\nCatalog record: https://awesomecua.github.io/resources/url-2e1927450da2183a5abe/"}
{"id": "url:30c831a1198b9bc2a0f5", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B", "citation_url": "https://awesomecua.github.io/resources/url-30c831a1198b9bc2a0f5/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-4B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "20c6679849cd81a871809aa7f96145855739c9170608dfa6ec3653e52bd795cb", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B"}], "name": "OS-Atlas Base 4B", "publication": null, "published_at": null, "record_sha256": "10e52e7fdd61a22417c6be4b5f2134a94562671164ef010729bcf86b0c3cb12c", "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-4B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "resource_id": "url:30c831a1198b9bc2a0f5", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B", "topics": ["grounding"], "updated_at": null}, "text": "OS-Atlas Base 4B\nResource type: model\nReview status: curated\nScope: core\n\nOS-Atlas base checkpoint with about 4.1B parameters, fine-tuned from InternVL2-4B and loaded in Transformers through custom InternVL chat model code.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Atlas-Base-4B\"}\n\nPrimary source: https://huggingface.co/OS-Copilot/OS-Atlas-Base-4B\nCatalog record: https://awesomecua.github.io/resources/url-30c831a1198b9bc2a0f5/"}
{"id": "url:342b081c3668a8bbe155", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e", "citation_url": "https://awesomecua.github.io/resources/url-342b081c3668a8bbe155/", "details": {"benchmark_protocol": {"actions": [], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md", "population": "Hub test split at revision b8eca829", "qualifier": "exact", "unit": "task-definition records", "value": 25}], "evaluation": "The published task schema contains rubrics and human_reference_operations. No automated scoring protocol or single numeric agent metric is established by the inspected card; execution uses separately configured Android tooling.", "evidence_urls": ["https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md", "https://huggingface.co/api/datasets/mikewang/mobile_eval_e", "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"], "focus": ["long-horizon multi-app Android task execution"], "notes": "Features are task_id, instruction, type, apps, scenario, rubrics and human-reference operation strings. A task record is not an action count or recorded screenshot trajectory. Public Parquet inventory was checked but no sample rows were downloaded; companion README limits the tested agent platform to Android.", "observations": ["task instructions", "app/scenario metadata"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/tree/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83", "gated": false, "hub_id": "mikewang/mobile_eval_e", "hub_revision": "b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md"], "reason": "No data-license declaration is present in the inspected task card; companion code MIT does not establish a dataset grant."}}, "repository_url": "https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:15Z", "role": "primary", "sha256": "f8515fbce73ea0ef3d1d558d4567bb949244e6384e61c956ba3ddbc27294f0ae", "supports": ["name", "kind", "environments", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "f9b10eada9e0a64e26f1a7d477768f218b807103199855d893b3ab07feddc755", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/mikewang/mobile_eval_e"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:15Z", "role": "primary", "sha256": "9fdf43753fbfac9868d8bec57fa5601db3e03ed5c30fb981f9da88f6b75ffcc5", "supports": ["summary", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:31:38Z", "role": "primary", "sha256": "e96b86015c763c780e0430a8d0867b96cae947f50e888b3441c85abcaa42de78", "supports": ["summary", "topics", "details.benchmark_protocol"], "url": "https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-18T04:27:15Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mikewang/mobile_eval_e"}], "name": "Mobile-Eval-E task definitions", "publication": null, "published_at": null, "record_sha256": "fbd9746715f2ea9784d08159669cbce8ec77d75313e7951ec0974499996df956", "relations": [{"evidence_url": "https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md", "target_id": "arxiv:2501.11733", "type": "documented_by"}], "resource_id": "url:342b081c3668a8bbe155", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/mikewang/mobile_eval_e", "topics": ["evaluation", "long-horizon", "cross-app"], "updated_at": null}, "text": "Mobile-Eval-E task definitions\nResource type: benchmark\nReview status: curated\nScope: core\n\nTask definitions for Mobile-Agent-E’s long-horizon, multi-app Android evaluation, packaged as a 25-record test split with rubrics and human-reference operation strings.\nPlatforms: android\nTopics: evaluation, long-horizon, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md\", \"population\": \"Hub test split at revision b8eca829\", \"qualifier\": \"exact\", \"unit\": \"task-definition records\", \"value\": 25}], \"evaluation\": \"The published task schema contains rubrics and human_reference_operations. No automated scoring protocol or single numeric agent metric is established by the inspected card; execution uses separately configured Android tooling.\", \"evidence_urls\": [\"https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md\", \"https://huggingface.co/api/datasets/mikewang/mobile_eval_e\", \"https://raw.githubusercontent.com/X-PLUG/MobileAgent/HEAD/Mobile-Agent-E/README.md\", \"https://export.arxiv.org/api/query?id_list=2412.04454%2C2412.04626%2C2412.05467%2C2412.09362%2C2412.09605%2C2412.10342%2C2412.10742%2C2412.10840%2C2412.13103%2C2412.13194%2C2412.13501%2C2412.14161%2C2412.14470%2C2412.16256%2C2412.17589%2C2412.18116%2C2412.18426%2C2412.19723%2C2501.01149%2C2501.04575%2C2501.07572%2C2501.11733%2C2501.12326%2C2501.13896%2C2501.16150%2C2502.01600%2C2502.01977%2C2502.02982%2C2502.06395%2C2502.07949%2C2502.08047%2C2502.08226%2C2502.11357%2C2502.13053%2C2502.13923%2C2502.14282%2C2502.18906%2C2503.00401%2C2503.02268%2C2503.03196%2C2503.04730%2C2503.05143%2C2503.06470%2C2503.09263%2C2503.11069%2C2503.11170%2C2503.12687%2C2503.15661%2C2503.15937%2C2503.16465&max_results=50\"], \"focus\": [\"long-horizon multi-app Android task execution\"], \"notes\": \"Features are task_id, instruction, type, apps, scenario, rubrics and human-reference operation strings. A task record is not an action count or recorded screenshot trajectory. Public Parquet inventory was checked but no sample rows were downloaded; companion README limits the tested agent platform to Android.\", \"observations\": [\"task instructions\", \"app/scenario metadata\"], \"setting\": \"interactive\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/mikewang/mobile_eval_e/tree/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83\", \"gated\": false, \"hub_id\": \"mikewang/mobile_eval_e\", \"hub_revision\": \"b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/mikewang/mobile_eval_e/raw/b8eca82988ed8d4f9dec01eaa9d51abf4b35aa83/README.md\"], \"reason\": \"No data-license declaration is present in the inspected task card; companion code MIT does not establish a dataset grant.\"}}, \"repository_url\": \"https://github.com/X-PLUG/MobileAgent/tree/main/Mobile-Agent-E\"}\n\nThe public Hub manifest contains one test Parquet file. This resource is the benchmark’s task-definition package, not a bundled phone environment or agent. Card metadata supplies rubrics/reference-operation fields but no data-license declaration or verified automated scorer; the code repository’s MIT license is not transferred to these records.\n\nPrimary source: https://huggingface.co/datasets/mikewang/mobile_eval_e\nCatalog record: https://awesomecua.github.io/resources/url-342b081c3668a8bbe155/"}
{"id": "url:347925d0f440d6396080", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Shaon Barman", "Sarah Chasins", "Ras Bodik", "Sumit Gulwani"], "canonical_source_url": "https://doi.org/10.1145/2983990.2984020", "citation_url": "https://awesomecua.github.io/resources/url-347925d0f440d6396080/", "details": {"publication": {"authors": ["Shaon Barman", "Sarah Chasins", "Ras Bodik", "Sumit Gulwani"], "date_basis": "Supplied month precision: 2016-11. year=2016, month=November", "date_parts": [2016, 11], "doi": "10.1145/2983990.2984020", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/", "https://2016.splashcon.org/track/splash-2016-oopsla", "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved. Page range, volume and exact publication day are not established by the retained citation sources.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf", "publisher": null, "series": null, "title": "Ringer: Web Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/2983990.2984020", "venue": "OOPSLA 2016", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "633aa15e2e14346241de5385cc19d99f1fcdc9ef79fd06d27176dbfa8f38149a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "94c6b6bbd127126c9dd259bd346a7757d75366b4d805bdee2959c4c73226df74", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://2016.splashcon.org/track/splash-2016-oopsla"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "45f65483c3801dac142f4657c4cbe45e17b6a3e081497731e984c1e670c6e8f6", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:00Z", "role": "primary", "sha256": "bfbb69f57aaed39e3a210fae832e1fd98ed62ba1c01217b792238fe57e1a0ad1", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "45f65483c3801dac142f4657c4cbe45e17b6a3e081497731e984c1e670c6e8f6", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title", "environments", "name", "scope", "topics"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "633aa15e2e14346241de5385cc19d99f1fcdc9ef79fd06d27176dbfa8f38149a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:36Z", "role": "primary", "sha256": "94c6b6bbd127126c9dd259bd346a7757d75366b4d805bdee2959c4c73226df74", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://2016.splashcon.org/track/splash-2016-oopsla"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:00Z", "role": "primary", "sha256": "bfbb69f57aaed39e3a210fae832e1fd98ed62ba1c01217b792238fe57e1a0ad1", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "summary"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}], "identifiers": {"doi": "10.1145/2983990.2984020"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:24Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/2983990.2984020"}, {"rel": "publication_pdf", "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"}], "name": "Ringer: Web Automation by Demonstration", "publication": {"authors": ["Shaon Barman", "Sarah Chasins", "Ras Bodik", "Sumit Gulwani"], "date_basis": "Supplied month precision: 2016-11. year=2016, month=November", "date_parts": [2016, 11], "doi": "10.1145/2983990.2984020", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/bibtex/", "https://2016.splashcon.org/track/splash-2016-oopsla", "https://www.microsoft.com/en-us/research/publication/ringer-web-automation-demonstration/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved. Page range, volume and exact publication day are not established by the retained citation sources.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2016/12/oopsla16-web.pdf", "publisher": null, "series": null, "title": "Ringer: Web Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/2983990.2984020", "venue": "OOPSLA 2016", "volume": null}, "published_at": null, "record_sha256": "01a77c7512ce863a9931b100645ec00f5044b615b3474b58fe3e5a44f3e9a95a", "relations": [], "resource_id": "url:347925d0f440d6396080", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/2983990.2984020", "topics": ["web-navigation", "recording", "robustness"], "updated_at": null}, "text": "Ringer: Web Automation by Demonstration\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Ringer, which converts demonstrated webpage interactions into scripts combining actions, target selectors, and waiting conditions. It evaluates replay robustness on tasks recorded from real websites.\nPlatforms: web\nTopics: web-navigation, recording, robustness\nRecord publication date (recorded precision): 2016-11\nAuthors: Shaon Barman, Sarah Chasins, Ras Bodik, Sumit Gulwani\n\nPublication edition: Ringer: Web Automation by Demonstration\nPublication authors: Shaon Barman, Sarah Chasins, Ras Bodik, Sumit Gulwani\nPublication edition date (recorded precision): 2016-11\nPublication venue: OOPSLA 2016\nPublication doi: 10.1145/2983990.2984020\nPublication date basis: Supplied month precision: 2016-11. year=2016, month=November\nPublication notes: The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved. Page range, volume and exact publication day are not established by the retained citation sources.\nPublication source: https://doi.org/10.1145/2983990.2984020\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The original manuscript and MSR BibTeX print Sarah Chasins and Ras Bodik. The official conference profiles expand these to Sarah E. Chasins and Rastislav Bodík; the printed publication names are preserved.\n\nPrimary source: https://doi.org/10.1145/2983990.2984020\nCatalog record: https://awesomecua.github.io/resources/url-347925d0f440d6396080/"}
{"id": "url:35b9879a351cc3b3d7d5", "metadata": {"aliases": ["url:cab7bd78abb1faa469e7", "Reallm-Labs/InfiGUI-R1-3B", "https://huggingface.co/Reallm-Labs/InfiGUI-R1-3B"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B", "citation_url": "https://awesomecua.github.io/resources/url-35b9879a351cc3b3d7d5/", "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-R1-3B"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["identity"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "5fd1399c88bdf406e7689765ccf0599b3b7e0ddbed082fa92bffdd7b120bd41a", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Reallm-Labs/InfiGUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:05:41Z", "role": "primary", "sha256": "73496fb30838ba9ea291bb6d1e14004ac3318ac03a1ae7b5d6044e0dc14fd46f", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:15Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/Reallm-Labs/InfiGUI-R1-3B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:16Z", "role": "primary", "sha256": "213a5181a3516cf15d5f7b746986bc1e36cd6579acabd8aa1c2fb16a23a51ee0", "supports": ["aliases"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-R1-3B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:05:41Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B"}], "name": "InfiGUI-R1 3B", "publication": null, "published_at": null, "record_sha256": "0da16af5afbd81b5e9f2763c0a8b2f6ec3b53b1dc3af63da54e26beb105267da", "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/raw/main/README.md", "target_id": "arxiv:2504.14239", "type": "documented_by"}], "resource_id": "url:35b9879a351cc3b3d7d5", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/InfiX-ai/InfiGUI-R1-3B", "topics": ["grounding", "reinforcement-learning", "self-reflection"], "updated_at": null}, "text": "InfiGUI-R1 3B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL checkpoint trained with the Actor2Reasoner procedure to combine interface grounding with planning and recovery reasoning.\nRecorded aliases: url:cab7bd78abb1faa469e7, Reallm-Labs/InfiGUI-R1-3B, https://huggingface.co/Reallm-Labs/InfiGUI-R1-3B\nPlatforms: web, desktop, android\nTopics: grounding, reinforcement-learning, self-reflection\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/InfiX-ai/InfiGUI-R1-3B/tree/main\", \"gated\": false, \"hub_id\": \"InfiX-ai/InfiGUI-R1-3B\"}\n\nThe official Hub API resolves the former Reallm-Labs model namespace to InfiX-ai. Two safetensors shards and model/tokenizer files are public and ungated. The card declares Apache-2.0 for this checkpoint; base-model terms remain separate. Its older loading example still uses the prior namespace.\n\nPrimary source: https://huggingface.co/InfiX-ai/InfiGUI-R1-3B\nCatalog record: https://awesomecua.github.io/resources/url-35b9879a351cc3b3d7d5/"}
{"id": "url:392a602c660ffa91be46", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Toby Jia-Jun Li", "Amos Azaria", "Brad A. Myers"], "canonical_source_url": "https://doi.org/10.1145/3025453.3025483", "citation_url": "https://awesomecua.github.io/resources/url-392a602c660ffa91be46/", "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Amos Azaria", "Brad A. Myers"], "date_basis": "Supplied year precision: 2017. Publications: SUGILITE, CHI 2017 Only publication year is asserted; event dates are kept separate.", "date_parts": [2017], "doi": "10.1145/3025453.3025483", "event_dates": null, "evidence_urls": ["https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution. Publication month/day are not independently identified by the retained sources.", "pages": "6038-6049", "pdf_url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf", "publisher": null, "series": null, "title": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/3025453.3025483", "venue": "Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "cc23c19284f0a558c4b99b8b051abca8ed61e274e9e1f1fed1ae91487f4368ad", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "cc23c19284f0a558c4b99b8b051abca8ed61e274e9e1f1fed1ae91487f4368ad", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}], "identifiers": {"doi": "10.1145/3025453.3025483"}, "kind": "paper", "last_verified_at": "2026-10-17T09:54:39Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3025453.3025483"}, {"rel": "publication_pdf", "url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"}], "name": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration", "publication": {"authors": ["Toby Jia-Jun Li", "Amos Azaria", "Brad A. Myers"], "date_basis": "Supplied year precision: 2017. Publications: SUGILITE, CHI 2017 Only publication year is asserted; event dates are kept separate.", "date_parts": [2017], "doi": "10.1145/3025453.3025483", "event_dates": null, "evidence_urls": ["https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution. Publication month/day are not independently identified by the retained sources.", "pages": "6038-6049", "pdf_url": "https://toby.li/files/TobyLi-CHI2017-Sugilite.pdf", "publisher": null, "series": null, "title": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration", "type": "conference-paper", "url": "https://doi.org/10.1145/3025453.3025483", "venue": "Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems", "volume": null}, "published_at": null, "record_sha256": "0cfa98dba34fd4680eae70536df7630c50c325dcb4740da19cd4bfd45c764ce9", "relations": [], "resource_id": "url:392a602c660ffa91be46", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/3025453.3025483", "topics": ["mobile-navigation", "recording", "human-in-the-loop", "cross-app"], "updated_at": null}, "text": "SUGILITE: Creating Multimodal Smartphone Automation by Demonstration\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces SUGILITE, which combines spoken instructions, demonstrated Android actions, and UI hierarchies to generate reusable automation scripts. Its interface includes context checking and repairs for unfamiliar situations and changing app interfaces.\nPlatforms: android\nTopics: mobile-navigation, recording, human-in-the-loop, cross-app\nRecord publication date (recorded precision): 2017\nAuthors: Toby Jia-Jun Li, Amos Azaria, Brad A. Myers\n\nPublication edition: SUGILITE: Creating Multimodal Smartphone Automation by Demonstration\nPublication authors: Toby Jia-Jun Li, Amos Azaria, Brad A. Myers\nPublication edition date (recorded precision): 2017\nPublication venue: Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems\nPublication pages: 6038-6049\nPublication doi: 10.1145/3025453.3025483\nPublication date basis: Supplied year precision: 2017. Publications: SUGILITE, CHI 2017 Only publication year is asserted; event dates are kept separate.\nPublication notes: Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution. Publication month/day are not independently identified by the retained sources.\nPublication source: https://doi.org/10.1145/3025453.3025483\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. Related to later APPINITE and PUMICE work but distinguished here by its 2017 three-author publication and base demonstration-automation contribution.\n\nPrimary source: https://doi.org/10.1145/3025453.3025483\nCatalog record: https://awesomecua.github.io/resources/url-392a602c660ffa91be46/"}
{"id": "url:3932d69c6b6f9b2a7ba8", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b", "citation_url": "https://awesomecua.github.io/resources/url-3932d69c6b6f9b2a7ba8/", "details": {"disabled": false, "download_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/tree/main", "gated": false, "hub_id": "xTRam1/plan-and-act-planner-70b", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"], "reason": "The model card declares MIT in its metadata header (license: mit), beside a template comment asking the publisher to update it if needed. It is recorded as the publisher's declaration and does not override the Llama-3.3-70B-Instruct base model terms."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "4ef2d84d2b5d9d2972907e6135ea49f1ef088beaf6ef6042dc7d8fb4297a9a3a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "d2f757790ded81244e7c9b9e7350ec9fdc304609e387a95c7918db705bda535c", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/xTRam1/plan-and-act-planner-70b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "feee31aa4d0801cb364578099a13e4f9cdcaa4eb5fbf4312b3fc3f8edd479a14", "supports": ["environments"], "url": "https://raw.githubusercontent.com/SqueezeAILab/plan-and-act/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b"}], "name": "Plan-and-Act Planner 70B", "publication": null, "published_at": null, "record_sha256": "1808588c7a2deaba9f0be7e761e10ed9521adc77abf0fddd8aaf86209b0051ed", "relations": [{"evidence_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "resource_id": "url:3932d69c6b6f9b2a7ba8", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/xTRam1/plan-and-act-planner-70b", "topics": ["planning", "web-navigation", "long-horizon"], "updated_at": null}, "text": "Plan-and-Act Planner 70B\nResource type: model\nReview status: curated\nScope: core\n\nThe planner component of Plan-and-Act, generating structured high-level plans that an executor turns into web actions.\nPlatforms: web\nTopics: planning, web-navigation, long-horizon\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/xTRam1/plan-and-act-planner-70b/tree/main\", \"gated\": false, \"hub_id\": \"xTRam1/plan-and-act-planner-70b\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/xTRam1/plan-and-act-planner-70b/raw/main/README.md\"], \"reason\": \"The model card declares MIT in its metadata header (license: mit), beside a template comment asking the publisher to update it if needed. It is recorded as the publisher's declaration and does not override the Llama-3.3-70B-Instruct base model terms.\"}}}\n\nThe Hub manifest contains 17 safetensors shards and tokenizer/configuration files. Card metadata declares MIT and Llama-3.3-70B-Instruct as the base, but retains template comments asking the publisher to update those fields if needed. These are recorded as publisher declarations, without claiming they override upstream model terms.\n\nPrimary source: https://huggingface.co/xTRam1/plan-and-act-planner-70b\nCatalog record: https://awesomecua.github.io/resources/url-3932d69c6b6f9b2a7ba8/"}
{"id": "url:3b6dfd52f98bfe76bc7f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/microsoft/WebTailBench", "citation_url": "https://awesomecua.github.io/resources/url-3b6dfd52f98bfe76bc7f/", "details": {"benchmark_protocol": {"actions": ["browser interaction"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "population": "main V1/V2 suite (shared identities)", "qualifier": "exact", "unit": "task identities", "value": 609}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "population": "refusal suite", "qualifier": "exact", "unit": "tasks", "value": 111}], "evaluation": "Task-specific precomputed rubrics; refusal tasks are a separate evaluation population.", "evidence_urls": ["https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"], "focus": ["live web task completion", "unsafe-request refusal"], "notes": "V2 refreshes task strings and rubrics while preserving V1 task IDs. Counts are task definitions, not independently verified runnable website states.", "observations": ["web pages"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/microsoft/WebTailBench/tree/main", "gated": false, "hub_id": "microsoft/WebTailBench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"], "reason": "The benchmark's Hub dataset card declares MIT in its metadata header (license: mit)."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c4cfc361d1d47d7855269af4110483363964e23f9c65244a1182830eccccbb7f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "93469a20d333219dde89a225891a54ee716120d559eb8ac9f69a1a7ee84f57a3", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/microsoft/WebTailBench"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/microsoft/WebTailBench"}], "name": "WebTailBench", "publication": null, "published_at": null, "record_sha256": "e260a5452ddd8d317ec09cc5bbd6ecff8460610feecc3c23b074001b550d776b", "relations": [{"evidence_url": "https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md", "target_id": "arxiv:2511.19663", "type": "documented_by"}], "resource_id": "url:3b6dfd52f98bfe76bc7f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/microsoft/WebTailBench", "topics": ["evaluation", "web-navigation", "safety-security"], "updated_at": null}, "text": "WebTailBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nLive-web task definitions and rubrics for evaluating computer-use agents, with preserved original tasks, a refreshed V2 task suite and a separate refusal set.\nPlatforms: web\nTopics: evaluation, web-navigation, safety-security\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"browser interaction\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md\", \"population\": \"main V1/V2 suite (shared identities)\", \"qualifier\": \"exact\", \"unit\": \"task identities\", \"value\": 609}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md\", \"population\": \"refusal suite\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 111}], \"evaluation\": \"Task-specific precomputed rubrics; refusal tasks are a separate evaluation population.\", \"evidence_urls\": [\"https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md\"], \"focus\": [\"live web task completion\", \"unsafe-request refusal\"], \"notes\": \"V2 refreshes task strings and rubrics while preserving V1 task IDs. Counts are task definitions, not independently verified runnable website states.\", \"observations\": [\"web pages\"], \"setting\": \"interactive\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/microsoft/WebTailBench/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/WebTailBench\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/microsoft/WebTailBench/raw/main/README.md\"], \"reason\": \"The benchmark's Hub dataset card declares MIT in its metadata header (license: mit).\"}}}\n\nPublic TSV files contain V1, V2 and refusal task definitions. The card declares MIT for the dataset. V1 and V2 share the same 609 task identities and must not be summed as distinct tasks. The May 2026 refresh supersedes calendar-bound task wording; live-site availability is not established by this metadata review.\n\nPrimary source: https://huggingface.co/datasets/microsoft/WebTailBench\nCatalog record: https://awesomecua.github.io/resources/url-3b6dfd52f98bfe76bc7f/"}
{"id": "url:3dab53c6c38c9e7217ea", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data", "citation_url": "https://awesomecua.github.io/resources/url-3dab53c6c38c9e7217ea/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"], "formats": ["Parquet", "LeRobot metadata"], "modalities": ["screenshots", "instructions", "actions"], "notes": "No sample or episode count is inferred from the size tag or package manifest. Main-branch files were inspected; the card’s v3.0 usage example was not executed.", "splits": [], "supervision": ["drag trajectories", "click episodes"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi-data"}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "0e3a2a923865ae24d0de590143e9e32564b0e20db059579582d6181cdcf75bf1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "86fadf9a32185f018f6a8e49d52b39ee5c99c177075510361202dc25a214191f", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-pi-data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data"}], "name": "ShowUI-π training data", "publication": null, "published_at": null, "record_sha256": "97b4b6b9ead0982576e80b047ed3571a94eefe1e72ea9e980a6bb13a4cfe5d94", "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "resource_id": "url:3dab53c6c38c9e7217ea", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/showlab/ShowUI-pi-data", "topics": ["grounding", "action-prediction", "imitation-learning"], "updated_at": null}, "text": "ShowUI-π training data\nResource type: dataset\nReview status: curated\nScope: core\n\nDrag-and-click episode data for the ShowUI-π model, spanning desktop editing, presentation, handwriting and other interface tasks.\nPlatforms: desktop\nTopics: grounding, action-prediction, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/showlab/ShowUI-pi-data/raw/main/README.md\"], \"formats\": [\"Parquet\", \"LeRobot metadata\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\"], \"notes\": \"No sample or episode count is inferred from the size tag or package manifest. Main-branch files were inspected; the card’s v3.0 usage example was not executed.\", \"splits\": [], \"supervision\": [\"drag trajectories\", \"click episodes\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/showlab/ShowUI-pi-data/tree/main\", \"gated\": false, \"hub_id\": \"showlab/ShowUI-pi-data\"}\n\nPublic Parquet chunks and LeRobot metadata are present. The card names a v3.0 revision in its usage example; this review verifies the main manifest and does not assert that tag’s compatibility. No data license is declared. Hub size-category bands are not treated as episode counts.\n\nPrimary source: https://huggingface.co/datasets/showlab/ShowUI-pi-data\nCatalog record: https://awesomecua.github.io/resources/url-3dab53c6c38c9e7217ea/"}
{"id": "url:41abb1b7206689cf5057", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data", "citation_url": "https://awesomecua.github.io/resources/url-41abb1b7206689cf5057/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE-Data"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "7782422de88830a497d9eb530bd2f0fb6283904cbb25dafcf2ff4d03358a377a", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/KDEGroup/UI-AGILE-Data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:43:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data"}], "name": "UI-AGILE-Data", "publication": null, "published_at": null, "record_sha256": "22a60305ade21cf9d6d2cc25bebf7d3781a3902ab1ba85f85f0883cd93e1fbc4", "relations": [{"evidence_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "resource_id": "url:41abb1b7206689cf5057", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null}, "text": "UI-AGILE-Data\nResource type: dataset\nReview status: curated\nScope: core\n\nUI-AGILE training annotations and screenshots together with converted GUI-grounding and AndroidControl evaluation files.\nPlatforms: android\nTopics: grounding, reinforcement-learning, evaluation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/tree/main\", \"gated\": false, \"hub_id\": \"KDEGroup/UI-AGILE-Data\"}\n\nPublic JSON, image archives and Parquet evaluation files are present. The card has no data-license declaration; the model’s MIT declaration is not inherited. The publisher describes fixes to AndroidControl action parameters, but those corrections were not independently reproduced or applied to upstream resources.\n\nPrimary source: https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data\nCatalog record: https://awesomecua.github.io/resources/url-41abb1b7206689cf5057/"}
{"id": "url:422010da7a5c845a18f6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/hkust-nlp/GUIMid", "citation_url": "https://awesomecua.github.io/resources/url-422010da7a5c845a18f6/", "details": {"availability": "Ungated mixture/post-training JSON and image-archive manifests; some per-domain images and training code remain pending.", "dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md", "https://raw.githubusercontent.com/hkust-nlp/GUIMid/HEAD/README.md"], "formats": ["JSON", "tar.gz image archives"], "modalities": ["text", "images"], "notes": "Do not treat Hub size bands or the inconsistent introductory task count as a sample total. Some constituent-domain image packages remain forthcoming.", "splits": ["train configurations for named mid-training domains"], "supervision": ["mid-training instruction supervision", "GUI post-training trajectories"]}, "gated": false, "hub_id": "hkust-nlp/GUIMid", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"], "reason": "Publisher dataset-card declaration for the released packaging; original source datasets retain their own terms."}}, "release_revision": "b6a360815c843b5733674191b6e6d7fa68b2ea8c"}, "environments": ["web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "1cdf702c819ecf370f4191c6c71fae7b579cfbaf766908b42156964a69eddacc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "ededa2b3c8f21af8188152b028758c800db76cdbb8461ba82b8574cab375914e", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/hkust-nlp/GUIMid"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "31b836da66c6907ff9a75ebc7f06da0db67e7d9fe9ca92d0c770f5f15f90a39e", "supports": ["identity"], "url": "https://api.github.com/repos/hkust-nlp/GUIMid"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:19Z", "role": "primary", "sha256": "d9f63a03088c37f7145725c7a182100d1593fceadd26a9bf817d434ca8d6297e", "supports": ["notes", "relations"], "url": "https://raw.githubusercontent.com/hkust-nlp/GUIMid/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T11:23:07Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for the released packaging; original source datasets retain their own terms.", "declared": "mit", "retrieved_at": "2026-10-17T11:23:07Z", "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/hkust-nlp/GUIMid"}], "name": "GUIMid training data", "publication": null, "published_at": null, "record_sha256": "f9b2cef24fdedd36ec1b44bb7c42b4569858a04dc440af980cbaa9e6d391ac6b", "relations": [{"evidence_url": "https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md", "target_id": "arxiv:2504.10127", "type": "documented_by"}], "resource_id": "url:422010da7a5c845a18f6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/hkust-nlp/GUIMid", "topics": ["imitation-learning", "planning", "screen-understanding"], "updated_at": null}, "text": "GUIMid training data\nResource type: dataset\nReview status: curated\nScope: core\n\nA released GUI-agent mid-training mixture and supporting domain datasets, together with a post-training trajectory file and image archives for web and mobile interaction.\nPlatforms: web, android\nTopics: imitation-learning, planning, screen-understanding\n\nRecorded specifications: {\"availability\": \"Ungated mixture/post-training JSON and image-archive manifests; some per-domain images and training code remain pending.\", \"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md\", \"https://raw.githubusercontent.com/hkust-nlp/GUIMid/HEAD/README.md\"], \"formats\": [\"JSON\", \"tar.gz image archives\"], \"modalities\": [\"text\", \"images\"], \"notes\": \"Do not treat Hub size bands or the inconsistent introductory task count as a sample total. Some constituent-domain image packages remain forthcoming.\", \"splits\": [\"train configurations for named mid-training domains\"], \"supervision\": [\"mid-training instruction supervision\", \"GUI post-training trajectories\"]}, \"gated\": false, \"hub_id\": \"hkust-nlp/GUIMid\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/hkust-nlp/GUIMid/raw/main/README.md\"], \"reason\": \"Publisher dataset-card declaration for the released packaging; original source datasets retain their own terms.\"}}, \"release_revision\": \"b6a360815c843b5733674191b6e6d7fa68b2ea8c\"}\n\nThe public manifest includes GUIMid.json, mavis.tar.gz, GUI_trajectory.json, traj.tar.gz and domain-specific JSON files. The card says image data for some individual domains is still forthcoming and its overview inconsistently says nine tasks while later enumerating seven vision-language plus four language-only domains; no aggregate task/sample count is inferred. Training code remains marked coming soon in the repository. MIT is the publisher dataset declaration, not a relicensing of all upstream content.\n\nPrimary source: https://huggingface.co/datasets/hkust-nlp/GUIMid\nCatalog record: https://awesomecua.github.io/resources/url-422010da7a5c845a18f6/"}
{"id": "url:4784f63cd7117f64a2a3", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj", "citation_url": "https://awesomecua.github.io/resources/url-4784f63cd7117f64a2a3/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj", "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"], "formats": ["ZIP"], "modalities": ["screenshots", "accessibility trees", "text"], "notes": "Observation modes follow publisher archive labels rather than inspected record schemas. No count or training-use permission is inferred.", "splits": [], "supervision": ["evaluation rollout logs"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj/tree/main", "gated": false, "hub_id": "OS-Copilot/ScienceBoard-Traj", "languages": [], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0a97233bd9132fdddc6ddb26fd3cf9f2af06ae086429f367a61cfab0c16c4f44", "supports": ["name", "kind", "identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:35:58Z", "role": "primary", "sha256": "4c3fe93aa740dbf4a279a246ddd836f2590a2a36f39b820abe166e38cb484577", "supports": ["summary", "environments", "relations", "topics", "scope"], "url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "8ee436f67e708bb917992764bc60a017c9db2edb21bdb99356b35c69f73724d7", "supports": ["environments", "notes"], "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0a97233bd9132fdddc6ddb26fd3cf9f2af06ae086429f367a61cfab0c16c4f44", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj"}], "name": "ScienceBoard evaluation trajectories", "publication": null, "published_at": null, "record_sha256": "4b949e205116a5e0807fbfee2e3e7441c8821e13bd8b5cc23cd2d707a14d6193", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "resource_id": "url:4784f63cd7117f64a2a3", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj", "topics": ["evaluation", "science", "workflow"], "updated_at": null}, "text": "ScienceBoard evaluation trajectories\nResource type: dataset\nReview status: curated\nScope: core\n\nArchives of ScienceBoard agent evaluation runs, separated by agent configuration and screenshot, accessibility-tree or set-of-marks observation mode.\nPlatforms: desktop, linux\nTopics: evaluation, science, workflow\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/api/datasets/OS-Copilot/ScienceBoard-Traj\", \"https://raw.githubusercontent.com/OS-Copilot/ScienceBoard/HEAD/README.md\"], \"formats\": [\"ZIP\"], \"modalities\": [\"screenshots\", \"accessibility trees\", \"text\"], \"notes\": \"Observation modes follow publisher archive labels rather than inspected record schemas. No count or training-use permission is inferred.\", \"splits\": [], \"supervision\": [\"evaluation rollout logs\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/ScienceBoard-Traj\", \"modalities\": [\"image\", \"text\"]}\n\nThe manifest exposes six ZIP archives but no README card or data-license declaration. The official project announces evaluation-result trajectories; this record does not label them as a training corpus or infer a trajectory count from archive names. Internal schemas and complete payload contents were not inspected.\n\nPrimary source: https://huggingface.co/datasets/OS-Copilot/ScienceBoard-Traj\nCatalog record: https://awesomecua.github.io/resources/url-4784f63cd7117f64a2a3/"}
{"id": "url:4a0b1692589594745a3c", "metadata": {"aliases": ["OpenCUA 72B Preview", "OpenCUA-72B-preview", "url:8c792416d297293422cf", "xlangai/OpenCUA-72B-preview", "https://huggingface.co/xlangai/OpenCUA-72B-preview"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/xlangai/OpenCUA-72B", "citation_url": "https://awesomecua.github.io/resources/url-4a0b1692589594745a3c/", "details": {"availability": "Public ungated checkpoint with 82 safetensors shards and model index.", "base_model": "Qwen/Qwen2.5-VL-72B-Instruct", "disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-72B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-72B", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"], "reason": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized."}}, "release_revision": "3cc4af5337a1d0010f934284df3c482797266547"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ebe551af1008a2307d16cf68715fc1861f8c501c4981660dbe08f691cf73d787", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:12Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:35Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["identity", "notes"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:04Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B-preview"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:05Z", "role": "primary", "sha256": "ff0ee40df2f44a2b3aa1b8117c31752013a40fdb6a85d66cdc596e096583ac5d", "supports": ["aliases"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-72B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T09:14:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized.", "declared": "mit", "retrieved_at": "2026-10-17T09:14:12Z", "url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-72B"}], "name": "OpenCUA 72B", "publication": null, "published_at": null, "record_sha256": "130af9d8c2c32d3c079ecfc523fd5295ca48c8037a6fb45a51d39cac5cf3d489", "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "github:xlang-ai/opencua", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "resource_id": "url:4a0b1692589594745a3c", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/xlangai/OpenCUA-72B", "topics": ["grounding", "planning", "action-prediction", "self-reflection"], "updated_at": null}, "text": "OpenCUA 72B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-based computer-use checkpoint that combines screenshot history and reflective reasoning to predict executable desktop actions.\nRecorded aliases: OpenCUA 72B Preview, OpenCUA-72B-preview, url:8c792416d297293422cf, xlangai/OpenCUA-72B-preview, https://huggingface.co/xlangai/OpenCUA-72B-preview\nPlatforms: desktop, linux, windows, macos\nTopics: grounding, planning, action-prediction, self-reflection\n\nRecorded specifications: {\"availability\": \"Public ungated checkpoint with 82 safetensors shards and model index.\", \"base_model\": \"Qwen/Qwen2.5-VL-72B-Instruct\", \"disabled\": false, \"download_url\": \"https://huggingface.co/xlangai/OpenCUA-72B/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/OpenCUA-72B\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/xlangai/OpenCUA-72B/raw/main/README.md\"], \"reason\": \"Publisher model card explicitly declares MIT for the released model; other code and training-data scopes are not generalized.\"}}, \"release_revision\": \"3cc4af5337a1d0010f934284df3c482797266547\"}\n\nThe former OpenCUA-72B-preview API redirects to the canonical xlangai/OpenCUA-72B repository. The public manifest lists 82 safetensors weight shards plus an index and custom model/tokenizer code. The card declares MIT for the released model and describes absolute coordinates after smart resizing. Weight files and remote code were not downloaded or executed.\n\nPrimary source: https://huggingface.co/xlangai/OpenCUA-72B\nCatalog record: https://awesomecua.github.io/resources/url-4a0b1692589594745a3c/"}
{"id": "url:4a47a6df845735984bf1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://hcompany.ai/newsroom/meet-holotab", "citation_url": "https://awesomecua.github.io/resources/url-4a47a6df845735984bf1/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "7d67747fdfe89bcc69d2b60b41f4f2dccf758913ea6d491949cbeac20de372ef", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://hcompany.ai/newsroom/meet-holotab"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/newsroom/meet-holotab"}], "name": "HoloTab", "publication": null, "published_at": null, "record_sha256": "8af80533b1e6238ff558fee4de93ba2a01e84e0e4eed562022f9ad79c0456e75", "relations": [{"evidence_url": "https://hcompany.ai/newsroom/meet-holotab", "target_id": "organization:h", "type": "published_by"}], "resource_id": "url:4a47a6df845735984bf1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://hcompany.ai/newsroom/meet-holotab", "topics": ["web-navigation"], "updated_at": null}, "text": "HoloTab\nResource type: product\nReview status: curated\nScope: core\n\nH browser extension for executing tasks through website interfaces.\nPlatforms: web\nTopics: web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://hcompany.ai/newsroom/meet-holotab\nCatalog record: https://awesomecua.github.io/resources/url-4a47a6df845735984bf1/"}
{"id": "url:4d32ad3682ae2d71c178", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B", "citation_url": "https://awesomecua.github.io/resources/url-4d32ad3682ae2d71c178/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B/tree/main", "gated": false, "hub_id": "ByteDance-Seed/UI-TARS-1.5-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "73b8964fb8ae2c25009dd4621549a3549d357c4c5c766ef7ed441acaef7efbc9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B"}], "name": "UI-TARS 1.5 7B", "publication": null, "published_at": null, "record_sha256": "fae261b185079b41575a39a2a2e9d8a5b64c9e9a2389e7dd5d8d485da3d24723", "relations": [{"evidence_url": "https://huggingface.co/api/models/ByteDance-Seed/UI-TARS-1.5-7B", "target_id": "github:bytedance/ui-tars", "type": "variant_of"}], "resource_id": "url:4d32ad3682ae2d71c178", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B", "topics": ["grounding", "planning"], "updated_at": null}, "text": "UI-TARS 1.5 7B\nResource type: model\nReview status: curated\nScope: core\n\nUI-TARS 1.5 GUI-agent checkpoint in the Qwen2.5-VL architecture with about 8.3B parameters stored in FP32, released under Apache-2.0 for English image-text-to-text use in Transformers.\nPlatforms: cross-platform\nTopics: grounding, planning\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B/tree/main\", \"gated\": false, \"hub_id\": \"ByteDance-Seed/UI-TARS-1.5-7B\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/ByteDance-Seed/UI-TARS-1.5-7B\nCatalog record: https://awesomecua.github.io/resources/url-4d32ad3682ae2d71c178/"}
{"id": "url:4e71dc456c4112eed727", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL", "citation_url": "https://awesomecua.github.io/resources/url-4e71dc456c4112eed727/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "prompts", "value": 12600}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "screenshots", "value": 11205}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "population": "RL training corpus", "qualifier": "exact", "unit": "desktop applications", "value": 87}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "Publisher-reported counts cover a selection and reformatting of GroundCUA, not a new disjoint image collection. No independent image-overlap audit was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "target bounding boxes"]}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "34f0025f686d0e48e2baa5244056a4362018aaad77ac54abea7f1f6eb25c4276", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL"}], "name": "GroundGUI RL Dataset", "publication": null, "published_at": null, "record_sha256": "fef0e67025b973ce361d8f2d2681c99d59df38bff24a4ca85f150d4d3784c33f", "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "resource_id": "url:4e71dc456c4112eed727", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/PrentisAI/GroundGUI_RL", "topics": ["grounding", "reinforcement-learning", "data-generation"], "updated_at": null}, "text": "GroundGUI RL Dataset\nResource type: dataset\nReview status: curated\nScope: core\n\nDesktop grounding prompts selected from GroundCUA and formatted for click-based reinforcement learning with target-box rewards.\nPlatforms: desktop\nTopics: grounding, reinforcement-learning, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md\", \"population\": \"RL training corpus\", \"qualifier\": \"exact\", \"unit\": \"prompts\", \"value\": 12600}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md\", \"population\": \"RL training corpus\", \"qualifier\": \"exact\", \"unit\": \"screenshots\", \"value\": 11205}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md\", \"population\": \"RL training corpus\", \"qualifier\": \"exact\", \"unit\": \"desktop applications\", \"value\": 87}], \"evidence_urls\": [\"https://huggingface.co/datasets/PrentisAI/GroundGUI_RL/raw/main/README.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"coordinates\"], \"notes\": \"Publisher-reported counts cover a selection and reformatting of GroundCUA, not a new disjoint image collection. No independent image-overlap audit was performed.\", \"splits\": [\"train\"], \"supervision\": [\"instruction-to-click grounding\", \"target bounding boxes\"]}}\n\nThe card declares MIT and attributes the source examples to GroundCUA. Its separate claim about UI-Vision overlap is not independently validated by this catalog.\n\nPrimary source: https://huggingface.co/datasets/PrentisAI/GroundGUI_RL\nCatalog record: https://awesomecua.github.io/resources/url-4e71dc456c4112eed727/"}
{"id": "url:4f47abde8cd28a1ec1e1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live", "citation_url": "https://awesomecua.github.io/resources/url-4f47abde8cd28a1ec1e1/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live/tree/main", "gated": false, "hub_id": "stanfordnlp/nnetnav-live", "languages": [], "modalities": ["text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "d8f8571dfc583d3a46d5b3a6a285e77f17057b36a8111d416d711513aa4ecf16", "supports": ["name", "kind"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-live"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:34Z", "role": "primary", "sha256": "3095f7262b7d4ff916c6f58c55e6fc9902c08f77d09bf3951199945501148bb3", "supports": ["name", "summary", "environments", "topics", "scope", "relations"], "url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T11:11:53Z", "role": "primary", "sha256": "d8f8571dfc583d3a46d5b3a6a285e77f17057b36a8111d416d711513aa4ecf16", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/stanfordnlp/nnetnav-live"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:11:53Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:11:53Z", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live"}], "name": "NNetNav Live data", "publication": null, "published_at": null, "record_sha256": "70a809ef7cb7d4961ad3bb5a81e85aec87319d11c57982da73c003382860e1a6", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MurtyShikhar/NNetnav/HEAD/README.md", "target_id": "github:murtyshikhar/nnetnav", "type": "component_of"}], "resource_id": "url:4f47abde8cd28a1ec1e1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/stanfordnlp/nnetnav-live", "topics": ["data-generation", "web-navigation"], "updated_at": null}, "text": "NNetNav Live data\nResource type: dataset\nReview status: curated\nScope: core\n\nBrowser-interaction data released with NNetNav for its live-website setting, with separate training and test files.\nPlatforms: web\nTopics: data-generation, web-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/stanfordnlp/nnetnav-live/tree/main\", \"gated\": false, \"hub_id\": \"stanfordnlp/nnetnav-live\", \"modalities\": [\"text\"]}\n\nThe Hub card supplies a license but no detailed schema. Dataset purpose is supported by the project's README; exact sample counts and annotation quality remain unverified.\n\nPrimary source: https://huggingface.co/datasets/stanfordnlp/nnetnav-live\nCatalog record: https://awesomecua.github.io/resources/url-4f47abde8cd28a1ec1e1/"}
{"id": "url:511903ebfac45e6751b2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey", "citation_url": "https://awesomecua.github.io/resources/url-511903ebfac45e6751b2/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "population": "recommended hflqf88888/GUIOdyssey release", "qualifier": "exact", "unit": "episodes", "value": 8334}], "evidence_urls": ["https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md", "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"], "formats": ["JSON annotations", "PNG screenshots"], "modalities": ["screenshots", "instructions", "actions", "bounding boxes", "step context"], "notes": "The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.", "splits": ["random", "task", "device", "app"], "supervision": ["cross-app trajectories", "action grounding"]}, "download_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/tree/main", "sample_count": 8334, "sample_unit": "episodes reported by the current release card"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "4585b765d7c8b3e9bb6187d36d576edde454d4c0158a6aa8b3fdbedf34c89a15", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "78c19ee712487af60df1913f11572958564e59a1a2cc663b1b30d971855d7a28", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:39:52Z", "role": "primary", "sha256": "fd627ee0a6ee0596e0ee8147996b9a7a3436f754fc95ed63cf5f408d778b2831", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:39:52Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T11:39:52Z", "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey"}], "name": "GUIOdyssey — current data release", "publication": null, "published_at": null, "record_sha256": "64dd5b6eac53a45510ec7689fe55dd24d45a2595917e47d9270122e9e270ce6b", "relations": [{"evidence_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md", "target_id": "arxiv:2406.08451", "type": "documented_by"}], "resource_id": "url:511903ebfac45e6751b2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/hflqf88888/GUIOdyssey", "topics": ["mobile-navigation", "cross-app", "data-generation"], "updated_at": null}, "text": "GUIOdyssey — current data release\nResource type: dataset\nReview status: curated\nScope: core\n\nCross-app Android interaction episodes with fine-grained step instructions, screen descriptions, action context, and target boxes.\nPlatforms: android\nTopics: mobile-navigation, cross-app, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md\", \"population\": \"recommended hflqf88888/GUIOdyssey release\", \"qualifier\": \"exact\", \"unit\": \"episodes\", \"value\": 8334}], \"evidence_urls\": [\"https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/README.md\", \"https://huggingface.co/datasets/hflqf88888/GUIOdyssey/raw/main/README.md\", \"https://raw.githubusercontent.com/OpenGVLab/GUI-Odyssey/HEAD/introduction.md\"], \"formats\": [\"JSON annotations\", \"PNG screenshots\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\", \"bounding boxes\", \"step context\"], \"notes\": \"The repository gives conflicting episode totals (8,334 and 8,834). This profile uses 8,334 from the specifically recommended current dataset card. No independent archive recount was performed. Coordinates are normalized to 0–1000.\", \"splits\": [\"random\", \"task\", \"device\", \"app\"], \"supervision\": [\"cross-app trajectories\", \"action grounding\"]}, \"download_url\": \"https://huggingface.co/datasets/hflqf88888/GUIOdyssey/tree/main\", \"sample_count\": 8334, \"sample_unit\": \"episodes reported by the current release card\"}\n\nThe project recommends this release over its earlier OpenGVLab/GUI-Odyssey upload. Episode count follows the current data card; conflicting repository prose is documented in the profile.\n\nPrimary source: https://huggingface.co/datasets/hflqf88888/GUIOdyssey\nCatalog record: https://awesomecua.github.io/resources/url-511903ebfac45e6751b2/"}
{"id": "url:52a3261eb6a19b402e88", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Tianlin Shi", "Andrej Karpathy", "Linxi Fan", "Jonathan Hernandez", "Percy Liang"], "canonical_source_url": "https://proceedings.mlr.press/v70/shi17a.html", "citation_url": "https://awesomecua.github.io/resources/url-52a3261eb6a19b402e88/", "details": {"publication": {"authors": ["Tianlin Shi", "Andrej Karpathy", "Linxi Fan", "Jonathan Hernandez", "Percy Liang"], "date_basis": "Publisher citation_publication_date metadata explicitly states 2017/07/17; this is separate from the conference interval.", "date_parts": [2017, 7, 17], "doi": null, "event_dates": "06--11 Aug (conference interval reported in the publisher BibTeX month field)", "evidence_urls": ["https://proceedings.mlr.press/v70/shi17a.html", "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained. DOI not supplied in the retained publisher citation. Publisher metadata supplies the publication date separately from the conference interval.", "pages": "3135-3144", "pdf_url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "World of Bits: An Open-Domain Platform for Web-Based Agents", "type": "conference-paper", "url": "https://proceedings.mlr.press/v70/shi17a.html", "venue": "Proceedings of the 34th International Conference on Machine Learning", "volume": "70"}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "749d2a150e4c2bcde1b219a0ecb86c95546134a057892544e8495b5e45323c44", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.event_dates", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.publisher", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume"], "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "087008efb9d056df5fd9fbc724413d6a3f20cb14c7f1597d00bacac53d8342d6", "supports": ["details.publication.notes", "details.publication.pdf_url"], "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "749d2a150e4c2bcde1b219a0ecb86c95546134a057892544e8495b5e45323c44", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.series", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "details.publication.volume", "kind", "name", "scope", "summary", "topics"], "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:39Z", "role": "primary", "sha256": "087008efb9d056df5fd9fbc724413d6a3f20cb14c7f1597d00bacac53d8342d6", "supports": ["details.publication.notes", "details.publication.pdf_url", "environments"], "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}], "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://proceedings.mlr.press/v70/shi17a.html"}, {"rel": "publication_pdf", "url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"}], "name": "World of Bits: An Open-Domain Platform for Web-Based Agents", "publication": {"authors": ["Tianlin Shi", "Andrej Karpathy", "Linxi Fan", "Jonathan Hernandez", "Percy Liang"], "date_basis": "Publisher citation_publication_date metadata explicitly states 2017/07/17; this is separate from the conference interval.", "date_parts": [2017, 7, 17], "doi": null, "event_dates": "06--11 Aug (conference interval reported in the publisher BibTeX month field)", "evidence_urls": ["https://proceedings.mlr.press/v70/shi17a.html", "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained. DOI not supplied in the retained publisher citation. Publisher metadata supplies the publication date separately from the conference interval.", "pages": "3135-3144", "pdf_url": "https://proceedings.mlr.press/v70/shi17a/shi17a.pdf", "publisher": "PMLR", "series": "Proceedings of Machine Learning Research", "title": "World of Bits: An Open-Domain Platform for Web-Based Agents", "type": "conference-paper", "url": "https://proceedings.mlr.press/v70/shi17a.html", "venue": "Proceedings of the 34th International Conference on Machine Learning", "volume": "70"}, "published_at": "2017-07-17", "record_sha256": "c8b28333b0a0f9b95bd44f78e0c71099972c085161fd4ef8ef88dba60513f91e", "relations": [], "resource_id": "url:52a3261eb6a19b402e88", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://proceedings.mlr.press/v70/shi17a.html", "topics": ["web-navigation", "reinforcement-learning", "evaluation", "data-generation"], "updated_at": null}, "text": "World of Bits: An Open-Domain Platform for Web-Based Agents\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces World of Bits, a web-task platform combining human demonstrations, keyboard and mouse actions, and cached HTTP traffic for repeatable task environments. It evaluates agents trained through behavioral cloning and reinforcement learning.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning, evaluation, data-generation\nPublished: 2017-07-17\nAuthors: Tianlin Shi, Andrej Karpathy, Linxi Fan, Jonathan Hernandez, Percy Liang\n\nPublication edition: World of Bits: An Open-Domain Platform for Web-Based Agents\nPublication authors: Tianlin Shi, Andrej Karpathy, Linxi Fan, Jonathan Hernandez, Percy Liang\nPublication edition date (recorded precision): 2017-07-17\nPublication venue: Proceedings of the 34th International Conference on Machine Learning\nPublication publisher: PMLR\nPublication volume: 70\nPublication pages: 3135-3144\nPublication date basis: Publisher citation_publication_date metadata explicitly states 2017/07/17; this is separate from the conference interval.\nPublication event dates: 06--11 Aug (conference interval reported in the publisher BibTeX month field)\nPublication notes: Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained. DOI not supplied in the retained publisher citation. Publisher metadata supplies the publication date separately from the conference interval.\nPublication source: https://proceedings.mlr.press/v70/shi17a.html\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. Publisher citation uses Tianlin Shi and Linxi Fan; the PDF also prints the nicknames Tim and Jim. The ordered publisher author names are retained.\n\nPrimary source: https://proceedings.mlr.press/v70/shi17a.html\nCatalog record: https://awesomecua.github.io/resources/url-52a3261eb6a19b402e88/"}
{"id": "url:52b7de0e43b1afc9c226", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B", "citation_url": "https://awesomecua.github.io/resources/url-52b7de0e43b1afc9c226/", "details": {"disabled": false, "download_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/tree/main", "gated": false, "hub_id": "SenseLLM/SpiritSight-Agent-8B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "cc2564a3b2011a06f435e16b95cbf1e22072fd39e76fba844657773541586b37", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/SenseLLM/SpiritSight-Agent-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "916bde91ff00aba22ff13ed3d8b4ca679bc8c198632f1d3bc9469660287cb5d3", "supports": ["licenses.weights"], "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B"}], "name": "SpiritSight-Agent 8B base", "publication": null, "published_at": null, "record_sha256": "eead6752395a0faa07ff8ea5759689487564e3cf58706589db77eabf71fe1dbb", "relations": [{"evidence_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/raw/main/README.md", "target_id": "arxiv:2503.03196", "type": "documented_by"}], "resource_id": "url:52b7de0e43b1afc9c226", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/SenseLLM/SpiritSight-Agent-8B", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "SpiritSight-Agent 8B base\nResource type: model\nReview status: curated\nScope: core\n\nAn InternVL2-based visual GUI checkpoint distributed for grounding and interface navigation, with the publisher recommending task-specific fine-tuning of the base release.\nPlatforms: cross-platform\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/SenseLLM/SpiritSight-Agent-8B/tree/main\", \"gated\": false, \"hub_id\": \"SenseLLM/SpiritSight-Agent-8B\"}\n\nThe Hub manifest exposes four safetensors shards under SpiritSight-Agent-8B-base. The card declares Apache-2.0 and separately links InternVL terms in its model table. This is the published base checkpoint; benchmark-specific fine-tuning is discussed separately. No model code was executed.\n\nPrimary source: https://huggingface.co/SenseLLM/SpiritSight-Agent-8B\nCatalog record: https://awesomecua.github.io/resources/url-52b7de0e43b1afc9c226/"}
{"id": "url:546e6c95825d1318c98d", "metadata": {"aliases": ["url:a778c3411858c38c0f39", "OpenMobile-2/OpenMobile-Data", "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2", "citation_url": "https://awesomecua.github.io/resources/url-546e6c95825d1318c98d/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "SFT collection", "qualifier": "exact", "unit": "trajectories", "value": 11643}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "SFT collection", "qualifier": "exact", "unit": "decision steps", "value": 118582}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md", "population": "RL collection", "qualifier": "exact", "unit": "executable tasks", "value": 2475}], "evidence_urls": ["https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md"], "formats": ["JSON", "JSONL", "PNG"], "modalities": ["screenshots", "text", "tool calls"], "notes": "These are publisher-reported totals at three distinct grains, not values to sum. Earlier screenshots may be collapsed; each SFT example keeps at most three images. The card and repository disagree on the singular/plural RL filename, so consumers should inspect the actual file manifest.", "splits": ["sft/train", "rl/train"], "supervision": ["next-action supervision", "verifiable task outcomes"]}}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "abacf9a10759a9e7a8eeec5a266c96ed83742bd9296a19be95d09c9717e18b7c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "485279038266d9454df75e27a80a0a0d71c24e90700126d2dcfa0da921b1273d", "supports": ["relations", "environments"], "url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:18Z", "role": "primary", "sha256": "14b1211ed709147f4022f311c17303ca70faaffa22be143617eb431d966eec1a", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/OpenMobile-2/OpenMobile-Data"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:20Z", "role": "primary", "sha256": "14b1211ed709147f4022f311c17303ca70faaffa22be143617eb431d966eec1a", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/OpenMobile-2/OpenMobile-Data-v2"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2"}], "name": "OpenMobile-Data-v2", "publication": null, "published_at": null, "record_sha256": "28360ef4b5df0ffb6ea25d4151209411f8348429e43752d2280e8f8834a34fc1", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenMobile2/OpenMobile2-Code/HEAD/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "resource_id": "url:546e6c95825d1318c98d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2", "topics": ["mobile-navigation", "data-generation", "reinforcement-learning", "tool-integration"], "updated_at": null}, "text": "OpenMobile-Data-v2\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile-agent supervision from emulator and simulator trajectories, plus executable tasks with starting states and success checks for reinforcement learning.\nRecorded aliases: url:a778c3411858c38c0f39, OpenMobile-2/OpenMobile-Data, https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data\nPlatforms: android, web\nTopics: mobile-navigation, data-generation, reinforcement-learning, tool-integration\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md\", \"population\": \"SFT collection\", \"qualifier\": \"exact\", \"unit\": \"trajectories\", \"value\": 11643}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md\", \"population\": \"SFT collection\", \"qualifier\": \"exact\", \"unit\": \"decision steps\", \"value\": 118582}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md\", \"population\": \"RL collection\", \"qualifier\": \"exact\", \"unit\": \"executable tasks\", \"value\": 2475}], \"evidence_urls\": [\"https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2/raw/main/README.md\"], \"formats\": [\"JSON\", \"JSONL\", \"PNG\"], \"modalities\": [\"screenshots\", \"text\", \"tool calls\"], \"notes\": \"These are publisher-reported totals at three distinct grains, not values to sum. Earlier screenshots may be collapsed; each SFT example keeps at most three images. The card and repository disagree on the singular/plural RL filename, so consumers should inspect the actual file manifest.\", \"splits\": [\"sft/train\", \"rl/train\"], \"supervision\": [\"next-action supervision\", \"verifiable task outcomes\"]}}\n\nThe reviewed card contains no explicit license declaration, so data-license status remains unknown. The repository code license is not inherited. The current code README points to the v2 dataset; an older project-page URL was not counted as a separate asset.\n\nPrimary source: https://huggingface.co/datasets/OpenMobile-2/OpenMobile-Data-v2\nCatalog record: https://awesomecua.github.io/resources/url-546e6c95825d1318c98d/"}
{"id": "url:54997a3c67e54abdf319", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/neulab/UIX-Qwen2", "citation_url": "https://awesomecua.github.io/resources/url-54997a3c67e54abdf319/", "details": {"disabled": false, "download_url": "https://huggingface.co/neulab/UIX-Qwen2/tree/main", "gated": false, "hub_id": "neulab/UIX-Qwen2"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "19bd4d126f3ccb4cb3049125d384c76964bd3169c051369ac0c7b64cef9f48b4", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/neulab/UIX-Qwen2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "fe0c03fe73529a29788d559103c4aed48dca6ed9ba3bfd5f151f902fa8858b38", "supports": ["identity", "summary"], "url": "https://raw.githubusercontent.com/neulab/multiui/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "91f90e329b76c2ded26a62c41f648a73e685116aa732c86c201592e46a941c44", "supports": ["licenses.weights"], "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "odc-by", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/neulab/UIX-Qwen2"}], "name": "UIX-Qwen2", "publication": null, "published_at": null, "record_sha256": "ee16c457fbd222baf56810d86531dd9e906a9f1a5c84f0337f02b465e1e168d2", "relations": [{"evidence_url": "https://huggingface.co/neulab/UIX-Qwen2/raw/main/README.md", "target_id": "arxiv:2410.13824", "type": "documented_by"}], "resource_id": "url:54997a3c67e54abdf319", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/neulab/UIX-Qwen2", "topics": ["screen-understanding", "grounding"], "updated_at": null}, "text": "UIX-Qwen2\nResource type: model\nReview status: curated\nScope: core\n\nA MultiUI-trained visual-language checkpoint for text-rich webpage understanding and interface tasks.\nPlatforms: web\nTopics: screen-understanding, grounding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/neulab/UIX-Qwen2/tree/main\", \"gated\": false, \"hub_id\": \"neulab/UIX-Qwen2\"}\n\nFour public safetensors shards are present. The model card declares odc-by, a data-oriented license designation; it is preserved as a publisher declaration rather than replaced with the code or base-model license. No independent legal scope or evaluation result is asserted.\n\nPrimary source: https://huggingface.co/neulab/UIX-Qwen2\nCatalog record: https://awesomecua.github.io/resources/url-54997a3c67e54abdf319/"}
{"id": "url:56bda04bd8d3a438deb3", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef", "citation_url": "https://awesomecua.github.io/resources/url-56bda04bd8d3a438deb3/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "ScreenRef image inventory", "qualifier": "exact", "unit": "distinct screenshots", "value": 585981}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "refusal subset of full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 22475}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "coordinates"], "notes": "The publisher reports these inventories in the companion annotation card. The refusal subset is included in the full recipe. Screenshot access is gated; no image download or independent corpus recount was performed.", "splits": ["train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef/tree/main", "gated": "auto", "hub_id": "PrentisAI/ScreenRef", "languages": ["en"], "license_name": "restricted-research-only", "modalities": ["image", "text"], "size_category_reported": ["100K<n<1M"]}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["name", "kind", "details.gated", "details.hub_id", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "634b479c7d089e50735f6d4a25f78a63f4d0c2a8c1955015e7641b7afd61c317", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/PrentisAI/ScreenRef"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "other", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef"}], "name": "ScreenRef", "publication": null, "published_at": null, "record_sha256": "e0a3812c93bc8cba20ac3185bd564ea186b9ae543d0c658aa7eb9b0e25659327", "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "resource_id": "url:56bda04bd8d3a438deb3", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef", "topics": ["grounding", "data-generation", "screen-understanding"], "updated_at": null}, "text": "ScreenRef\nResource type: dataset\nReview status: curated\nScope: core\n\nGUI screenshots paired with localization instructions and targets, including absent-target examples; image access is gated under publisher terms.\nPlatforms: desktop, web, android, linux\nTopics: grounding, data-generation, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\", \"population\": \"ScreenRef image inventory\", \"qualifier\": \"exact\", \"unit\": \"distinct screenshots\", \"value\": 585981}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\", \"population\": \"full recipe\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 633369}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\", \"population\": \"refusal subset of full recipe\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 22475}], \"evidence_urls\": [\"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"coordinates\"], \"notes\": \"The publisher reports these inventories in the companion annotation card. The refusal subset is included in the full recipe. Screenshot access is gated; no image download or independent corpus recount was performed.\", \"splits\": [\"train\"], \"supervision\": [\"instruction-to-click grounding\", \"bounding boxes\", \"absent-target abstention\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef/tree/main\", \"gated\": \"auto\", \"hub_id\": \"PrentisAI/ScreenRef\", \"languages\": [\"en\"], \"license_name\": \"restricted-research-only\", \"modalities\": [\"image\", \"text\"], \"size_category_reported\": [\"100K<n<1M\"]}\n\nThe public Hub API reports gated=auto and a Restricted Research-Only declaration. The companion card distinguishes annotations from screenshots and limits reuse. Access terms were inspected as public metadata; no access agreement was accepted or gated files downloaded.\n\nPrimary source: https://huggingface.co/datasets/PrentisAI/ScreenRef\nCatalog record: https://awesomecua.github.io/resources/url-56bda04bd8d3a438deb3/"}
{"id": "url:570654ff67ecefa2b955", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/runamu/MONDAY", "citation_url": "https://awesomecua.github.io/resources/url-570654ff67ecefa2b955/", "details": {"availability": "Ungated Parquet annotation files for train, validation, test and test_unseen_os; videos external.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "video annotation records", "value": 19725}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "validation split", "qualifier": "exact", "unit": "video annotation records", "value": 495}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "video annotation records", "value": 100}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "test_unseen_os split", "qualifier": "exact", "unit": "video annotation records", "value": 50}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "publisher overview of MONDAY", "qualifier": "approximately", "unit": "detected scenes", "value": 333000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "population": "publisher overview of MONDAY", "qualifier": "approximately", "unit": "identified actions", "value": 313000}], "evidence_urls": ["https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["text", "timestamps", "bounding boxes", "action labels"], "notes": "Split counts are video-annotation rows; scene and action totals are rounded publisher reports and are not added to row counts. Source videos/images are not included in the Hub files.", "splits": ["train", "validation", "test", "test_unseen_os"], "supervision": ["video-derived scene and action annotations"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md", "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"], "reason": "No dataset-license declaration in the inspected card; Apache-2.0 applies to companion code and is not transferred to YouTube content or this annotation package."}}, "release_revision": "d7bd929cf629bf0767f154cbdd984eea1eba01e9"}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "62a346fd0d328deb5b56a652c8581850b0c8120f5c77e53b33f24b9baf937adb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "f55c319a70c8ff988049363dc8fc266a9f1b1a09b8ac7096ee573f962e8f9492", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/runamu/MONDAY"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:53Z", "role": "primary", "sha256": "cb3d2329571a80b6edeab98c04f8b63c7f4e976929788d6a8283238b0b7fdadc", "supports": ["notes", "details.dataset_profile", "licenses.data"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:09Z", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["notes", "licenses.data"], "url": "https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/runamu/MONDAY"}], "name": "MONDAY annotations", "publication": null, "published_at": null, "record_sha256": "6ecca0c71e3240ba4684cc236a9f30e80745e38a6d501dd91f8db348d43cac14", "relations": [{"evidence_url": "https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md", "target_id": "arxiv:2505.12632", "type": "documented_by"}], "resource_id": "url:570654ff67ecefa2b955", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/runamu/MONDAY", "topics": ["data-generation", "mobile-navigation", "video", "imitation-learning"], "updated_at": null}, "text": "MONDAY annotations\nResource type: dataset\nReview status: curated\nScope: core\n\nVideo-derived mobile navigation annotations that pair scene boundaries and screen or element boxes with action labels for training and evaluating GUI agents.\nPlatforms: android, ios\nTopics: data-generation, mobile-navigation, video, imitation-learning\n\nRecorded specifications: {\"availability\": \"Ungated Parquet annotation files for train, validation, test and test_unseen_os; videos external.\", \"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"train split\", \"qualifier\": \"exact\", \"unit\": \"video annotation records\", \"value\": 19725}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"validation split\", \"qualifier\": \"exact\", \"unit\": \"video annotation records\", \"value\": 495}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"test split\", \"qualifier\": \"exact\", \"unit\": \"video annotation records\", \"value\": 100}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"test_unseen_os split\", \"qualifier\": \"exact\", \"unit\": \"video annotation records\", \"value\": 50}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"publisher overview of MONDAY\", \"qualifier\": \"approximately\", \"unit\": \"detected scenes\", \"value\": 333000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"population\": \"publisher overview of MONDAY\", \"qualifier\": \"approximately\", \"unit\": \"identified actions\", \"value\": 313000}], \"evidence_urls\": [\"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"https://raw.githubusercontent.com/runamu/monday/HEAD/README.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"text\", \"timestamps\", \"bounding boxes\", \"action labels\"], \"notes\": \"Split counts are video-annotation rows; scene and action totals are rounded publisher reports and are not added to row counts. Source videos/images are not included in the Hub files.\", \"splits\": [\"train\", \"validation\", \"test\", \"test_unseen_os\"], \"supervision\": [\"video-derived scene and action annotations\"]}, \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/runamu/MONDAY/raw/main/README.md\", \"https://raw.githubusercontent.com/runamu/monday/HEAD/README.md\", \"https://raw.githubusercontent.com/runamu/monday/HEAD/LICENSE\"], \"reason\": \"No dataset-license declaration in the inspected card; Apache-2.0 applies to companion code and is not transferred to YouTube content or this annotation package.\"}}, \"release_revision\": \"d7bd929cf629bf0767f154cbdd984eea1eba01e9\"}\n\nThe ungated Hub publishes four Parquet splits containing video identifiers and annotations. Original videos and extracted screenshots must be obtained separately; the code README explains that consumers download source videos from YouTube. A Windows Mobile unseen-OS test split is documented but is not labeled desktop Windows here. The companion code has an Apache-2.0 file, while the data card declares no data license and the repository defers third-party materials to original terms. Training scripts, pretrained checkpoints and the full generation pipeline remain planned updates.\n\nPrimary source: https://huggingface.co/datasets/runamu/MONDAY\nCatalog record: https://awesomecua.github.io/resources/url-570654ff67ecefa2b955/"}
{"id": "url:57f17b7d0c18b2d7c8bf", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool", "citation_url": "https://awesomecua.github.io/resources/url-57f17b7d0c18b2d7c8bf/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "de16b4fd611982cfaa160cb0fbdc1b0f0559b90983e452964527704a6de4e72d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool"}], "name": "Claude Computer Use", "publication": null, "published_at": null, "record_sha256": "ecdd8fc06b08675bf72bb4cf26f741c110bc7cc6f6cf9dfa18c7f4eed1862032", "relations": [{"evidence_url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool", "target_id": "organization:anthropic", "type": "published_by"}], "resource_id": "url:57f17b7d0c18b2d7c8bf", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool", "topics": ["planning", "grounding"], "updated_at": null}, "text": "Claude Computer Use\nResource type: model\nReview status: curated\nScope: core\n\nClaude tool capability for screenshot-based computer interaction.\nPlatforms: cross-platform\nTopics: planning, grounding\n\nPrimary source: https://docs.anthropic.com/en/docs/agents-and-tools/tool-use/computer-use-tool\nCatalog record: https://awesomecua.github.io/resources/url-57f17b7d0c18b2d7c8bf/"}
{"id": "url:5857c840372512cfe3fc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/docs/smolagents/examples/web_browser", "citation_url": "https://awesomecua.github.io/resources/url-5857c840372512cfe3fc/", "details": {"license_review": {"document": {"evidence_urls": ["https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/docs/source/en/examples/web_browser.md", "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/LICENSE"], "reason": "The retained repository-root file contains Apache-2.0 terms, but the reviewed Markdown has no license notice and the retained sources do not explicitly establish that this license applies to the tutorial document. Document reuse terms remain unknown; this is not a claim that no license exists. Code, data, weights and hosted-service terms are not inferred."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:59.429802+00:00", "role": "primary", "sha256": "0bd3804794c37f263db5cba3429523721854e95fe1477127f9fb83d4d0510f50", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "licenses.document"], "url": "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/docs/source/en/examples/web_browser.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:14.002361+00:00", "role": "primary", "sha256": "736281d4d96fc310050a09e830801f739729cc01217fef0be1fd45874e4c68f6", "supports": ["identity", "name", "summary", "environments", "topics", "scope", "notes"], "url": "https://huggingface.co/docs/smolagents/v1.26.0/examples/web_browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:17.450466+00:00", "role": "primary", "sha256": "cf08eff3ce4ca2fd85f85ba1eebca477277bca0cb2a54613e172741392061365", "supports": ["notes"], "url": "https://api.github.com/repos/huggingface/smolagents/git/ref/tags/v1.26.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:51:48.091044+00:00", "role": "primary", "sha256": "9def26af20ef4a6c4091e5873b53737940b360e6ff7db2d606add92d6586c8fb", "supports": ["notes"], "url": "https://huggingface.co/docs/smolagents/v1.8.0/examples/web_browser"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:43:02.729855+00:00", "role": "primary", "sha256": "c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4", "supports": ["licenses.document"], "url": "https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/LICENSE"}], "identifiers": {}, "kind": "learning", "last_verified_at": "2026-10-20T13:42:59.429802+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/docs/smolagents/examples/web_browser"}], "name": "Web Browser Automation with Agents (smolagents tutorial)", "publication": null, "published_at": null, "record_sha256": "68ed1c459adb3bac75c4497377d00a15af195205614614f15eeb0c5d32f0d2d3", "relations": [], "resource_id": "url:5857c840372512cfe3fc", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/docs/smolagents/examples/web_browser", "topics": ["web-navigation", "screen-understanding"], "updated_at": null}, "text": "Web Browser Automation with Agents (smolagents tutorial)\nResource type: learning\nReview status: curated\nScope: core\n\nVersioned tutorial building a vision-guided CodeAgent with Selenium/Helium browser tools, screenshot callbacks and example web-navigation tasks.\nPlatforms: web\nTopics: web-navigation, screen-understanding\n\nRecorded specifications: {\"license_review\": {\"document\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/docs/source/en/examples/web_browser.md\", \"https://raw.githubusercontent.com/huggingface/smolagents/12c1bc820eca50ace6f80a21d90426d41d74f845/LICENSE\"], \"reason\": \"The retained repository-root file contains Apache-2.0 terms, but the reviewed Markdown has no license notice and the retained sources do not explicitly establish that this license applies to the tutorial document. Document reuse terms remain unknown; this is not a claim that no license exists. Code, data, weights and hosted-service terms are not inferred.\"}}}\n\nThe reviewed v1.26.0 page matches all eight code blocks in its commit-pinned Markdown source. It uses InferenceClientModel and a vision model, unlike the retained v1.8.0 example. The shown setup starts a local visible Chrome session, and its installation command does not pin dependency versions. Runtime API compatibility, model-provider access and example outcomes were not tested. The inspected repository root contains Apache-2.0 terms, but explicit applicability to this Markdown tutorial was not established from retained sources; document reuse terms remain unknown. Model weights, visited websites and hosted-service terms are separate. No installation or notebook execution occurred.\n\nPrimary source: https://huggingface.co/docs/smolagents/examples/web_browser\nCatalog record: https://awesomecua.github.io/resources/url-5857c840372512cfe3fc/"}
{"id": "url:58a209f2b3b970916396", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth", "citation_url": "https://awesomecua.github.io/resources/url-58a209f2b3b970916396/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "population": "collection described in the dataset card; not an independently counted release-row total", "qualifier": "more-than", "unit": "tasks", "value": 6000}], "evidence_urls": ["https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "https://huggingface.co/api/datasets/sunblaze-ucb/AgentSynth"], "formats": ["JSONL", "ZIP trajectory packages"], "modalities": ["screenshots", "task instructions", "LLM thoughts", "commands"], "notes": "Card maps osworld/train to all_tasks.jsonl and insta/train to task_sequences.jsonl. Six difficulty levels and trajectory screenshots/thoughts/commands are documented. ZIP file labels are not trajectory counts; configurations and composed-task levels are not assumed disjoint.", "splits": ["train (osworld configuration)", "train (insta configuration)"], "supervision": ["synthetic tasks", "agent trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/tree/9ec3f3139f9786e81aee14b7fff1eb487cc7e991", "gated": false, "hub_id": "sunblaze-ucb/AgentSynth", "hub_revision": "9ec3f3139f9786e81aee14b7fff1eb487cc7e991", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/sunblaze-ucb/AgentSynth"}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "c1669d8e3caa27f8cd47c17b243b6a8d42807438d8a6a36074668adc57d33757", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:20Z", "role": "primary", "sha256": "0d3797d75dd4089ece8c8d4624f2a726024d25be255255d8770f78de21d75906", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/sunblaze-ucb/AgentSynth"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "42e8b8dead7caea3f8b21e93a075e8c895dfc30a25e1470bf97a90940bbf0106", "supports": ["details.repository_url"], "url": "https://raw.githubusercontent.com/sunblaze-ucb/AgentSynth/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth"}], "name": "AgentSynth tasks and trajectories", "publication": null, "published_at": null, "record_sha256": "300cad97f5f3ff3501b56565b2b32db1689d8c15c81c6f7932cc2b19b6b85f80", "relations": [{"evidence_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md", "target_id": "arxiv:2506.14205", "type": "documented_by"}], "resource_id": "url:58a209f2b3b970916396", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/sunblaze-ucb/AgentSynth", "topics": ["data-generation", "long-horizon", "imitation-learning"], "updated_at": null}, "text": "AgentSynth tasks and trajectories\nResource type: dataset\nReview status: curated\nScope: core\n\nPublished AgentSynth task definitions and recorded GUI trajectories, with OSWorld and InSTA training configurations and packaged screenshots, thoughts and commands.\nPlatforms: desktop, web\nTopics: data-generation, long-horizon, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md\", \"population\": \"collection described in the dataset card; not an independently counted release-row total\", \"qualifier\": \"more-than\", \"unit\": \"tasks\", \"value\": 6000}], \"evidence_urls\": [\"https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md\", \"https://huggingface.co/api/datasets/sunblaze-ucb/AgentSynth\"], \"formats\": [\"JSONL\", \"ZIP trajectory packages\"], \"modalities\": [\"screenshots\", \"task instructions\", \"LLM thoughts\", \"commands\"], \"notes\": \"Card maps osworld/train to all_tasks.jsonl and insta/train to task_sequences.jsonl. Six difficulty levels and trajectory screenshots/thoughts/commands are documented. ZIP file labels are not trajectory counts; configurations and composed-task levels are not assumed disjoint.\", \"splits\": [\"train (osworld configuration)\", \"train (insta configuration)\"], \"supervision\": [\"synthetic tasks\", \"agent trajectories\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/tree/9ec3f3139f9786e81aee14b7fff1eb487cc7e991\", \"gated\": false, \"hub_id\": \"sunblaze-ucb/AgentSynth\", \"hub_revision\": \"9ec3f3139f9786e81aee14b7fff1eb487cc7e991\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/sunblaze-ucb/AgentSynth/raw/9ec3f3139f9786e81aee14b7fff1eb487cc7e991/README.md\"], \"reason\": \"Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.\"}}, \"repository_url\": \"https://github.com/sunblaze-ucb/AgentSynth\"}\n\nPublic Hub manifest contains JSONL task files and trajectory ZIP packages. Counts describe the publisher-reported task collection, not independently counted archive rows or numbers inferred from ZIP filenames.\n\nPrimary source: https://huggingface.co/datasets/sunblaze-ucb/AgentSynth\nCatalog record: https://awesomecua.github.io/resources/url-58a209f2b3b970916396/"}
{"id": "url:5d61c1b5fc8dd650518d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/yifeizhou/pae-llava-7b", "citation_url": "https://awesomecua.github.io/resources/url-5d61c1b5fc8dd650518d/", "details": {"disabled": false, "download_url": "https://huggingface.co/yifeizhou/pae-llava-7b/tree/main", "gated": false, "hub_id": "yifeizhou/pae-llava-7b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c0994582f81415c9b0d37e4e5018752e48e511178c4b3898c207a042992db845", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "1b76d2aebbe15954eff2b8859b884ec6504f8b67ed0b1045d8b3a83c0d132d3a", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/yifeizhou/pae-llava-7b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "3e5f24b3175afeba8635dacc7b66eba1ce89fea49996a20281ff2df5a41f1e24", "supports": ["environments", "summary"], "url": "https://raw.githubusercontent.com/amazon-science/PAE/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c0994582f81415c9b0d37e4e5018752e48e511178c4b3898c207a042992db845", "supports": ["licenses.weights"], "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-nc-4.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/yifeizhou/pae-llava-7b"}], "name": "PAE LLaVA 7B", "publication": null, "published_at": null, "record_sha256": "4843d0d6a71941171650108364e3f14f254e429303cdd71ffc0ba9241ad60268", "relations": [{"evidence_url": "https://huggingface.co/yifeizhou/pae-llava-7b/raw/main/README.md", "target_id": "arxiv:2412.13194", "type": "documented_by"}], "resource_id": "url:5d61c1b5fc8dd650518d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/yifeizhou/pae-llava-7b", "topics": ["web-navigation", "reinforcement-learning"], "updated_at": null}, "text": "PAE LLaVA 7B\nResource type: model\nReview status: curated\nScope: core\n\nA released LLaVA checkpoint for visual browser interaction from the PAE autonomous task-proposal and policy-learning system.\nPlatforms: web\nTopics: web-navigation, reinforcement-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/yifeizhou/pae-llava-7b/tree/main\", \"gated\": false, \"hub_id\": \"yifeizhou/pae-llava-7b\"}\n\nThe model is publicly ungated and has seven safetensors shards plus tokenizer and configuration files. Its card declares CC-BY-NC-4.0 for this model. This weight declaration does not settle the separate PAE dataset or code licensing questions.\n\nPrimary source: https://huggingface.co/yifeizhou/pae-llava-7b\nCatalog record: https://awesomecua.github.io/resources/url-5d61c1b5fc8dd650518d/"}
{"id": "url:5e2bd1b3398974b79c9c", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Hcompany/WebClick", "citation_url": "https://awesomecua.github.io/resources/url-5e2bd1b3398974b79c9c/", "details": {"benchmark_protocol": {"actions": ["click target or bounding box"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "population": "test split", "qualifier": "exact", "unit": "examples", "value": 1639}], "evaluation": "Assess localization against annotated targets; agentbrowse, humanbrowse and calendars are distinct source buckets.", "evidence_urls": ["https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"], "focus": ["instruction-conditioned web localization"], "notes": "The card reports more than 100 websites but does not provide a precise website total. Grounding accuracy does not establish task completion.", "observations": ["web screenshots", "instructions"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/WebClick/tree/main", "gated": false, "hub_id": "Hcompany/WebClick"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "958bbf0eb9f61afd6c576673fbe6f7d84e6c4b8a9d4d8c4fc0e1224b5d8da30b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "c9576957a3cb14785bf274691b50f90b04567a442ed60ee8c39f0f73108359f4", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/WebClick"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/WebClick"}], "name": "WebClick", "publication": null, "published_at": null, "record_sha256": "43b21d15f638ea8f0053b7bc3657e7f449b023468dcb23b0b85e31aba93f78d5", "relations": [{"evidence_url": "https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "resource_id": "url:5e2bd1b3398974b79c9c", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Hcompany/WebClick", "topics": ["grounding", "evaluation", "web-navigation"], "updated_at": null}, "text": "WebClick\nResource type: benchmark\nReview status: curated\nScope: core\n\nA web-interface localization benchmark with screenshot instructions and target boxes, including agent browsing, human browsing and calendar cases.\nPlatforms: web\nTopics: grounding, evaluation, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"click target or bounding box\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md\", \"population\": \"test split\", \"qualifier\": \"exact\", \"unit\": \"examples\", \"value\": 1639}], \"evaluation\": \"Assess localization against annotated targets; agentbrowse, humanbrowse and calendars are distinct source buckets.\", \"evidence_urls\": [\"https://huggingface.co/datasets/Hcompany/WebClick/raw/main/README.md\"], \"focus\": [\"instruction-conditioned web localization\"], \"notes\": \"The card reports more than 100 websites but does not provide a precise website total. Grounding accuracy does not establish task completion.\", \"observations\": [\"web screenshots\", \"instructions\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/Hcompany/WebClick/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/WebClick\"}\n\nFour public test Parquet shards are present. The dataset card declares Apache-2.0. This is an offline localization test, not an end-to-end web-task success benchmark; reported screenshots and samples are not summed as different populations.\n\nPrimary source: https://huggingface.co/datasets/Hcompany/WebClick\nCatalog record: https://awesomecua.github.io/resources/url-5e2bd1b3398974b79c9c/"}
{"id": "url:60a95f34072aeb2771cc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/meituan/EvoCUA-32B-20260105", "citation_url": "https://awesomecua.github.io/resources/url-60a95f34072aeb2771cc/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/meituan/EvoCUA-32B-20260105/tree/main", "gated": false, "hub_id": "meituan/EvoCUA-32B-20260105", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "f42d9537a05f2562b57e5d508f95c3d29a9107d06574318c5715c7da5fb2a86d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "f42d9537a05f2562b57e5d508f95c3d29a9107d06574318c5715c7da5fb2a86d", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/meituan/EvoCUA-32B-20260105"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/meituan/EvoCUA-32B-20260105"}], "name": "EvoCUA 32B 20260105", "publication": null, "published_at": null, "record_sha256": "7ff92082986c3049219acfc037682b2b01a0e6dbb0c799c686b8217b202cbc17", "relations": [{"evidence_url": "https://huggingface.co/api/models/meituan/EvoCUA-32B-20260105", "target_id": "github:meituan/evocua", "type": "variant_of"}], "resource_id": "url:60a95f34072aeb2771cc", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/meituan/EvoCUA-32B-20260105", "topics": ["reinforcement-learning", "planning"], "updated_at": null}, "text": "EvoCUA 32B 20260105\nResource type: model\nReview status: curated\nScope: core\n\nLarger EvoCUA computer-use agent checkpoint (20260105 release) with about 33B BF16 parameters in the Qwen3-VL architecture, tagged for OSWorld and for English and Chinese.\nPlatforms: desktop\nTopics: reinforcement-learning, planning\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/meituan/EvoCUA-32B-20260105/tree/main\", \"gated\": false, \"hub_id\": \"meituan/EvoCUA-32B-20260105\", \"languages\": [\"en\", \"zh\"]}\n\nPrimary source: https://huggingface.co/meituan/EvoCUA-32B-20260105\nCatalog record: https://awesomecua.github.io/resources/url-60a95f34072aeb2771cc/"}
{"id": "url:611df37e5f3caa6befd7", "metadata": {"aliases": ["url:5f4cb613bda51b45d9c7", "inclusionAI/UI-Venus-2-9B", "https://huggingface.co/inclusionAI/UI-Venus-2-9B"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b", "citation_url": "https://awesomecua.github.io/resources/url-611df37e5f3caa6befd7/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b/tree/main", "gated": false, "hub_id": "inclusionAI/UI-Venus-2-9B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "0097be2c0e5c3f3c8aace10de835b0d7c1075531cc7823287c3f5488aebaea08", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:02Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:03Z", "role": "primary", "sha256": "3ec7929329e8e1b09f31dee4a3e67128d68cddc95220a5944f4cd6102dae327e", "supports": ["aliases"], "url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b"}], "name": "UI-Venus 2 9B", "publication": null, "published_at": null, "record_sha256": "1dd4209f3dad9c5e0d6679a71bc3a197c0555430897a149594da117a54610527", "relations": [{"evidence_url": "https://huggingface.co/api/models/inclusionAI/UI-Venus-2-9b", "target_id": "github:inclusionai/ui-venus", "type": "variant_of"}], "resource_id": "url:611df37e5f3caa6befd7", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/inclusionAI/UI-Venus-2-9b", "topics": ["grounding", "planning"], "updated_at": null}, "text": "UI-Venus 2 9B\nResource type: model\nReview status: curated\nScope: core\n\nUI-Venus 2 GUI-agent checkpoint fine-tuned from Qwen3.5-9B, with about 9.4B BF16 parameters in the Qwen3.5 multimodal architecture, loaded through Transformers for image-text-to-text.\nRecorded aliases: url:5f4cb613bda51b45d9c7, inclusionAI/UI-Venus-2-9B, https://huggingface.co/inclusionAI/UI-Venus-2-9B\nPlatforms: cross-platform\nTopics: grounding, planning\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/inclusionAI/UI-Venus-2-9b/tree/main\", \"gated\": false, \"hub_id\": \"inclusionAI/UI-Venus-2-9B\"}\n\nPrimary source: https://huggingface.co/inclusionAI/UI-Venus-2-9b\nCatalog record: https://awesomecua.github.io/resources/url-611df37e5f3caa6befd7/"}
{"id": "url:61a5321db10fc8a312b9", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/ritzzai/GUI-R1", "citation_url": "https://awesomecua.github.io/resources/url-61a5321db10fc8a312b9/", "details": {"disabled": false, "download_url": "https://huggingface.co/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "3eff86951de7c36dbd4af006ab3f1350213bdfea384494e59c8abc007793c979", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["environments"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.weights"], "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/ritzzai/GUI-R1"}], "name": "GUI-R1 checkpoint bundle", "publication": null, "published_at": null, "record_sha256": "4b9d659a634e312071d76dd7a87e66a34938caf9f531654a350388776aba3fb4", "relations": [{"evidence_url": "https://huggingface.co/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "resource_id": "url:61a5321db10fc8a312b9", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/ritzzai/GUI-R1", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null}, "text": "GUI-R1 checkpoint bundle\nResource type: model\nReview status: curated\nScope: core\n\nA single Hub release containing GUI-R1 3B and 7B action-model checkpoints trained with reinforcement learning across interface platforms.\nPlatforms: desktop, windows, linux, macos, android, web\nTopics: grounding, reinforcement-learning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/ritzzai/GUI-R1/tree/main\", \"gated\": false, \"hub_id\": \"ritzzai/GUI-R1\"}\n\nThe manifest contains two safetensors shards in GUI-R1-3B and four in GUI-R1-7B. Both are counted as one Hub resource rather than duplicate root URLs. The model card declares MIT; this does not automatically apply to source training data or upstream base models.\n\nPrimary source: https://huggingface.co/ritzzai/GUI-R1\nCatalog record: https://awesomecua.github.io/resources/url-61a5321db10fc8a312b9/"}
{"id": "url:6262c757080f8838e62c", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "canonical_source_url": "https://doi.org/10.1145/1753326.1753555", "citation_url": "https://awesomecua.github.io/resources/url-6262c757080f8838e62c/", "details": {"publication": {"authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "date_basis": "Supplied month precision: 2010-04. Date Issued: April 2010", "date_parts": [2010, 4], "doi": "10.1145/1753326.1753555", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Distinct visual-testing contribution from the 2009 Sikuli paper. The institutional abstract and published-version citation supply the retained evidence. Exact publication day and operating-system coverage are not established in the accessible institutional abstract.", "pages": "1535-1544", "pdf_url": null, "publisher": null, "series": null, "title": "GUI Testing Using Computer Vision", "type": "conference-paper", "url": "https://doi.org/10.1145/1753326.1753555", "venue": "Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI '10)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbace77c54d1b4ed5dcf825dcb0f4525c32e33ac885e70312e8bb4a731a94e80", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"}], "identifiers": {"doi": "10.1145/1753326.1753555"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1753326.1753555"}], "name": "GUI Testing Using Computer Vision", "publication": {"authors": ["Tsung-Hsiang Chang", "Tom Yeh", "Robert C. Miller"], "date_basis": "Supplied month precision: 2010-04. Date Issued: April 2010", "date_parts": [2010, 4], "doi": "10.1145/1753326.1753555", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/6a6097ba-1d97-43cf-bc70-04b5af02efd5"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Distinct visual-testing contribution from the 2009 Sikuli paper. The institutional abstract and published-version citation supply the retained evidence. Exact publication day and operating-system coverage are not established in the accessible institutional abstract.", "pages": "1535-1544", "pdf_url": null, "publisher": null, "series": null, "title": "GUI Testing Using Computer Vision", "type": "conference-paper", "url": "https://doi.org/10.1145/1753326.1753555", "venue": "Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI '10)", "volume": null}, "published_at": null, "record_sha256": "7a0e25df3c31170d528bbf5083c959cdd138b7ca53379919bd710c0e97f4302c", "relations": [], "resource_id": "url:6262c757080f8838e62c", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/1753326.1753555", "topics": ["evaluation", "grounding", "automation"], "updated_at": null}, "text": "GUI Testing Using Computer Vision\nResource type: paper\nReview status: curated\nScope: core\n\nPresents visual GUI test scripts that use images to specify actions and expected screen feedback. Recording both input events and screenshots can generate tests from demonstrations, supporting unit, regression, and test-driven workflows.\nPlatforms: unspecified\nTopics: evaluation, grounding, automation\nRecord publication date (recorded precision): 2010-04\nAuthors: Tsung-Hsiang Chang, Tom Yeh, Robert C. Miller\n\nPublication edition: GUI Testing Using Computer Vision\nPublication authors: Tsung-Hsiang Chang, Tom Yeh, Robert C. Miller\nPublication edition date (recorded precision): 2010-04\nPublication venue: Proceedings of the 28th International Conference on Human Factors in Computing Systems (CHI '10)\nPublication pages: 1535-1544\nPublication doi: 10.1145/1753326.1753555\nPublication date basis: Supplied month precision: 2010-04. Date Issued: April 2010\nPublication notes: Distinct visual-testing contribution from the 2009 Sikuli paper. The institutional abstract and published-version citation supply the retained evidence. Exact publication day and operating-system coverage are not established in the accessible institutional abstract.\nPublication source: https://doi.org/10.1145/1753326.1753555\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This is a separate visual-testing contribution from the 2009 Sikuli paper, not another version of the same article. MIT PDF download returned HTTP 429; its complete institutional abstract and citation metadata remain the retained review source.\n\nPrimary source: https://doi.org/10.1145/1753326.1753555\nCatalog record: https://awesomecua.github.io/resources/url-6262c757080f8838e62c/"}
{"id": "url:628b5251757bb3888e89", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["James Allen", "Nathanael Chambers", "George Ferguson", "Lucian Galescu", "Hyuckchul Jung", "Mary Swift", "William Taysom"], "canonical_source_url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "citation_url": "https://awesomecua.github.io/resources/url-628b5251757bb3888e89/", "details": {"publication": {"authors": ["James Allen", "Nathanael Chambers", "George Ferguson", "Lucian Galescu", "Hyuckchul Jung", "Mary Swift", "William Taysom"], "date_basis": "Supplied year precision: 2007. Papers section: AAAI-2007", "date_parts": [2007], "doi": null, "event_dates": null, "evidence_urls": ["https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "https://www.cs.rochester.edu/research/cisd/projects/plow/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes. DOI and volume number are not supplied in the retained primary sources. No exact publication month/day is established.", "pages": "1514-1519", "pdf_url": "https://cdn.aaai.org/AAAI/2007/AAAI07-240.pdf", "publisher": null, "series": null, "title": "PLOW: A Collaborative Task Learning Agent", "type": "conference-paper", "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "venue": "Twenty-Second AAAI National Conference on Artificial Intelligence", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:26Z", "role": "primary", "sha256": "795d275c9f85180561fd101652221d30fb98daca2c2a00faf2b57c397bdfd68d", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "c0b3571efd5f2e137e42b38107de29aca6668819bec84109912d514867d85b8c", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.rochester.edu/research/cisd/projects/plow/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:26Z", "role": "primary", "sha256": "795d275c9f85180561fd101652221d30fb98daca2c2a00faf2b57c397bdfd68d", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:24Z", "role": "primary", "sha256": "c0b3571efd5f2e137e42b38107de29aca6668819bec84109912d514867d85b8c", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "environments", "kind"], "url": "https://www.cs.rochester.edu/research/cisd/projects/plow/"}], "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf"}, {"rel": "publication_pdf", "url": "https://cdn.aaai.org/AAAI/2007/AAAI07-240.pdf"}], "name": "PLOW: A Collaborative Task Learning Agent", "publication": {"authors": ["James Allen", "Nathanael Chambers", "George Ferguson", "Lucian Galescu", "Hyuckchul Jung", "Mary Swift", "William Taysom"], "date_basis": "Supplied year precision: 2007. Papers section: AAAI-2007", "date_parts": [2007], "doi": null, "event_dates": null, "evidence_urls": ["https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "https://www.cs.rochester.edu/research/cisd/projects/plow/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes. DOI and volume number are not supplied in the retained primary sources. No exact publication month/day is established.", "pages": "1514-1519", "pdf_url": "https://cdn.aaai.org/AAAI/2007/AAAI07-240.pdf", "publisher": null, "series": null, "title": "PLOW: A Collaborative Task Learning Agent", "type": "conference-paper", "url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "venue": "Twenty-Second AAAI National Conference on Artificial Intelligence", "volume": null}, "published_at": null, "record_sha256": "027cbc2eaa97616fd73c40b8f56879880b6e27abb2234d06e1630730ec0f30be", "relations": [], "resource_id": "url:628b5251757bb3888e89", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf", "topics": ["web-navigation", "human-in-the-loop", "planning"], "updated_at": null}, "text": "PLOW: A Collaborative Task Learning Agent\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes PLOW, which learns executable web-task models from a session combining demonstrations, natural-language explanations, and dialogue. Browser instrumentation connects user actions with DOM structure to support task generalization.\nPlatforms: web\nTopics: web-navigation, human-in-the-loop, planning\nRecord publication date (recorded precision): 2007\nAuthors: James Allen, Nathanael Chambers, George Ferguson, Lucian Galescu, Hyuckchul Jung, Mary Swift, William Taysom\n\nPublication edition: PLOW: A Collaborative Task Learning Agent\nPublication authors: James Allen, Nathanael Chambers, George Ferguson, Lucian Galescu, Hyuckchul Jung, Mary Swift, William Taysom\nPublication edition date (recorded precision): 2007\nPublication venue: Twenty-Second AAAI National Conference on Artificial Intelligence\nPublication pages: 1514-1519\nPublication date basis: Supplied year precision: 2007. Papers section: AAAI-2007\nPublication notes: The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes. DOI and volume number are not supplied in the retained primary sources. No exact publication month/day is established.\nPublication source: https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The ordered names follow the original published paper, without expanding initials or replacing names from secondary indexes.\n\nPrimary source: https://aaai.org/Papers/AAAI/2007/AAAI07-240.pdf\nCatalog record: https://awesomecua.github.io/resources/url-628b5251757bb3888e89/"}
{"id": "url:631e74223e198ef40202", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data", "citation_url": "https://awesomecua.github.io/resources/url-631e74223e198ef40202/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md", "population": "Linux, macOS, Windows, Android, iOS, and Web", "qualifier": "exact", "unit": "platform categories", "value": 6}], "evidence_urls": ["https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md"], "formats": ["JSONL annotations", "JSON manifest", "split image archives"], "modalities": ["screenshots", "text", "actions"], "notes": "The six categories include Web, so they are not six operating systems. Grounding and trajectory annotations follow different collection processes. No sample total or train/test division is inferred from archive-part counts or Hub size tags.", "splits": [], "supervision": ["GUI grounding", "GUI action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/tree/main", "gated": false, "hub_id": "OpenGVLab/ScaleCUA-Data", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "dca4b89176c7653e83545363e4868bd2d5f0bf618df007b68a7e25407262f1ac", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/OpenGVLab/ScaleCUA-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:37Z", "role": "primary", "sha256": "8c545d66cc7199a9258bd91ed48930691d0d99555bc30ee457fae9bf656fd18d", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "dca4b89176c7653e83545363e4868bd2d5f0bf618df007b68a7e25407262f1ac", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OpenGVLab/ScaleCUA-Data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data"}], "name": "ScaleCUA Data", "publication": null, "published_at": null, "record_sha256": "dcb3675c3e140b91acd366164d2a923b7d62f233dbcd200df0bd80b7de0f3762", "relations": [{"evidence_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "resource_id": "url:631e74223e198ef40202", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data", "topics": ["data-generation"], "updated_at": null}, "text": "ScaleCUA Data\nResource type: dataset\nReview status: curated\nScope: core\n\nCross-platform training data released with ScaleCUA, with grounding and navigation annotations for Ubuntu, Windows, macOS, Android, iPhone and web screenshots.\nPlatforms: cross-platform\nTopics: data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md\", \"population\": \"Linux, macOS, Windows, Android, iOS, and Web\", \"qualifier\": \"exact\", \"unit\": \"platform categories\", \"value\": 6}], \"evidence_urls\": [\"https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/raw/main/README.md\"], \"formats\": [\"JSONL annotations\", \"JSON manifest\", \"split image archives\"], \"modalities\": [\"screenshots\", \"text\", \"actions\"], \"notes\": \"The six categories include Web, so they are not six operating systems. Grounding and trajectory annotations follow different collection processes. No sample total or train/test division is inferred from archive-part counts or Hub size tags.\", \"splits\": [], \"supervision\": [\"GUI grounding\", \"GUI action trajectories\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data/tree/main\", \"gated\": false, \"hub_id\": \"OpenGVLab/ScaleCUA-Data\"}\n\nPrimary source: https://huggingface.co/datasets/OpenGVLab/ScaleCUA-Data\nCatalog record: https://awesomecua.github.io/resources/url-631e74223e198ef40202/"}
{"id": "url:63ef8c7e455951a20281", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://scribehow.com/", "citation_url": "https://awesomecua.github.io/resources/url-63ef8c7e455951a20281/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "bb658d64714578f8ce5d78b014474781f865e42ccc078be3b61539bc3780b235", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://scribehow.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://scribehow.com/"}], "name": "Scribe", "publication": null, "published_at": null, "record_sha256": "7cd07ba0aedb68687d49f019120dfd227eb60785f42bf29472ccdcb8ab7df954", "relations": [], "resource_id": "url:63ef8c7e455951a20281", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://scribehow.com/", "topics": ["recording", "workflow"], "updated_at": null}, "text": "Scribe\nResource type: product\nReview status: curated\nScope: adjacent\n\nWorkflow recording and documentation product adjacent to agent data collection.\nPlatforms: desktop, web\nTopics: recording, workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://scribehow.com/\nCatalog record: https://awesomecua.github.io/resources/url-63ef8c7e455951a20281/"}
{"id": "url:66096ad4057a3d3726f5", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data", "citation_url": "https://awesomecua.github.io/resources/url-66096ad4057a3d3726f5/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data", "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart image archives"], "modalities": ["GUI screenshots", "bounding boxes", "grounding annotations"], "notes": "Public Hub revision 5a0b35fb includes six bbox JSON files and associated image archives. The training configuration names six input paths with sampling_strategy=all; those input paths are not six disjoint datasets or a verified example count. The card’s source-list wording and mismatched hyperlinks remain unresolved, so its approximately 1M screenshots/10M elements headline is not adopted as a release inventory. No data license is established by the card, and the model repository license is not transferred to these data.", "splits": ["training"], "supervision": ["GUI grounding with bounding-box supervision"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/tree/main", "gated": false, "hub_id": "cckevinn/GUI-Actor-Data"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "e42bbcf8d466c977f9b2150feb73d8cbda2ff907c279630ad4730a8b1f5d665c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "d6271cff879b1a2a70490cf3dd95765bef44ecfca531d8ef3545f101673b4423", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dafdebf2fa1e0a5db871addd103e50cb55625f0108c757708029e885920ee4e", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:12Z", "role": "primary", "sha256": "3521c849b29d7fc88430a8e73a8b2aeb2e613d7fa16ca054dd5bf4ec011b36ca", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data"}], "name": "GUI-Actor data", "publication": null, "published_at": null, "record_sha256": "978d632046475da1b0a7a77f92f202933923ddabf67317d1bdf686b378355113", "relations": [{"evidence_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md", "target_id": "arxiv:2506.03143", "type": "documented_by"}], "resource_id": "url:66096ad4057a3d3726f5", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/cckevinn/GUI-Actor-Data", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "GUI-Actor data\nResource type: dataset\nReview status: curated\nScope: core\n\nA processed screenshot and bounding-box collection released for training GUI-Actor visual grounding models.\nPlatforms: unspecified\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/raw/main/README.md\", \"https://huggingface.co/api/datasets/cckevinn/GUI-Actor-Data\", \"https://raw.githubusercontent.com/microsoft/GUI-Actor/HEAD/data/data_config.yaml\"], \"formats\": [\"JSON annotations\", \"ZIP image packages\", \"multipart image archives\"], \"modalities\": [\"GUI screenshots\", \"bounding boxes\", \"grounding annotations\"], \"notes\": \"Public Hub revision 5a0b35fb includes six bbox JSON files and associated image archives. The training configuration names six input paths with sampling_strategy=all; those input paths are not six disjoint datasets or a verified example count. The card’s source-list wording and mismatched hyperlinks remain unresolved, so its approximately 1M screenshots/10M elements headline is not adopted as a release inventory. No data license is established by the card, and the model repository license is not transferred to these data.\", \"splits\": [\"training\"], \"supervision\": [\"GUI grounding with bounding-box supervision\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/cckevinn/GUI-Actor-Data/tree/main\", \"gated\": false, \"hub_id\": \"cckevinn/GUI-Actor-Data\"}\n\nPublic JSON annotations and image archives are present. The card has no data-license declaration. Its source list contains mismatched hyperlinks and says six datasets while naming fewer distinct entries; no source-derivation relationships or aggregate count are imported from that list.\n\nPrimary source: https://huggingface.co/datasets/cckevinn/GUI-Actor-Data\nCatalog record: https://awesomecua.github.io/resources/url-66096ad4057a3d3726f5/"}
{"id": "url:69d6afe38a9a128ba9f7", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.askui.com/", "citation_url": "https://awesomecua.github.io/resources/url-69d6afe38a9a128ba9f7/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "df8615d0687cba27e5558ea3f7ecb1823871e3804522603dd62b52dcf127ac89", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.askui.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.askui.com/"}], "name": "AskUI", "publication": null, "published_at": null, "record_sha256": "9364fb661841291df1fa44aeb460b530bd7af91847b65b9cb1f4401b762b1270", "relations": [{"evidence_url": "https://www.askui.com/", "target_id": "organization:askui", "type": "published_by"}], "resource_id": "url:69d6afe38a9a128ba9f7", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.askui.com/", "topics": ["workflow"], "updated_at": null}, "text": "AskUI\nResource type: product\nReview status: curated\nScope: core\n\nAgentic test-automation platform running plain-language test cases from pixels rather than selectors on desktop apps, Citrix, in-car head units and mobile devices, deployable on-premise or air-gapped.\nPlatforms: desktop, web\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.askui.com/\nCatalog record: https://awesomecua.github.io/resources/url-69d6afe38a9a128ba9f7/"}
{"id": "url:6a212bc12f82c824f6b9", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/wanlilll/WeaveBench", "citation_url": "https://awesomecua.github.io/resources/url-6a212bc12f82c824f6b9/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/wanlilll/WeaveBench/tree/main", "gated": false, "hub_id": "wanlilll/WeaveBench", "languages": ["en"], "modalities": [], "size_category_reported": ["n<1K"]}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e65e5793fe5b251ef7793a0618f4e8139fe5c92c2cf8093d0cc6142e8ece9d02", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/wanlilll/WeaveBench"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e65e5793fe5b251ef7793a0618f4e8139fe5c92c2cf8093d0cc6142e8ece9d02", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/wanlilll/WeaveBench"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:27Z", "url": "https://huggingface.co/datasets/wanlilll/WeaveBench"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/wanlilll/WeaveBench"}], "name": "WeaveBench", "publication": null, "published_at": null, "record_sha256": "240a9e270d4de9deecb43d2a4faba4729eab1b0653ddaa7da499d30026b2455f", "relations": [], "resource_id": "url:6a212bc12f82c824f6b9", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/wanlilll/WeaveBench", "topics": ["evaluation", "tool-integration"], "updated_at": null}, "text": "WeaveBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nComputer-use benchmark mixing graphical interaction and code execution.\nPlatforms: desktop, linux\nTopics: evaluation, tool-integration\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/wanlilll/WeaveBench/tree/main\", \"gated\": false, \"hub_id\": \"wanlilll/WeaveBench\", \"languages\": [\"en\"], \"size_category_reported\": [\"n<1K\"]}\n\nPrimary source: https://huggingface.co/datasets/wanlilll/WeaveBench\nCatalog record: https://awesomecua.github.io/resources/url-6a212bc12f82c824f6b9/"}
{"id": "url:6aa1aa58a7a888395189", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xlangai/CUA-Gym", "citation_url": "https://awesomecua.github.io/resources/url-6aa1aa58a7a888395189/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/CUA-Gym/tree/main", "gated": false, "hub_id": "xlangai/CUA-Gym", "languages": ["en"], "modalities": ["tabular", "text"], "size_category_reported": ["10K<n<100K"]}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "43b8ebde3b971e1b8b61d789c786ead59b45a2490bc6ed19e204bafc03a31f44", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/xlangai/CUA-Gym"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:25Z", "role": "primary", "sha256": "962cd9162bea83f7adfedb3da50129bcf4203ae65176bdf53f0af6788b9603ef", "supports": ["relations"], "url": "https://huggingface.co/datasets/xlangai/CUA-Gym/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "43b8ebde3b971e1b8b61d789c786ead59b45a2490bc6ed19e204bafc03a31f44", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/CUA-Gym"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/xlangai/CUA-Gym"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/CUA-Gym"}], "name": "CUA-Gym", "publication": null, "published_at": null, "record_sha256": "0253712607421c37afb1c830a3ee6d9e2d5bc4b0eccd077aabd374a4b340c0ae", "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/CUA-Gym/raw/main/README.md", "target_id": "arxiv:2605.25624", "type": "documented_by"}], "resource_id": "url:6aa1aa58a7a888395189", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xlangai/CUA-Gym", "topics": ["evaluation", "data-generation"], "updated_at": null}, "text": "CUA-Gym\nResource type: dataset\nReview status: curated\nScope: core\n\nComputer-use environment and task resources for agent research.\nPlatforms: desktop\nTopics: evaluation, data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xlangai/CUA-Gym/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/CUA-Gym\", \"languages\": [\"en\"], \"modalities\": [\"tabular\", \"text\"], \"size_category_reported\": [\"10K<n<100K\"]}\n\nPrimary source: https://huggingface.co/datasets/xlangai/CUA-Gym\nCatalog record: https://awesomecua.github.io/resources/url-6aa1aa58a7a888395189/"}
{"id": "url:6cd150ec23361b4d2e84", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/yiye2023/GUIAct", "citation_url": "https://awesomecua.github.io/resources/url-6cd150ec23361b4d2e84/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 67000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-step portion", "qualifier": "approximately", "unit": "action instructions", "value": 15000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "instructions", "actions"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": ["train", "test"], "supervision": ["single-step actions", "multi-step navigation"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIAct/tree/main", "gated": false, "hub_id": "yiye2023/GUIAct", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["web", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "624ccdd6e6c5b29b588a4289d1d94c1a1c0803868bd586175f2c9ebc3bfdd0b6", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIAct"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "98b45ea81164d1e1a1dd82255207053b15cd6c69d922a1c5cf3387ce604d4b74", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIAct"}], "name": "GUIAct", "publication": null, "published_at": null, "record_sha256": "38ae9d7d12c6aad34682aa0ecd2fb15099ddf54a3ce6435a5aedc3798963f23f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "resource_id": "url:6cd150ec23361b4d2e84", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/yiye2023/GUIAct", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "GUIAct\nResource type: dataset\nReview status: curated\nScope: core\n\nWeb and Android GUI action-instruction data, with single-step, multi-step and smartphone partitions.\nPlatforms: web, android\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"population\": \"single-step portion\", \"qualifier\": \"approximately\", \"unit\": \"action instructions\", \"value\": 67000}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"population\": \"multi-step portion\", \"qualifier\": \"approximately\", \"unit\": \"action instructions\", \"value\": 15000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\"], \"formats\": [\"JSON\", \"Parquet\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\"], \"notes\": \"Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.\", \"splits\": [\"train\", \"test\"], \"supervision\": [\"single-step actions\", \"multi-step navigation\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/yiye2023/GUIAct/tree/main\", \"gated\": false, \"hub_id\": \"yiye2023/GUIAct\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"https://huggingface.co/datasets/yiye2023/GUIAct/raw/main/README.md\"], \"reason\": \"Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict.\"}}}\n\nPublic JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.\n\nPrimary source: https://huggingface.co/datasets/yiye2023/GUIAct\nCatalog record: https://awesomecua.github.io/resources/url-6cd150ec23361b4d2e84/"}
{"id": "url:6d959b58bd4a2e98e417", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B", "citation_url": "https://awesomecua.github.io/resources/url-6d959b58bd4a2e98e417/", "details": {"disabled": false, "download_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-Agent-7B"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "3d4e2deda3e8ecc2e6bbba732c01a84dee2443bb7132ef1aeae444f001d23d69", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "20e47378595ed007766a135c37fe6709d8853a780ecd5ef56792f0b88c05aeff", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/HanXiao1999/UI-Genie-Agent-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "3d4e2deda3e8ecc2e6bbba732c01a84dee2443bb7132ef1aeae444f001d23d69", "supports": ["licenses.weights"], "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B"}], "name": "UI-Genie Agent 7B", "publication": null, "published_at": null, "record_sha256": "186f6aaf10073d7239c17a3dcaa47e2bf441727da787baad322c0bdff1a06456", "relations": [{"evidence_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "resource_id": "url:6d959b58bd4a2e98e417", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B", "topics": ["mobile-navigation", "imitation-learning", "action-prediction"], "updated_at": null}, "text": "UI-Genie Agent 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL mobile action model trained using existing and reward-guided synthetic trajectories from the UI-Genie framework.\nPlatforms: android\nTopics: mobile-navigation, imitation-learning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B/tree/main\", \"gated\": false, \"hub_id\": \"HanXiao1999/UI-Genie-Agent-7B\"}\n\nFour public safetensors shards are present and the card declares MIT. Its training-data metadata/link uses UI-Genie-Agent-5k while prose labels it 16k; no precise training-set relationship or count is asserted across those inconsistent references. Larger announced models are not counted as released.\n\nPrimary source: https://huggingface.co/HanXiao1999/UI-Genie-Agent-7B\nCatalog record: https://awesomecua.github.io/resources/url-6d959b58bd4a2e98e417/"}
{"id": "url:72f7fe0211c6551b945f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B", "citation_url": "https://awesomecua.github.io/resources/url-72f7fe0211c6551b945f/", "details": {"disabled": false, "download_url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebShepherd_3B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "6f28e86a99b286a575c5e7f8d24d6901655971e9a091493950b59948f4b9ad9b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/LangAGI-Lab/WebShepherd_3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "647d9e7bdad5fe2c03a8cfcd34d93f3b2bf7c1553b517a4a843bc0113fa47456", "supports": ["identity", "summary", "relations"], "url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:49:45Z", "role": "primary", "sha256": "252ed290ec84ffd8fa686fb72df68afe7bbd8c19b80538c0b318db60cd107d09", "supports": ["details"], "url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B/raw/main/config.json"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "6f28e86a99b286a575c5e7f8d24d6901655971e9a091493950b59948f4b9ad9b", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/LangAGI-Lab/WebShepherd_3B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:45:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B"}], "name": "WebShepherd 3B", "publication": null, "published_at": null, "record_sha256": "bfab07eca2539ccfddb1c74e4ea32bc6f0f28fae4e9b85ebd20553e68a75de52", "relations": [{"evidence_url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "resource_id": "url:72f7fe0211c6551b945f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/LangAGI-Lab/WebShepherd_3B", "topics": ["reward-models", "web-navigation"], "updated_at": null}, "text": "WebShepherd 3B\nResource type: model\nReview status: curated\nScope: core\n\nA public text-based process-reward checkpoint from the Web-Shepherd collection for scoring web-agent decisions against task checklists.\nPlatforms: web\nTopics: reward-models, web-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/LangAGI-Lab/WebShepherd_3B/tree/main\", \"gated\": false, \"hub_id\": \"LangAGI-Lab/WebShepherd_3B\"}\n\nThe official project-linked collection identifies this public 3B checkpoint. The README uses a separate unavailable WebShepherd/web-shepherd-base URL; redirect equivalence is not asserted. Two safetensors shards and a Qwen2 causal-language-model configuration are present. No model card or weight license is published in the inspected manifest. The separately announced multimodal 3B model is not assumed released.\n\nPrimary source: https://huggingface.co/LangAGI-Lab/WebShepherd_3B\nCatalog record: https://awesomecua.github.io/resources/url-72f7fe0211c6551b945f/"}
{"id": "url:7405463ba56f9a41e79a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B", "citation_url": "https://awesomecua.github.io/resources/url-7405463ba56f9a41e79a/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-0.8B", "gated": false, "hub_revision": "c36a096cb0cb3320099eb8fde82608cadf2248ee", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "6bd2608de2715e3d76a0c9e9b8fd7540372bfd2f9864db0445d8e548c9afb518", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-0.8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "8130a68ede5391bd158ce2cc49c36c5f8c8ebf22b897db6f68e27babbbc66965", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B"}], "name": "Seg2Click Qwen3.5 0.8B", "publication": null, "published_at": null, "record_sha256": "ca85a8d502f6dc0c712513c8bc0bbf8b7241e2457fcf055c8372e60293166e67", "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "url:7405463ba56f9a41e79a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B", "topics": ["grounding"], "updated_at": null}, "text": "Seg2Click Qwen3.5 0.8B\nResource type: model\nReview status: curated\nScope: core\n\nCompact Seg2Click checkpoint with a merged Qwen3.5 backbone and an action-mask decoder.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.5-0.8B\", \"gated\": false, \"hub_revision\": \"c36a096cb0cb3320099eb8fde82608cadf2248ee\", \"listed_weight_files\": [\"model.safetensors\", \"phase1_mask_decoder.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.\n\nPrimary source: https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-0.8B\nCatalog record: https://awesomecua.github.io/resources/url-7405463ba56f9a41e79a/"}
{"id": "url:74eeb7c9fb20674fe6e2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://openai.com/index/computer-using-agent/", "citation_url": "https://awesomecua.github.io/resources/url-74eeb7c9fb20674fe6e2/", "details": {}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "14f573af0c4ea3e5922b8831b7f03724936963e293d7c0aaa3fccbb71dbb0cfa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://openai.com/index/computer-using-agent/"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/computer-using-agent/"}], "name": "OpenAI CUA", "publication": null, "published_at": null, "record_sha256": "feeb9a3de318b09d5f94f1abb75c19d1cebfac259f4837b35231c3dcedcc9108", "relations": [{"evidence_url": "https://openai.com/index/computer-using-agent/", "target_id": "organization:openai", "type": "published_by"}], "resource_id": "url:74eeb7c9fb20674fe6e2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openai.com/index/computer-using-agent/", "topics": ["planning", "grounding"], "updated_at": null}, "text": "OpenAI CUA\nResource type: model\nReview status: curated\nScope: core\n\nComputer-using model introduced with the Operator research preview.\nPlatforms: web, desktop\nTopics: planning, grounding\n\nPrimary source: https://openai.com/index/computer-using-agent/\nCatalog record: https://awesomecua.github.io/resources/url-74eeb7c9fb20674fe6e2/"}
{"id": "url:75917f2c2b2802aafd3e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env", "citation_url": "https://awesomecua.github.io/resources/url-75917f2c2b2802aafd3e/", "details": {"artifact_type": "Ubuntu VMware image", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/tree/main", "gated": false, "hub_id": "OS-Copilot/ScienceBoard-Env", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md", "https://huggingface.co/api/models/OS-Copilot/ScienceBoard-Env"], "reason": "This resource is a prebuilt Ubuntu VMware environment archive (VM.zip), not neural model weights. MIT Hub metadata in a model namespace cannot establish a weights license or relicense the guest OS and bundled applications."}}}, "environments": ["linux", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "8ee436f67e708bb917992764bc60a017c9db2edb21bdb99356b35c69f73724d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.weights"], "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "9ffaf62b54a610afe186b3576b55fdd8287289e0cdfb88371bd9849a616f8460", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/ScienceBoard-Env"}], "identifiers": {}, "kind": "infrastructure", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env"}], "name": "ScienceBoard VMware environment", "publication": null, "published_at": null, "record_sha256": "749f4d26ce4a847448b29d5af01c3c451dbfeea99e9e8261a9e8c698e8025a18", "relations": [{"evidence_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md", "target_id": "arxiv:2505.19897", "type": "documented_by"}], "resource_id": "url:75917f2c2b2802aafd3e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OS-Copilot/ScienceBoard-Env", "topics": ["sandbox", "science", "evaluation"], "updated_at": null}, "text": "ScienceBoard VMware environment\nResource type: infrastructure\nReview status: curated\nScope: core\n\nA prebuilt Ubuntu virtual-machine archive containing the software environment used for ScienceBoard scientific computer-use tasks.\nPlatforms: linux, desktop\nTopics: sandbox, science, evaluation\n\nRecorded specifications: {\"artifact_type\": \"Ubuntu VMware image\", \"disabled\": false, \"download_url\": \"https://huggingface.co/OS-Copilot/ScienceBoard-Env/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/ScienceBoard-Env\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/OS-Copilot/ScienceBoard-Env/raw/main/README.md\", \"https://huggingface.co/api/models/OS-Copilot/ScienceBoard-Env\"], \"reason\": \"This resource is a prebuilt Ubuntu VMware environment archive (VM.zip), not neural model weights. MIT Hub metadata in a model namespace cannot establish a weights license or relicense the guest OS and bundled applications.\"}}}\n\nThe public Hub manifest contains VM.zip. This is an environment image, not neural model weights despite its Hub namespace. The card declares MIT metadata for the hosted release, but that does not relicense Ubuntu or included applications. The archive was not downloaded or booted; VMware’s own terms are separate.\n\nPrimary source: https://huggingface.co/OS-Copilot/ScienceBoard-Env\nCatalog record: https://awesomecua.github.io/resources/url-75917f2c2b2802aafd3e/"}
{"id": "url:772161f06c2b3d4978b5", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/osunlp/WebJudge-7B", "citation_url": "https://awesomecua.github.io/resources/url-772161f06c2b3d4978b5/", "details": {"disabled": false, "download_url": "https://huggingface.co/osunlp/WebJudge-7B/tree/main", "gated": false, "hub_id": "osunlp/WebJudge-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "425f11651cdae09997c1b038cdc9c091778c35a51b8de88140fc6dc935b8bb64", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "2eebc1f909263be946057306b41d733e29f0f32337dae02da4e989b5477cf31f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/osunlp/WebJudge-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "425f11651cdae09997c1b038cdc9c091778c35a51b8de88140fc6dc935b8bb64", "supports": ["licenses.weights"], "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/WebJudge-7B"}], "name": "WebJudge 7B", "publication": null, "published_at": null, "record_sha256": "e6c549642c4c8f6357500f45583b6de457eb5cabf4eab8f14a093a22338d10c0", "relations": [{"evidence_url": "https://huggingface.co/osunlp/WebJudge-7B/raw/main/README.md", "target_id": "arxiv:2504.01382", "type": "documented_by"}], "resource_id": "url:772161f06c2b3d4978b5", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/osunlp/WebJudge-7B", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null}, "text": "WebJudge 7B\nResource type: model\nReview status: curated\nScope: core\n\nA visual trajectory judge that retains selected intermediate screenshots to evaluate whether a web agent completed its task.\nPlatforms: web\nTopics: evaluation, reward-models, web-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/osunlp/WebJudge-7B/tree/main\", \"gated\": false, \"hub_id\": \"osunlp/WebJudge-7B\"}\n\nFour public safetensors shards are present. Apache-2.0 is declared for the weights. This is an evaluator checkpoint rather than a navigation policy; human-agreement rates are not treated as task-success scores.\n\nPrimary source: https://huggingface.co/osunlp/WebJudge-7B\nCatalog record: https://awesomecua.github.io/resources/url-772161f06c2b3d4978b5/"}
{"id": "url:79287058342289aa5d45", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k", "citation_url": "https://awesomecua.github.io/resources/url-79287058342289aa5d45/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "population": "default/train in the dataset card metadata", "qualifier": "exact", "unit": "training rows", "value": 701864}], "evidence_urls": ["https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"], "formats": [], "modalities": ["screenshots", "instructions", "functionality labels", "bounding boxes"], "notes": "Rows are not unique screenshots. The card documents image-downscaling conventions that affect coordinates; preserve the device and original image-size fields.", "splits": ["train"], "supervision": ["element grounding", "functionality captioning"]}, "download_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/tree/main", "modalities": ["image", "text", "bounding-box"], "sample_count": 701864, "sample_unit": "training rows reported in the dataset card"}, "environments": ["web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.sample_count", "details.sample_unit", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "c73b46acc2ff034a443fb8dbde327de1888535267199c9fb1819adf66898f1d7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k"}], "name": "AutoGUI v1 training data", "publication": null, "published_at": null, "record_sha256": "72a004c7bbca74a9530002753679be9e4c676ba7d377ba3743227ba60d50dc3a", "relations": [{"evidence_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md", "target_id": "arxiv:2502.01977", "type": "documented_by"}], "resource_id": "url:79287058342289aa5d45", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "AutoGUI v1 training data\nResource type: dataset\nReview status: curated\nScope: core\n\nScreenshot-based grounding and captioning samples with UI functionality descriptions, target coordinates, and bounding boxes.\nPlatforms: web, cross-platform\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md\", \"population\": \"default/train in the dataset card metadata\", \"qualifier\": \"exact\", \"unit\": \"training rows\", \"value\": 701864}], \"evidence_urls\": [\"https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/raw/main/README.md\"], \"formats\": [], \"modalities\": [\"screenshots\", \"instructions\", \"functionality labels\", \"bounding boxes\"], \"notes\": \"Rows are not unique screenshots. The card documents image-downscaling conventions that affect coordinates; preserve the device and original image-size fields.\", \"splits\": [\"train\"], \"supervision\": [\"element grounding\", \"functionality captioning\"]}, \"download_url\": \"https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k/tree/main\", \"modalities\": [\"image\", \"text\", \"bounding-box\"], \"sample_count\": 701864, \"sample_unit\": \"training rows reported in the dataset card\"}\n\nThe row count is not a unique-screen count. The data card documents image scaling conventions that affect coordinate interpretation.\n\nPrimary source: https://huggingface.co/datasets/AutoGUI/AutoGUI-v1-702k\nCatalog record: https://awesomecua.github.io/resources/url-79287058342289aa5d45/"}
{"id": "url:7abb418096432f56c685", "metadata": {"aliases": ["url:f4bcd696942ee04dcc42", "bevaya/ScreenSpot", "https://huggingface.co/datasets/bevaya/ScreenSpot"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot", "citation_url": "https://awesomecua.github.io/resources/url-7abb418096432f56c685/", "details": {"benchmark_protocol": {"actions": ["predicted click point; a predicted box is reduced to its center"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "population": "Hub test split at 0be08781; not unique screenshots", "qualifier": "exact", "unit": "instruction–target examples", "value": 1272}], "evaluation": "The SeeClick evaluator scores whether the predicted point falls inside the target box, with platform and text/icon breakdowns. The Hub packaging uses xyxy boxes, whereas the original evaluator reads xywh and converts it; adapt the loader to the chosen release.", "evidence_urls": ["https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md", "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot", "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"], "focus": ["single-step GUI element grounding"], "notes": "Offline, zero-shot evaluation across mobile, desktop, and web. The card explicitly excludes training use in its intended-use description. Public API reports ungated Parquet files and canonical Hub id bevaya/ScreenSpot; the catalog identity is preserved. Apache-2.0 is the Hub data declaration. No runtime or underlying screenshot rights audit was performed.", "observations": ["screenshot", "English instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot/tree/main", "gated": false, "hub_id": "bevaya/ScreenSpot", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:00Z", "role": "primary", "sha256": "5ab2240c064502f75a6726311e10e7724ab7055347697f883b15a6ee8e7d00a6", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "eb72113eefd4868114f8a8105eba2d3385da9034a5b3a4860661e03ae7dcae15", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "9e2d2a2c076032225911ae1032b548eafa6331ea81a2779632610f0f1c82dfd9", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/bevaya/ScreenSpot"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:56Z", "role": "primary", "sha256": "af0c4babeff84158ff196b54bf47242923bbf502d2805a1dd4eb52dba44be89c", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/rootsautomation/ScreenSpot"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot"}], "name": "ScreenSpot", "publication": null, "published_at": null, "record_sha256": "274a4e9078df315e603f7d75eb4141bced8b5fc5fa2beb1214aca63199dd772e", "relations": [], "resource_id": "url:7abb418096432f56c685", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/rootsautomation/ScreenSpot", "topics": ["grounding", "evaluation"], "updated_at": null}, "text": "ScreenSpot\nResource type: benchmark\nReview status: curated\nScope: core\n\nGUI grounding test set of 1,272 screenshot instructions from iOS, Android, macOS, Windows and web interfaces, each labeled with the target element's bounding box.\nRecorded aliases: url:f4bcd696942ee04dcc42, bevaya/ScreenSpot, https://huggingface.co/datasets/bevaya/ScreenSpot\nPlatforms: cross-platform\nTopics: grounding, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"predicted click point; a predicted box is reduced to its center\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md\", \"population\": \"Hub test split at 0be08781; not unique screenshots\", \"qualifier\": \"exact\", \"unit\": \"instruction–target examples\", \"value\": 1272}], \"evaluation\": \"The SeeClick evaluator scores whether the predicted point falls inside the target box, with platform and text/icon breakdowns. The Hub packaging uses xyxy boxes, whereas the original evaluator reads xywh and converts it; adapt the loader to the chosen release.\", \"evidence_urls\": [\"https://huggingface.co/datasets/bevaya/ScreenSpot/raw/0be08781e2e188582f6131625ae1598d443b4d5d/README.md\", \"https://huggingface.co/api/datasets/rootsautomation/ScreenSpot\", \"https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/pretrain/screenspot_test.py\"], \"focus\": [\"single-step GUI element grounding\"], \"notes\": \"Offline, zero-shot evaluation across mobile, desktop, and web. The card explicitly excludes training use in its intended-use description. Public API reports ungated Parquet files and canonical Hub id bevaya/ScreenSpot; the catalog identity is preserved. Apache-2.0 is the Hub data declaration. No runtime or underlying screenshot rights audit was performed.\", \"observations\": [\"screenshot\", \"English instruction\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/rootsautomation/ScreenSpot/tree/main\", \"gated\": false, \"hub_id\": \"bevaya/ScreenSpot\", \"languages\": [\"en\"], \"modalities\": [\"image\", \"text\"]}\n\nPrimary source: https://huggingface.co/datasets/rootsautomation/ScreenSpot\nCatalog record: https://awesomecua.github.io/resources/url-7abb418096432f56c685/"}
{"id": "url:7ae2670859e3796e8607", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browseract.com/", "citation_url": "https://awesomecua.github.io/resources/url-7ae2670859e3796e8607/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "463db02b607e7215a4421440c4d274ee3546bd9eefc50ed9e10f64d27d7f8f38", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browseract.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browseract.com/"}], "name": "BrowserAct", "publication": null, "published_at": null, "record_sha256": "417f687472aea7e436f9ae70ee3b97edc2df517b905a48560dfa9f48d995beb4", "relations": [{"evidence_url": "https://www.browseract.com/", "target_id": "organization:browseract", "type": "published_by"}], "resource_id": "url:7ae2670859e3796e8607", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browseract.com/", "topics": ["workflow"], "updated_at": null}, "text": "BrowserAct\nResource type: product\nReview status: curated\nScope: core\n\nManaged web scraping platform with reusable browser interaction workflows.\nPlatforms: web\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.browseract.com/\nCatalog record: https://awesomecua.github.io/resources/url-7ae2670859e3796e8607/"}
{"id": "url:7c3e8c4ba52428584079", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web", "citation_url": "https://awesomecua.github.io/resources/url-7c3e8c4ba52428584079/", "details": {"architecture": "LlamaForCausalLM", "availability": "Ungated complete weight/config/tokenizer manifest.", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt", "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md"], "reason": "Weight license text and card declaration are retained separately from framework code licensing."}}, "parameter_label": "8B", "release_revision": "09856c28ec324cf3fe4c513bdc4437a75b78185f", "training_data_note": "Publisher describes a 24K curated subset of WebLINX interactions and dialogue; not a newly counted independent dataset."}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "1a445c7894c65c9a7c4247dba446c206c13ac2ecdc129ce93a0de943f2984a18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "69fe17825db0d5f32ccaa84b67fabfa10018e1410da0c70ec0ea9509703feba2", "supports": ["identity", "release_status", "details.release_revision", "details.architecture"], "url": "https://huggingface.co/api/models/McGill-NLP/Llama-3-8B-Web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "6bcca341999ab474e2973116128177b77f97cd64028eeace224035358ce6474f", "supports": ["licenses.weights"], "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T12:34:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Weight license text and card declaration are retained separately from framework code licensing.", "declared": "Meta Llama 3 Community License Agreement", "retrieved_at": "2026-10-17T12:36:55Z", "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web"}], "name": "Llama-3-8B-Web", "publication": null, "published_at": null, "record_sha256": "92920abd12d0659622e9841e20d200f1e24f42fe1763879112027a7737a888e4", "relations": [{"evidence_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md", "target_id": "github:mcgill-nlp/weblinx", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md", "target_id": "arxiv:2402.05930", "type": "documented_by"}], "resource_id": "url:7c3e8c4ba52428584079", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/McGill-NLP/Llama-3-8B-Web", "topics": ["web-navigation", "imitation-learning", "action-prediction"], "updated_at": null}, "text": "Llama-3-8B-Web\nResource type: model\nReview status: curated\nScope: core\n\nA Llama 3 8B instruction-model checkpoint fine-tuned on WebLINX web interactions and dialogue, predicting browser actions from a textual state representation.\nPlatforms: web\nTopics: web-navigation, imitation-learning, action-prediction\n\nRecorded specifications: {\"architecture\": \"LlamaForCausalLM\", \"availability\": \"Ungated complete weight/config/tokenizer manifest.\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/META_LLAMA_3_COMMUNITY_LICENSE_AGREEMENT.txt\", \"https://huggingface.co/McGill-NLP/Llama-3-8B-Web/raw/main/README.md\"], \"reason\": \"Weight license text and card declaration are retained separately from framework code licensing.\"}}, \"parameter_label\": \"8B\", \"release_revision\": \"09856c28ec324cf3fe4c513bdc4437a75b78185f\", \"training_data_note\": \"Publisher describes a 24K curated subset of WebLINX interactions and dialogue; not a newly counted independent dataset.\"}\n\nThe ungated Hub manifest includes model.safetensors, pytorch_model.bin, config and tokenizer files; the two weight formats represent the same checkpoint and are not counted separately. The card describes an approximately 24K-example curated WebLINX training subset and evaluates dialogue-guided web navigation. Predictions still require parsing and execution in an external browser environment. The retained weight license is the Meta Llama 3 Community License Agreement, not WebLlama code MIT. No weights were downloaded, model used, or agreement accepted.\n\nPrimary source: https://huggingface.co/McGill-NLP/Llama-3-8B-Web\nCatalog record: https://awesomecua.github.io/resources/url-7c3e8c4ba52428584079/"}
{"id": "url:7c568d88a472f914d316", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train", "citation_url": "https://awesomecua.github.io/resources/url-7c568d88a472f914d316/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/tree/main", "gated": false, "hub_id": "LZXzju/UI-R1-3B-Train"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "640cb5908abefb770ea6bb8d12f22e9eca622224fa73e9f5c67a3a95f4f7b8e5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "482d1304ad53736b3e410694cfa427067ff8b45267374964e4a8ff3833afbbd3", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LZXzju/UI-R1-3B-Train"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments", "summary"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train"}], "name": "UI-R1 3B training data", "publication": null, "published_at": null, "record_sha256": "49ccc4140fd14fd115d066b92ee7f403146b2dac397f5c8007902109ec5e3c19", "relations": [{"evidence_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "resource_id": "url:7c568d88a472f914d316", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train", "topics": ["data-generation", "grounding", "action-prediction"], "updated_at": null}, "text": "UI-R1 3B training data\nResource type: dataset\nReview status: curated\nScope: core\n\nThe released mobile training subset for the original UI-R1 3B model, with action annotations and individual screenshot files.\nPlatforms: android\nTopics: data-generation, grounding, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train/tree/main\", \"gated\": false, \"hub_id\": \"LZXzju/UI-R1-3B-Train\"}\n\nThe public manifest includes train.json and a train_imgs directory. The card has no dataset-license declaration, and the model’s MIT metadata is not inherited. The official repository identifies this as v1 training data drawn from AndroidControl and ScreenSpot; later enlarged training mixtures are distinct.\n\nPrimary source: https://huggingface.co/datasets/LZXzju/UI-R1-3B-Train\nCatalog record: https://awesomecua.github.io/resources/url-7c568d88a472f914d316/"}
{"id": "url:80858580c1009eba0e98", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC", "citation_url": "https://awesomecua.github.io/resources/url-80858580c1009eba0e98/", "details": {"disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Genesis-7B-AC"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "62de4360ad62165bd4dc37d5897d6f5f88813956dff33bac63a24ded50cca118", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "c35278027cd0e36e2fbbba4861031eca2a52090f20b40fb5aa680b067bb13dac", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Genesis-7B-AC"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "62de4360ad62165bd4dc37d5897d6f5f88813956dff33bac63a24ded50cca118", "supports": ["licenses.weights"], "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC"}], "name": "OS-Genesis 7B AC", "publication": null, "published_at": null, "record_sha256": "578a2c27c25df2f7c0b99173ddcee07dfe2cad25c4dfcdb30ea3b263e15b1e04", "relations": [{"evidence_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/raw/main/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "resource_id": "url:80858580c1009eba0e98", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC", "topics": ["mobile-navigation", "action-prediction", "imitation-learning"], "updated_at": null}, "text": "OS-Genesis 7B AC\nResource type: model\nReview status: curated\nScope: core\n\nThe AndroidControl-oriented Qwen2-VL checkpoint trained with mobile trajectories generated through OS-Genesis reverse task synthesis.\nPlatforms: android\nTopics: mobile-navigation, action-prediction, imitation-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Genesis-7B-AC\"}\n\nFour public safetensors shards are present. The card declares Apache-2.0. This is the AC task-specific checkpoint; AndroidWorld and WebArena variants are separate releases and their results are not transferred to it.\n\nPrimary source: https://huggingface.co/OS-Copilot/OS-Genesis-7B-AC\nCatalog record: https://awesomecua.github.io/resources/url-80858580c1009eba0e98/"}
{"id": "url:826f3753905c5c43bf46", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B", "citation_url": "https://awesomecua.github.io/resources/url-826f3753905c5c43bf46/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-9B", "gated": false, "hub_revision": "76b7e9f7ce97675b7d4e1ea6d88a6adc9ced2886", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "f215590b7a7b510085f7f99a8c236816604d4c11a8f5ae544b4ee75771dd4f82", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-9B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "feeb30bda7f86e10d87bdc1b0148d0632aba6df1676931a3d6c65cbdf197df3e", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B"}], "name": "Seg2Click Qwen3.5 9B", "publication": null, "published_at": null, "record_sha256": "98b0a736998fee17470356c36d819866e55c838cb9f9d5b71b279c6ea9dce69c", "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "url:826f3753905c5c43bf46", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B", "topics": ["grounding"], "updated_at": null}, "text": "Seg2Click Qwen3.5 9B\nResource type: model\nReview status: curated\nScope: core\n\nSeg2Click checkpoint pairing a 9B Qwen3.5 backbone with a decoder for valid GUI click regions.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.5-9B\", \"gated\": false, \"hub_revision\": \"76b7e9f7ce97675b7d4e1ea6d88a6adc9ced2886\", \"listed_weight_files\": [\"model.safetensors\", \"phase1_mask_decoder.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.\n\nPrimary source: https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-9B\nCatalog record: https://awesomecua.github.io/resources/url-826f3753905c5c43bf46/"}
{"id": "url:82a30f45d9e855f423c7", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B", "citation_url": "https://awesomecua.github.io/resources/url-82a30f45d9e855f423c7/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Atlas-Base-7B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "a3355c4f884490efe2dc31ad393b869349c35130cc37370dafa19f4f0f5016b7", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B"}], "name": "OS-Atlas Base 7B", "publication": null, "published_at": null, "record_sha256": "3e09828657cc75103e165d349bc6c73c57d1e65b02457ba9d1a817e599d32849", "relations": [{"evidence_url": "https://huggingface.co/api/models/OS-Copilot/OS-Atlas-Base-7B", "target_id": "github:os-copilot/os-atlas", "type": "variant_of"}], "resource_id": "url:82a30f45d9e855f423c7", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B", "topics": ["grounding"], "updated_at": null}, "text": "OS-Atlas Base 7B\nResource type: model\nReview status: curated\nScope: core\n\nOS-Atlas base checkpoint fine-tuned from Qwen2-VL-7B-Instruct, using the standard Qwen2-VL architecture in Transformers with about 8.3B BF16 parameters.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Atlas-Base-7B\"}\n\nPrimary source: https://huggingface.co/OS-Copilot/OS-Atlas-Base-7B\nCatalog record: https://awesomecua.github.io/resources/url-82a30f45d9e855f423c7/"}
{"id": "url:849a86beaee006c6b147", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Writer/omniact", "citation_url": "https://awesomecua.github.io/resources/url-849a86beaee006c6b147/", "details": {"benchmark_protocol": {"actions": ["PyAutoGUI script: pointer movement/click/drag, key presses, hotkeys, and text entry"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "train split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 6788}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "test split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 2020}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "population": "val split, HF revision 5e4bcb26", "qualifier": "exact", "unit": "task examples", "value": 991}], "evaluation": "The bundled evaluator reports weighted sequence match and action match, comparing action types and applying coordinate, key, and writing penalties. Target element-box metadata is evaluation-only. The retained scorer has a click-penalty defect: predicted coordinates are read from gold_line, so score reproduction requires explicit scorer-version review.", "evidence_urls": ["https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md", "https://huggingface.co/api/datasets/Writer/omniact", "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/eval.py"], "focus": ["offline desktop/web action-script generation"], "notes": "This released split table totals 9,799 examples; it is not silently substituted for any paper-level collection size. Public files include train/test/val JSON indexes, task/image archives, and eval.py. Evaluation is offline script comparison, not execution success on a live desktop. MIT is the Hub data declaration; original app/content rights are not resolved by it.", "observations": ["screenshot", "natural-language task"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/Writer/omniact/tree/main", "gated": false, "hub_id": "Writer/omniact", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "2db0a5dd2b983dded3b90a8ecd66cadbb39074f03778a339822418531661b3e3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/Writer/omniact"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:00Z", "role": "primary", "sha256": "7b54ad557f89d621fe39c3d2fe9f02164a2bfcb7fef0db610d964586d3673693", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:06Z", "role": "primary", "sha256": "a786de678af46df5e2b6b1a2a69d2ee239fe8fd5b63663765d84d39e21fc2365", "supports": ["details.benchmark_protocol"], "url": "https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/eval.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "2db0a5dd2b983dded3b90a8ecd66cadbb39074f03778a339822418531661b3e3", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Writer/omniact"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:27Z", "url": "https://huggingface.co/datasets/Writer/omniact"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Writer/omniact"}], "name": "OmniAct", "publication": null, "published_at": null, "record_sha256": "2a888de585329e162205ead48da175cc24fcb9557b5808978d6fa07512821afe", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/Writer/omniact", "target_id": "arxiv:2402.17553", "type": "documented_by"}], "resource_id": "url:849a86beaee006c6b147", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Writer/omniact", "topics": ["evaluation"], "updated_at": null}, "text": "OmniAct\nResource type: benchmark\nReview status: curated\nScope: core\n\nDataset and benchmark for generating computer interaction programs.\nPlatforms: desktop, web\nTopics: evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"PyAutoGUI script: pointer movement/click/drag, key presses, hotkeys, and text entry\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md\", \"population\": \"train split, HF revision 5e4bcb26\", \"qualifier\": \"exact\", \"unit\": \"task examples\", \"value\": 6788}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md\", \"population\": \"test split, HF revision 5e4bcb26\", \"qualifier\": \"exact\", \"unit\": \"task examples\", \"value\": 2020}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md\", \"population\": \"val split, HF revision 5e4bcb26\", \"qualifier\": \"exact\", \"unit\": \"task examples\", \"value\": 991}], \"evaluation\": \"The bundled evaluator reports weighted sequence match and action match, comparing action types and applying coordinate, key, and writing penalties. Target element-box metadata is evaluation-only. The retained scorer has a click-penalty defect: predicted coordinates are read from gold_line, so score reproduction requires explicit scorer-version review.\", \"evidence_urls\": [\"https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/README.md\", \"https://huggingface.co/api/datasets/Writer/omniact\", \"https://huggingface.co/datasets/Writer/omniact/raw/5e4bcb26745afc8d58ba9519ea9c9fb832386d26/eval.py\"], \"focus\": [\"offline desktop/web action-script generation\"], \"notes\": \"This released split table totals 9,799 examples; it is not silently substituted for any paper-level collection size. Public files include train/test/val JSON indexes, task/image archives, and eval.py. Evaluation is offline script comparison, not execution success on a live desktop. MIT is the Hub data declaration; original app/content rights are not resolved by it.\", \"observations\": [\"screenshot\", \"natural-language task\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/Writer/omniact/tree/main\", \"gated\": false, \"hub_id\": \"Writer/omniact\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/datasets/Writer/omniact\nCatalog record: https://awesomecua.github.io/resources/url-849a86beaee006c6b147/"}
{"id": "url:84e5aeeffaf561baa966", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k", "citation_url": "https://awesomecua.github.io/resources/url-84e5aeeffaf561baa966/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md", "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"], "formats": ["JSONL"], "modalities": ["text"], "notes": "The public manifest contains reward_data.jsonl but no separately packaged image archive. The card declares MIT. The 517k identifier is not converted into a verified sample count, and complete image packaging is not inferred from the JSONL filename.", "splits": [], "supervision": ["trajectory rewards"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/tree/main", "gated": false, "hub_id": "HanXiao1999/UI-Genie-RM-517k"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "dddb70b2b6986525f3980cbe50c3241d17f840a81fda9715b01ac3b25dd35d63", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "ab4bf30b6188959f825f2747ca711b0f190dddfe1658319c881ca9c73be9ecc9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/HanXiao1999/UI-Genie-RM-517k"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T03:59:54Z", "role": "primary", "sha256": "e08569bd9d8a6ff44537d4fe69243ddfeb8a41ae5c9f87ef84af50a67c64c726", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "dddb70b2b6986525f3980cbe50c3241d17f840a81fda9715b01ac3b25dd35d63", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k"}], "name": "UI-Genie reward data", "publication": null, "published_at": null, "record_sha256": "b3d4948a5d44466bd837d2d337b094162fa6a856bf8d25de779f10eb10c163a7", "relations": [{"evidence_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md", "target_id": "arxiv:2505.21496", "type": "documented_by"}], "resource_id": "url:84e5aeeffaf561baa966", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k", "topics": ["data-generation", "mobile-navigation", "reward-models"], "updated_at": null}, "text": "UI-Genie reward data\nResource type: dataset\nReview status: curated\nScope: core\n\nMobile GUI reward annotations released for training UI-Genie’s trajectory-assessment model.\nPlatforms: android\nTopics: data-generation, mobile-navigation, reward-models\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/raw/main/README.md\", \"https://raw.githubusercontent.com/Euphoria16/UI-Genie/HEAD/README.md\"], \"formats\": [\"JSONL\"], \"modalities\": [\"text\"], \"notes\": \"The public manifest contains reward_data.jsonl but no separately packaged image archive. The card declares MIT. The 517k identifier is not converted into a verified sample count, and complete image packaging is not inferred from the JSONL filename.\", \"splits\": [], \"supervision\": [\"trajectory rewards\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k/tree/main\", \"gated\": false, \"hub_id\": \"HanXiao1999/UI-Genie-RM-517k\"}\n\nThe public manifest contains reward_data.jsonl but no separately packaged image archive. The card declares MIT. The 517k identifier is not converted into a verified sample count, and complete image packaging is not inferred from the JSONL filename.\n\nPrimary source: https://huggingface.co/datasets/HanXiao1999/UI-Genie-RM-517k\nCatalog record: https://awesomecua.github.io/resources/url-84e5aeeffaf561baa966/"}
{"id": "url:879d4e3e75e83582bde6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://ai.google.dev/gemini-api/docs/computer-use", "citation_url": "https://awesomecua.github.io/resources/url-879d4e3e75e83582bde6/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "e7062cfb4c0948d3a2274843f9f493f34cb40dfd634371408086df6141a0e036", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ai.google.dev/gemini-api/docs/computer-use"}], "name": "Gemini Computer Use", "publication": null, "published_at": null, "record_sha256": "f2bcf7f55cd235f683251ae65da30a3d24f6f080ba4d893b2591e971ad2cade1", "relations": [{"evidence_url": "https://ai.google.dev/gemini-api/docs/computer-use", "target_id": "organization:google", "type": "published_by"}], "resource_id": "url:879d4e3e75e83582bde6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://ai.google.dev/gemini-api/docs/computer-use", "topics": ["planning", "grounding"], "updated_at": null}, "text": "Gemini Computer Use\nResource type: model\nReview status: curated\nScope: core\n\nGemini capability for generating interface actions from screenshots.\nPlatforms: cross-platform\nTopics: planning, grounding\n\nPrimary source: https://ai.google.dev/gemini-api/docs/computer-use\nCatalog record: https://awesomecua.github.io/resources/url-879d4e3e75e83582bde6/"}
{"id": "url:8896d44370711136fd54", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.agentql.com/", "citation_url": "https://awesomecua.github.io/resources/url-8896d44370711136fd54/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "41ec857dacdb106ad99abbe1872f40a295056522ffaa1b1714b51dffb370f450", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.agentql.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.agentql.com/"}], "name": "AgentQL", "publication": null, "published_at": null, "record_sha256": "3e8cfaf196caf1ad49726547a436bf0ef3dc3f470d111f15f681405aa4c672a0", "relations": [], "resource_id": "url:8896d44370711136fd54", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.agentql.com/", "topics": ["grounding", "web-navigation"], "updated_at": null}, "text": "AgentQL\nResource type: product\nReview status: curated\nScope: core\n\nNatural-language querying and interaction tools for web interfaces.\nPlatforms: web\nTopics: grounding, web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.agentql.com/\nCatalog record: https://awesomecua.github.io/resources/url-8896d44370711136fd54/"}
{"id": "url:8933977bb0a57940ed23", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/showlab/ShowUI-web", "citation_url": "https://awesomecua.github.io/resources/url-8933977bb0a57940ed23/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "rows", "value": 21988}], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"], "formats": ["Parquet", "JSON", "TAR.GZ"], "modalities": ["screenshots", "text", "bounding boxes", "points"], "notes": "Row count is the card dataset_info value; one row may contain multiple instructions. The card summarizes screenshots as approximately 22K. No data license is declared.", "splits": ["train"], "supervision": ["element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-web/tree/main", "gated": false, "hub_id": "showlab/ShowUI-web"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "08586aa1ded1f1a1f984fb55697fdb1007d86c7fdf4988a2dc1fcf3fdfd49a47", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "12412480f1a3957069d9dcdc53abce4ad3e34be7dac7ca6fe1726d04764cadfd", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-web"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-web"}], "name": "ShowUI-web", "publication": null, "published_at": null, "record_sha256": "14dd48293e80de723094a6a2ad0dfdc057699ca74468ed5d65a658407e5c87d5", "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md", "target_id": "arxiv:2411.17465", "type": "documented_by"}], "resource_id": "url:8933977bb0a57940ed23", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/showlab/ShowUI-web", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "ShowUI-web\nResource type: dataset\nReview status: curated\nScope: core\n\nWeb-interface grounding examples with instructions, element boxes and points, paired with a separately packaged screenshot archive.\nPlatforms: web\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md\", \"population\": \"train split\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 21988}], \"evidence_urls\": [\"https://huggingface.co/datasets/showlab/ShowUI-web/raw/main/README.md\"], \"formats\": [\"Parquet\", \"JSON\", \"TAR.GZ\"], \"modalities\": [\"screenshots\", \"text\", \"bounding boxes\", \"points\"], \"notes\": \"Row count is the card dataset_info value; one row may contain multiple instructions. The card summarizes screenshots as approximately 22K. No data license is declared.\", \"splits\": [\"train\"], \"supervision\": [\"element grounding\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/showlab/ShowUI-web/tree/main\", \"gated\": false, \"hub_id\": \"showlab/ShowUI-web\"}\n\nPublic Parquet annotations, metadata JSON and images.tar.gz are present. The card refers to image.tar.gz in its extraction command, while the actual archive is plural. No data-license declaration is present; the ShowUI code or model licenses are not inherited.\n\nPrimary source: https://huggingface.co/datasets/showlab/ShowUI-web\nCatalog record: https://awesomecua.github.io/resources/url-8933977bb0a57940ed23/"}
{"id": "url:8c9aaafb918aaf8b4e92", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M", "citation_url": "https://awesomecua.github.io/resources/url-8c9aaafb918aaf8b4e92/", "details": {"availability": "Ungated annotation files and multipart image-archive manifest.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md", "population": "TongUI-collected GUI-Net-1M data, excluding added external mixtures", "qualifier": "approximately", "unit": "training trajectories", "value": 1000000}], "evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md", "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"], "formats": ["JSON annotations", "multipart image archives"], "modalities": ["text", "screenshots"], "notes": "The source reports an approximate trajectory total; it does not establish per-platform counts or a complete test-split inventory.", "splits": ["training collection"], "supervision": ["GUI trajectory supervision from multimodal tutorials"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"], "reason": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights."}}, "release_revision": "60d1e644fcff7d3b62093baec03fe7efeb0d98a1"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:30:54Z", "role": "primary", "sha256": "f30a18d1a533b8f71f436dae2ee7369f8971422bd76527e069499c5bba911912", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "314a77eb0ba43bedffbe6869601e41aea41993eaca59d86ba346652c853802d2", "supports": ["identity", "release_status", "details.release_revision", "details.dataset_profile.formats"], "url": "https://huggingface.co/api/datasets/Bofeee5675/GUI-Net-1M"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:36:55Z", "role": "primary", "sha256": "b51acb70c95d079ccbe1de0a685ad7c8a4b3b5bed11e6162e6f7c2f60d78805e", "supports": ["summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "c2df5c3a6dd576ae53f0dc0f72ecab07d93ed034b04336abd2072a6f0d812c64", "supports": ["licenses.data", "details.dataset_profile.formats", "notes"], "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:30:54Z", "licenses": {"code": null, "data": {"basis": "Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T12:32:07Z", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M"}], "name": "GUI-Net-1M", "publication": null, "published_at": null, "record_sha256": "a159799dfd5a5becd2c7dbcd4589d2f31829ef73caf144b626a8ce44e5e95eed", "relations": [{"evidence_url": "https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/README.md", "target_id": "arxiv:2504.12679", "type": "documented_by"}], "resource_id": "url:8c9aaafb918aaf8b4e92", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M", "topics": ["data-generation", "grounding", "imitation-learning"], "updated_at": null}, "text": "GUI-Net-1M\nResource type: dataset\nReview status: curated\nScope: core\n\nA TongUI-collected training corpus from multimodal web tutorials, distributed as annotation files and split screenshot archives for GUI trajectory supervision.\nPlatforms: unspecified\nTopics: data-generation, grounding, imitation-learning\n\nRecorded specifications: {\"availability\": \"Ungated annotation files and multipart image-archive manifest.\", \"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md\", \"population\": \"TongUI-collected GUI-Net-1M data, excluding added external mixtures\", \"qualifier\": \"approximately\", \"unit\": \"training trajectories\", \"value\": 1000000}], \"evidence_urls\": [\"https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md\", \"https://raw.githubusercontent.com/TongUI-agent/TongUI-agent/HEAD/docs/train.md\"], \"formats\": [\"JSON annotations\", \"multipart image archives\"], \"modalities\": [\"text\", \"screenshots\"], \"notes\": \"The source reports an approximate trajectory total; it does not establish per-platform counts or a complete test-split inventory.\", \"splits\": [\"training collection\"], \"supervision\": [\"GUI trajectory supervision from multimodal tutorials\"]}, \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M/raw/main/README.md\"], \"reason\": \"Publisher declaration for this dataset packaging; original web/tutorial materials retain applicable rights.\"}}, \"release_revision\": \"60d1e644fcff7d3b62093baec03fe7efeb0d98a1\"}\n\nThe ungated manifest lists JSON annotations and multipart image archives. Training documentation describes roughly one million collected training trajectories from Wikihow, Baidu Experience and GUI videos; this is distinct from larger mixtures that add external datasets. Platform-specific distribution and exact row counts were not established from this release metadata. Apache-2.0 is the publisher dataset declaration, without a grant of additional rights to upstream tutorial content. No image archive or example record was downloaded.\n\nPrimary source: https://huggingface.co/datasets/Bofeee5675/GUI-Net-1M\nCatalog record: https://awesomecua.github.io/resources/url-8c9aaafb918aaf8b4e92/"}
{"id": "url:8dfc84eb4016204fe70a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair", "citation_url": "https://awesomecua.github.io/resources/url-8dfc84eb4016204fe70a/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "population": "Hub split named test", "qualifier": "exact", "unit": "rows", "value": 9460}], "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "action history"], "notes": "Preserves the publisher’s split label despite the collection’s training purpose. Rows are not individual preference pairs because rejected is a list. Data license is unknown.", "splits": ["test"], "supervision": ["preferred and rejected continuations", "task checklists"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebPRMCollection_preference_pair"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "9c091c5fc625ce53f53f8262f5cee8d08194e4fd1b8eeed543ff1c3723c823fd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "28a1d5d12e84fc38515bd3c5031ae0cf315c5f8772b90d982d926b98f2857563", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_preference_pair"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair"}], "name": "WebPRM Collection preference data", "publication": null, "published_at": null, "record_sha256": "5df4263ba5a73a61089d4b099c8bdfadcd58043be229e0ce35a7894c2360d409", "relations": [{"evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "resource_id": "url:8dfc84eb4016204fe70a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair", "topics": ["web-navigation", "reward-models", "evaluation"], "updated_at": null}, "text": "WebPRM Collection preference data\nResource type: dataset\nReview status: curated\nScope: core\n\nChecklist-annotated web decision records with page observations, prior actions, one preferred continuation and rejected alternatives for process-reward learning.\nPlatforms: web\nTopics: web-navigation, reward-models, evaluation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md\", \"population\": \"Hub split named test\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 9460}], \"evidence_urls\": [\"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/raw/main/README.md\", \"https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"action history\"], \"notes\": \"Preserves the publisher’s split label despite the collection’s training purpose. Rows are not individual preference pairs because rejected is a list. Data license is unknown.\", \"splits\": [\"test\"], \"supervision\": [\"preferred and rejected continuations\", \"task checklists\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair/tree/main\", \"gated\": false, \"hub_id\": \"LangAGI-Lab/WebPRMCollection_preference_pair\"}\n\nTen public Parquet shards are present. The card names its split test and reports 9,460 rows; the project’s 40K-plus preference-annotation claim is a different grain and is not substituted for this row count. A row can contain multiple rejected alternatives. No data-license declaration is present.\n\nPrimary source: https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_preference_pair\nCatalog record: https://awesomecua.github.io/resources/url-8dfc84eb4016204fe70a/"}
{"id": "url:8ef3985dce697e5feea9", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/openbmb/AgentCPM-GUI", "citation_url": "https://awesomecua.github.io/resources/url-8ef3985dce697e5feea9/", "details": {"disabled": false, "download_url": "https://huggingface.co/openbmb/AgentCPM-GUI/tree/main", "gated": false, "hub_id": "openbmb/AgentCPM-GUI"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "4d76c8cc19f1ff4c850ba6cc566d50bbd2355e10921a2214e38c921cedfeed4c", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/openbmb/AgentCPM-GUI"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "9016a68e833a7717c922d9363a6d6c56cc7cf6a7087543f5694507668fe9ab75", "supports": ["licenses.weights"], "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/openbmb/AgentCPM-GUI"}], "name": "AgentCPM-GUI 8B checkpoint", "publication": null, "published_at": null, "record_sha256": "7ef4515b352d9e85cc1d17c66ccee4e66989d40582161405272d7ab7ab5f4fa2", "relations": [{"evidence_url": "https://huggingface.co/openbmb/AgentCPM-GUI/raw/main/README.md", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "resource_id": "url:8ef3985dce697e5feea9", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/openbmb/AgentCPM-GUI", "topics": ["mobile-navigation", "grounding", "reinforcement-learning"], "updated_at": null}, "text": "AgentCPM-GUI 8B checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nAn eight-billion-parameter MiniCPM-V-based model that consumes smartphone screenshots and predicts compact actions for Chinese and English Android app tasks.\nPlatforms: android\nTopics: mobile-navigation, grounding, reinforcement-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/openbmb/AgentCPM-GUI/tree/main\", \"gated\": false, \"hub_id\": \"openbmb/AgentCPM-GUI\"}\n\nThe Hub provides four safetensors shards and associated model files without gating. Weight metadata declares Apache-2.0, while the prose license section refers specifically to code; these scopes are kept distinct. The card describes Android-app operation, and the CAGUI evaluation data remains a separate existing catalog asset.\n\nPrimary source: https://huggingface.co/openbmb/AgentCPM-GUI\nCatalog record: https://awesomecua.github.io/resources/url-8ef3985dce697e5feea9/"}
{"id": "url:901ca855640457dfe2f1", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Toby Jia-Jun Li", "Oriana Riva"], "canonical_source_url": "https://doi.org/10.1145/3210240.3210339", "citation_url": "https://awesomecua.github.io/resources/url-901ca855640457dfe2f1/", "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Oriana Riva"], "date_basis": "Supplied month precision: 2018-06. Publication line: MobiSys 2018, June 2018", "date_parts": [2018, 6], "doi": "10.1145/3210240.3210339", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Final proceedings page range and volume are not established in the retained first-party citation. The paper states a 14-page extent, not a page range. Exact publication day is unknown.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf", "publisher": null, "series": null, "title": "Kite: Building Conversational Bots from Mobile Apps", "type": "conference-paper", "url": "https://doi.org/10.1145/3210240.3210339", "venue": "The 16th Annual International Conference on Mobile Systems, Applications, and Services", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:26Z", "role": "primary", "sha256": "e2f5403d77f563ec6a88458e32d057559a8f8ba4d2a80304e70c2c793fb60bb5", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes"], "url": "https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:27Z", "role": "primary", "sha256": "bfb67283f49dad7974d97e3b3da9d9ff24744d18e83c4d25cbe4f12ecf807071", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:27Z", "role": "primary", "sha256": "bfb67283f49dad7974d97e3b3da9d9ff24744d18e83c4d25cbe4f12ecf807071", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:59:26Z", "role": "primary", "sha256": "e2f5403d77f563ec6a88458e32d057559a8f8ba4d2a80304e70c2c793fb60bb5", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes"], "url": "https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/"}], "identifiers": {"doi": "10.1145/3210240.3210339"}, "kind": "paper", "last_verified_at": "2026-10-17T09:59:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3210240.3210339"}, {"rel": "publication_pdf", "url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"}], "name": "Kite: Building Conversational Bots from Mobile Apps", "publication": {"authors": ["Toby Jia-Jun Li", "Oriana Riva"], "date_basis": "Supplied month precision: 2018-06. Publication line: MobiSys 2018, June 2018", "date_parts": [2018, 6], "doi": "10.1145/3210240.3210339", "event_dates": null, "evidence_urls": ["https://www.microsoft.com/en-us/research/publication/kite-building-conversational-bots-mobile-apps/", "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Final proceedings page range and volume are not established in the retained first-party citation. The paper states a 14-page extent, not a page range. Exact publication day is unknown.", "pages": null, "pdf_url": "https://www.microsoft.com/en-us/research/wp-content/uploads/2018/05/kite_mobisys18.pdf", "publisher": null, "series": null, "title": "Kite: Building Conversational Bots from Mobile Apps", "type": "conference-paper", "url": "https://doi.org/10.1145/3210240.3210339", "venue": "The 16th Annual International Conference on Mobile Systems, Applications, and Services", "volume": null}, "published_at": null, "record_sha256": "f08bad499bd2fe1ac17d2631b24a0ef42b167413991f38e06e9fd37934b4b5e0", "relations": [], "resource_id": "url:901ca855640457dfe2f1", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/3210240.3210339", "topics": ["mobile-navigation", "workflow", "human-in-the-loop"], "updated_at": null}, "text": "Kite: Building Conversational Bots from Mobile Apps\nResource type: paper\nReview status: curated\nScope: core\n\nDerives task graphs from demonstrated Android app interactions and layers a conversational question-answer interface over those graphs. The evaluation examines task-model extraction and developer use of the resulting bot templates.\nPlatforms: android\nTopics: mobile-navigation, workflow, human-in-the-loop\nRecord publication date (recorded precision): 2018-06\nAuthors: Toby Jia-Jun Li, Oriana Riva\n\nPublication edition: Kite: Building Conversational Bots from Mobile Apps\nPublication authors: Toby Jia-Jun Li, Oriana Riva\nPublication edition date (recorded precision): 2018-06\nPublication venue: The 16th Annual International Conference on Mobile Systems, Applications, and Services\nPublication doi: 10.1145/3210240.3210339\nPublication date basis: Supplied month precision: 2018-06. Publication line: MobiSys 2018, June 2018\nPublication notes: Final proceedings page range and volume are not established in the retained first-party citation. The paper states a 14-page extent, not a page range. Exact publication day is unknown.\nPublication source: https://doi.org/10.1145/3210240.3210339\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. \n\nPrimary source: https://doi.org/10.1145/3210240.3210339\nCatalog record: https://awesomecua.github.io/resources/url-901ca855640457dfe2f1/"}
{"id": "url:90eea6f6d099a3297850", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/inclusionAI/GUI-G2-7B", "citation_url": "https://awesomecua.github.io/resources/url-90eea6f6d099a3297850/", "details": {"disabled": false, "download_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/tree/main", "gated": false, "hub_id": "inclusionAI/GUI-G2-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "c1f11fea30d3a64445e983a011db4efb9515a4221ad770e3a5223699e665eebe", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/inclusionAI/GUI-G2-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "cf53a868a010a20b165386edaaafed8918071c35a3babf1766fbb51f020cb4a1", "supports": ["licenses.weights"], "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/inclusionAI/GUI-G2-7B"}], "name": "GUI-G2 7B", "publication": null, "published_at": null, "record_sha256": "91786af398888d5d750c1be2231c1edfb1f74260f584983abd858dc4f4ae6432", "relations": [{"evidence_url": "https://huggingface.co/inclusionAI/GUI-G2-7B/raw/main/README.md", "target_id": "arxiv:2507.15846", "type": "documented_by"}], "resource_id": "url:90eea6f6d099a3297850", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/inclusionAI/GUI-G2-7B", "topics": ["grounding", "reinforcement-learning", "reward-models"], "updated_at": null}, "text": "GUI-G2 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL grounding checkpoint trained with Gaussian point and coverage rewards whose granularity adapts to interface-element scale.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning, reward-models\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/inclusionAI/GUI-G2-7B/tree/main\", \"gated\": false, \"hub_id\": \"inclusionAI/GUI-G2-7B\"}\n\nThe model card and four public safetensors shards establish this specific release. Apache-2.0 is declared for the model. The release targets element grounding; its grounding results do not establish long-horizon task-completion success.\n\nPrimary source: https://huggingface.co/inclusionAI/GUI-G2-7B\nCatalog record: https://awesomecua.github.io/resources/url-90eea6f6d099a3297850/"}
{"id": "url:964d70c40bfc5ca8e1ee", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xlangai/AgentNet", "citation_url": "https://awesomecua.github.io/resources/url-964d70c40bfc5ca8e1ee/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md", "population": "AgentNet collection", "qualifier": "approximately", "unit": "tasks", "value": 22600}], "evidence_urls": ["https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md"], "formats": ["JSONL", "image archives"], "modalities": ["screenshots", "instructions", "actions", "generated reasoning"], "notes": "The card distinguishes human task instructions from generated summaries, reasoning, reflections, and quality fields. Generated labels are not independent human quality acceptance records.", "splits": [], "supervision": ["computer-use trajectories", "synthesized reasoning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/AgentNet/tree/main", "gated": false, "hub_id": "xlangai/AgentNet", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "f23c588ea9fb27210f9d605e164a32527a24298ca516df98272ff7ca7110eac6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/xlangai/AgentNet"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:40:46Z", "role": "primary", "sha256": "327a77856df259db0a0c2405478f75a0d0f2bf17c51806abc7c509ad82c3ad57", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "f23c588ea9fb27210f9d605e164a32527a24298ca516df98272ff7ca7110eac6", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/AgentNet"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/xlangai/AgentNet"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/AgentNet"}], "name": "AgentNet", "publication": null, "published_at": null, "record_sha256": "b63e884e3021b52f4cc43f9752704d57e6f544da94451002e3a86323fd61497a", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/xlangai/AgentNet", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "resource_id": "url:964d70c40bfc5ca8e1ee", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xlangai/AgentNet", "topics": ["data-generation"], "updated_at": null}, "text": "AgentNet\nResource type: dataset\nReview status: curated\nScope: core\n\n22.6K human-annotated desktop computer-use tasks on Windows, macOS and Ubuntu, released with OpenCUA for training and evaluating vision-language-action models.\nPlatforms: desktop\nTopics: data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md\", \"population\": \"AgentNet collection\", \"qualifier\": \"approximately\", \"unit\": \"tasks\", \"value\": 22600}], \"evidence_urls\": [\"https://huggingface.co/datasets/xlangai/AgentNet/raw/main/README.md\"], \"formats\": [\"JSONL\", \"image archives\"], \"modalities\": [\"screenshots\", \"instructions\", \"actions\", \"generated reasoning\"], \"notes\": \"The card distinguishes human task instructions from generated summaries, reasoning, reflections, and quality fields. Generated labels are not independent human quality acceptance records.\", \"splits\": [], \"supervision\": [\"computer-use trajectories\", \"synthesized reasoning\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xlangai/AgentNet/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/AgentNet\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/datasets/xlangai/AgentNet\nCatalog record: https://awesomecua.github.io/resources/url-964d70c40bfc5ca8e1ee/"}
{"id": "url:978d1dcd70fde8661b9e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.browserbase.com/", "citation_url": "https://awesomecua.github.io/resources/url-978d1dcd70fde8661b9e/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "d86315819750e4a5fde12bce85a03c61cf68e704cb54146e6180027a37471b72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.browserbase.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.browserbase.com/"}], "name": "Browserbase Cloud", "publication": null, "published_at": null, "record_sha256": "a2d4d6f09d9874c5eb5234339b1d3f82d40284683d24c2c8d694fbeb85b4787d", "relations": [{"evidence_url": "https://www.browserbase.com/", "target_id": "organization:browserbase", "type": "published_by"}], "resource_id": "url:978d1dcd70fde8661b9e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.browserbase.com/", "topics": ["sandbox"], "updated_at": null}, "text": "Browserbase Cloud\nResource type: product\nReview status: curated\nScope: core\n\nManaged browser infrastructure for agent applications.\nPlatforms: web\nTopics: sandbox\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.browserbase.com/\nCatalog record: https://awesomecua.github.io/resources/url-978d1dcd70fde8661b9e/"}
{"id": "url:97c0b1896525632434e2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Sarah E. Chasins", "Maria Mueller", "Rastislav Bodik"], "canonical_source_url": "https://doi.org/10.1145/3242587.3242661", "citation_url": "https://awesomecua.github.io/resources/url-97c0b1896525632434e2/", "details": {"publication": {"authors": ["Sarah E. Chasins", "Maria Mueller", "Rastislav Bodik"], "date_basis": "Supplied year precision: 2018. Rousillon entry: UIST 2018", "date_parts": [2018], "doi": "10.1145/3242587.3242661", "event_dates": null, "evidence_urls": ["https://schasins.com/papers/", "https://schasins.com/assets/papers/rousillon.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This is browser interaction learned from demonstrations, rather than a generic server-side crawling library. Final proceedings page range and volume are not established. Manuscript pages 1-13 are not substituted for publisher pagination. Exact publication month/day is not inferred from conference dates.", "pages": null, "pdf_url": "https://schasins.com/assets/papers/rousillon.pdf", "publisher": null, "series": null, "title": "Rousillon: Scraping Distributed Hierarchical Web Data", "type": "conference-paper", "url": "https://doi.org/10.1145/3242587.3242661", "venue": "Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:38Z", "role": "primary", "sha256": "6b7e584bb3862e0de51ca164d0f311c2a853e97bd63d56658ce62f0b49d86557", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://schasins.com/papers/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:12Z", "role": "primary", "sha256": "764dd1f5fe1db398188f403b4b1f6e3658e3238e4d8aac9ceed360dcad87e818", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.url"], "url": "https://schasins.com/assets/papers/rousillon.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:12Z", "role": "primary", "sha256": "764dd1f5fe1db398188f403b4b1f6e3658e3238e4d8aac9ceed360dcad87e818", "supports": ["details.publication.authors", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "details.publication.url", "environments", "name", "scope", "summary", "topics"], "url": "https://schasins.com/assets/papers/rousillon.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:57:38Z", "role": "primary", "sha256": "6b7e584bb3862e0de51ca164d0f311c2a853e97bd63d56658ce62f0b49d86557", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://schasins.com/papers/"}], "identifiers": {"doi": "10.1145/3242587.3242661"}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3242587.3242661"}, {"rel": "publication_pdf", "url": "https://schasins.com/assets/papers/rousillon.pdf"}], "name": "Rousillon: Scraping Distributed Hierarchical Web Data", "publication": {"authors": ["Sarah E. Chasins", "Maria Mueller", "Rastislav Bodik"], "date_basis": "Supplied year precision: 2018. Rousillon entry: UIST 2018", "date_parts": [2018], "doi": "10.1145/3242587.3242661", "event_dates": null, "evidence_urls": ["https://schasins.com/papers/", "https://schasins.com/assets/papers/rousillon.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This is browser interaction learned from demonstrations, rather than a generic server-side crawling library. Final proceedings page range and volume are not established. Manuscript pages 1-13 are not substituted for publisher pagination. Exact publication month/day is not inferred from conference dates.", "pages": null, "pdf_url": "https://schasins.com/assets/papers/rousillon.pdf", "publisher": null, "series": null, "title": "Rousillon: Scraping Distributed Hierarchical Web Data", "type": "conference-paper", "url": "https://doi.org/10.1145/3242587.3242661", "venue": "Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology", "volume": null}, "published_at": null, "record_sha256": "654e5feeb40aa424c08b26730a3b40bac5ff46b166e08f93808717d5dc4245e7", "relations": [], "resource_id": "url:97c0b1896525632434e2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/3242587.3242661", "topics": ["web-navigation", "recording", "automation"], "updated_at": null}, "text": "Rousillon: Scraping Distributed Hierarchical Web Data\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Rousillon, which generalizes a demonstration of collecting one hierarchical data row into a browser-automation program that traverses related pages. The paper develops relation-selection and generalization methods and evaluates users building such scripts.\nPlatforms: web\nTopics: web-navigation, recording, automation\nRecord publication date (recorded precision): 2018\nAuthors: Sarah E. Chasins, Maria Mueller, Rastislav Bodik\n\nPublication edition: Rousillon: Scraping Distributed Hierarchical Web Data\nPublication authors: Sarah E. Chasins, Maria Mueller, Rastislav Bodik\nPublication edition date (recorded precision): 2018\nPublication venue: Proceedings of the 31st Annual ACM Symposium on User Interface Software and Technology\nPublication doi: 10.1145/3242587.3242661\nPublication date basis: Supplied year precision: 2018. Rousillon entry: UIST 2018\nPublication notes: This is browser interaction learned from demonstrations, rather than a generic server-side crawling library. Final proceedings page range and volume are not established. Manuscript pages 1-13 are not substituted for publisher pagination. Exact publication month/day is not inferred from conference dates.\nPublication source: https://doi.org/10.1145/3242587.3242661\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This is browser interaction learned from demonstrations, rather than a generic server-side crawling library.\n\nPrimary source: https://doi.org/10.1145/3242587.3242661\nCatalog record: https://awesomecua.github.io/resources/url-97c0b1896525632434e2/"}
{"id": "url:9a3506ca6ef034adb424", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset", "citation_url": "https://awesomecua.github.io/resources/url-9a3506ca6ef034adb424/", "details": {"availability": "Public ungated JSON artifact is listed by the Hub manifest.", "dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/BlitherBoom/AutoDroid-V2-Dataset", "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"], "formats": ["JSON in ShareGPT conversation format"], "modalities": ["text"], "notes": "The card has license metadata only. The project training README supplies formatting and purpose. No sample count or split inventory is inferred from the training configuration.", "splits": [], "supervision": ["supervised conversations for AutoDroid-V2 fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/tree/main", "gated": false, "hub_id": "BlitherBoom/AutoDroid-V2-Dataset", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"], "reason": "Publisher dataset-card license declaration; does not establish permissions for model weights or third-party application content."}}, "release_revision": "f041810dd32a2f2999647b7d4f4b8438482c01b8"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:33Z", "role": "primary", "sha256": "1b300edfbf44167da71639a56f8ab668dabd33769575fb313ebef5220fc1bddb", "supports": ["name", "kind", "identity", "release_status", "details.hub_id", "details.gated", "details.disabled", "details.release_revision", "details.availability", "details.download_url", "details.dataset_profile.modalities"], "url": "https://huggingface.co/api/datasets/BlitherBoom/AutoDroid-V2-Dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:15:33Z", "role": "primary", "sha256": "d8d7a46d41a1a37fe4f0a5f637bf55c649310185329127d8a2204632e480be17", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "8ba37983a2c2da53d96a52696c07e4bdbea4e987a1a07b557935d614c3be46d2", "supports": ["summary", "topics", "scope", "notes", "details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:14:58Z", "role": "primary", "sha256": "74508100843f8db6db6bb151e77ac1cc604a6b7ebd16d3225266d73434465136", "supports": ["environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/README_minimal.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T09:15:33Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card license declaration; does not establish permissions for model weights or third-party application content.", "declared": "mit", "retrieved_at": "2026-10-17T09:15:33Z", "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset"}], "name": "AutoDroid V2 training data", "publication": null, "published_at": null, "record_sha256": "dcf0174f16637a35fb37d27b8ee85d47c7ca4aa31d6b68100ba22d16fd00bd10", "relations": [{"evidence_url": "https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md", "target_id": "github:mobilellm/autodroid-v2", "type": "component_of"}], "resource_id": "url:9a3506ca6ef034adb424", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset", "topics": ["data-generation", "imitation-learning", "mobile-navigation"], "updated_at": null}, "text": "AutoDroid V2 training data\nResource type: dataset\nReview status: curated\nScope: core\n\nConversation-formatted supervision released for fine-tuning AutoDroid V2, with the project documenting a ShareGPT-style JSON configuration.\nPlatforms: android\nTopics: data-generation, imitation-learning, mobile-navigation\n\nRecorded specifications: {\"availability\": \"Public ungated JSON artifact is listed by the Hub manifest.\", \"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/api/datasets/BlitherBoom/AutoDroid-V2-Dataset\", \"https://raw.githubusercontent.com/MobileLLM/AutoDroid-V2/HEAD/step_3_training/README.md\"], \"formats\": [\"JSON in ShareGPT conversation format\"], \"modalities\": [\"text\"], \"notes\": \"The card has license metadata only. The project training README supplies formatting and purpose. No sample count or split inventory is inferred from the training configuration.\", \"splits\": [], \"supervision\": [\"supervised conversations for AutoDroid-V2 fine-tuning\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/tree/main\", \"gated\": false, \"hub_id\": \"BlitherBoom/AutoDroid-V2-Dataset\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset/raw/main/README.md\"], \"reason\": \"Publisher dataset-card license declaration; does not establish permissions for model weights or third-party application content.\"}}, \"release_revision\": \"f041810dd32a2f2999647b7d4f4b8438482c01b8\"}\n\nThe public Hub manifest lists AutoDroid-V2-Dataset.json; no training records were downloaded. The card contains only an MIT metadata declaration, so the description and intended training use come from the official repository training guide. Sample count, split sizes and third-party-content rights remain unverified; the training command max_samples value is not treated as a dataset count.\n\nPrimary source: https://huggingface.co/datasets/BlitherBoom/AutoDroid-V2-Dataset\nCatalog record: https://awesomecua.github.io/resources/url-9a3506ca6ef034adb424/"}
{"id": "url:9dc9d8d3b9a4b32d6f4b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1", "citation_url": "https://awesomecua.github.io/resources/url-9dc9d8d3b9a4b32d6f4b/", "details": {"disabled": false, "download_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/tree/main", "gated": false, "hub_id": "Yuqi-Zhou/GUI-G1-3B-v1"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "2f4b2084ca9cfd67f0952ea8c578c87f3a5d721064b5de3a191d7489373125aa", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Yuqi-Zhou/GUI-G1-3B-v1"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "b7a4628a4d65ae226b50086075f3977b8f69ad12d15238a194675875bb9113e8", "supports": ["licenses.weights"], "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1"}], "name": "GUI-G1 3B v1", "publication": null, "published_at": null, "record_sha256": "941a7511d91561bb8a926ca933dee3c0514d783e6fa82292600073c68b83ce20", "relations": [{"evidence_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/raw/main/README.md", "target_id": "arxiv:2505.15810", "type": "documented_by"}], "resource_id": "url:9dc9d8d3b9a4b32d6f4b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1", "topics": ["grounding", "reinforcement-learning"], "updated_at": null}, "text": "GUI-G1 3B v1\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-3B checkpoint released for R1-style visual grounding experiments, predicting interface target locations from screenshots and instructions.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1/tree/main\", \"gated\": false, \"hub_id\": \"Yuqi-Zhou/GUI-G1-3B-v1\"}\n\nTwo public safetensors shards and model/tokenizer configuration are present. The card declares MIT. Its inference example loads the v1 model but names a different 0.1K processor checkpoint; that example was not executed, and compatibility is not asserted.\n\nPrimary source: https://huggingface.co/Yuqi-Zhou/GUI-G1-3B-v1\nCatalog record: https://awesomecua.github.io/resources/url-9dc9d8d3b9a4b32d6f4b/"}
{"id": "url:a06622da7688e3718a7b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/openbmb/CAGUI", "citation_url": "https://awesomecua.github.io/resources/url-a06622da7688e3718a7b/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/openbmb/CAGUI/tree/main", "gated": false, "hub_id": "openbmb/CAGUI", "languages": ["zh"], "modalities": [], "size_category_reported": ["1K<n<10K"]}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "98699e13f7c7f5af812e1cbbd5e71eaceefa8aa73ca37167a772774a40818efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/openbmb/CAGUI"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "98699e13f7c7f5af812e1cbbd5e71eaceefa8aa73ca37167a772774a40818efd", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/openbmb/CAGUI"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-nc-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/openbmb/CAGUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/openbmb/CAGUI"}], "name": "CAGUI", "publication": null, "published_at": null, "record_sha256": "b6583f313d35c4ef2715c8281c5020f5d925337d618abca18f4de885c947acf7", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/openbmb/CAGUI", "target_id": "arxiv:2506.01391", "type": "documented_by"}], "resource_id": "url:a06622da7688e3718a7b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/openbmb/CAGUI", "topics": ["data-generation"], "updated_at": null}, "text": "CAGUI\nResource type: dataset\nReview status: curated\nScope: core\n\nChinese Android GUI benchmark released with AgentCPM-GUI, with a grounding track for widgets and OCR text and an agent track for multi-step tasks in Chinese apps.\nPlatforms: android\nTopics: data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/openbmb/CAGUI/tree/main\", \"gated\": false, \"hub_id\": \"openbmb/CAGUI\", \"languages\": [\"zh\"], \"size_category_reported\": [\"1K<n<10K\"]}\n\nPrimary source: https://huggingface.co/datasets/openbmb/CAGUI\nCatalog record: https://awesomecua.github.io/resources/url-a06622da7688e3718a7b/"}
{"id": "url:a13a187bc058fbb4f6af", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa", "citation_url": "https://awesomecua.github.io/resources/url-a13a187bc058fbb4f6af/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md", "population": "train split", "qualifier": "exact", "unit": "step records", "value": 164533}], "evidence_urls": ["https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md", "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"], "formats": ["Parquet"], "modalities": ["text", "accessibility-tree observations", "actions"], "notes": "The card count is flattened step records, not trajectories or independent tasks. Both successful and unsuccessful trajectories are included according to the project README; no license is declared.", "splits": ["train"], "supervision": ["action completion", "trajectory outcome"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/tree/main", "gated": false, "hub_id": "apurvaga/go-browse-wa", "languages": [], "modalities": ["tabular", "text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "494f4ac9ab9e94831dcf768be001b7284e30e00704674358baf3bb2060344361", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/apurvaga/go-browse-wa"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "18ab64833acf31113a83b2463359f1c7e37c4aaabfe11c33db2ba3e7d32818c6", "supports": ["details"], "url": "https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["summary", "environments", "relations"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "494f4ac9ab9e94831dcf768be001b7284e30e00704674358baf3bb2060344361", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/apurvaga/go-browse-wa"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/apurvaga/go-browse-wa"}], "name": "Go-Browse-WA processed data", "publication": null, "published_at": null, "record_sha256": "e9693efd37d71faffaeacd25770bf55340db7a78440650b14cff61348d79734f", "relations": [{"evidence_url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "resource_id": "url:a13a187bc058fbb4f6af", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/apurvaga/go-browse-wa", "topics": ["web-navigation", "imitation-learning"], "updated_at": null}, "text": "Go-Browse-WA processed data\nResource type: dataset\nReview status: curated\nScope: core\n\nFlattened web-interaction training records with accessibility-tree observations, assistant completions, trajectory rewards and step-link metadata.\nPlatforms: web\nTopics: web-navigation, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md\", \"population\": \"train split\", \"qualifier\": \"exact\", \"unit\": \"step records\", \"value\": 164533}], \"evidence_urls\": [\"https://huggingface.co/datasets/apurvaga/go-browse-wa/raw/main/README.md\", \"https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md\"], \"formats\": [\"Parquet\"], \"modalities\": [\"text\", \"accessibility-tree observations\", \"actions\"], \"notes\": \"The card count is flattened step records, not trajectories or independent tasks. Both successful and unsuccessful trajectories are included according to the project README; no license is declared.\", \"splits\": [\"train\"], \"supervision\": [\"action completion\", \"trajectory outcome\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/apurvaga/go-browse-wa/tree/main\", \"gated\": false, \"hub_id\": \"apurvaga/go-browse-wa\", \"modalities\": [\"tabular\", \"text\"]}\n\nThe public release contains nine Parquet shards. The short card supplies schema and split metadata but no license declaration; the official README describes successful and unsuccessful trajectories processed for fine-tuning. The raw screenshot/HTML corpus is a separate release and is not claimed to be included here.\n\nPrimary source: https://huggingface.co/datasets/apurvaga/go-browse-wa\nCatalog record: https://awesomecua.github.io/resources/url-a13a187bc058fbb4f6af/"}
{"id": "url:a349aace3cfde25628a8", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Hcompany/Holo3-35B-A3B", "citation_url": "https://awesomecua.github.io/resources/url-a349aace3cfde25628a8/", "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/tree/main", "gated": false, "hub_id": "Hcompany/Holo3-35B-A3B"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:07:18Z", "role": "primary", "sha256": "bc5324dd6ceea9a0cc6bf0639b99d24cdd172af3c5d247bf923fdce1dc11f582", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo3-35B-A3B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:07:19Z", "role": "primary", "sha256": "0681f87b059110a4e5dff8955e9feca6c5ea118b326f729f9c0ed0fe49e42d03", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:07:19Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:07:19Z", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo3-35B-A3B"}], "name": "Holo3 35B-A3B", "publication": null, "published_at": null, "record_sha256": "52ece178867f991dc7c21a9a808f331f675121eebb0e2f9af998e684b27ab597", "relations": [], "resource_id": "url:a349aace3cfde25628a8", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Hcompany/Holo3-35B-A3B", "topics": ["grounding", "planning", "action-prediction"], "updated_at": null}, "text": "Holo3 35B-A3B\nResource type: model\nReview status: curated\nScope: core\n\nA sparse Qwen3.5-based GUI-action checkpoint released for navigation and computer-use workflows.\nPlatforms: web, desktop, cross-platform\nTopics: grounding, planning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Hcompany/Holo3-35B-A3B/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/Holo3-35B-A3B\"}\n\nFourteen public safetensors shards are present. The card declares Apache-2.0 for this release and describes a 35B-total/3B-active MoE architecture. No unpublished enterprise benchmark or training corpus is added solely from its performance claims.\n\nPrimary source: https://huggingface.co/Hcompany/Holo3-35B-A3B\nCatalog record: https://awesomecua.github.io/resources/url-a349aace3cfde25628a8/"}
{"id": "url:a4cdb2ee0d07d3980651", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "canonical_source_url": "https://doi.org/10.1145/1622176.1622213", "citation_url": "https://awesomecua.github.io/resources/url-a4cdb2ee0d07d3980651/", "details": {"publication": {"authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "date_basis": "Supplied month precision: 2009-10. Date Issued: October 2009", "date_parts": [2009, 10], "doi": "10.1145/1622176.1622213", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Institutional abstract and published-version citation supply the retained evidence; the institution identifies an author final manuscript record. Exact publication day is not supplied. Operating-system-specific experiments are not asserted.", "pages": "183-192", "pdf_url": null, "publisher": null, "series": null, "title": "Sikuli: Using GUI screenshots for search and automation", "type": "conference-paper", "url": "https://doi.org/10.1145/1622176.1622213", "venue": "Proceedings of the 22nd Annual ACM Symposium on User Interface Software and Technology (UIST '09)", "volume": null}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:21Z", "role": "primary", "sha256": "dbaaf21982522b34857be014b39d6c43fedcda20f808efe40d04958c77243a5a", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"}], "identifiers": {"doi": "10.1145/1622176.1622213"}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1622176.1622213"}], "name": "Sikuli: Using GUI screenshots for search and automation", "publication": {"authors": ["Tom Yeh", "Tsung-Hsiang Chang", "Robert C. Miller"], "date_basis": "Supplied month precision: 2009-10. Date Issued: October 2009", "date_parts": [2009, 10], "doi": "10.1145/1622176.1622213", "event_dates": null, "evidence_urls": ["https://dspace.mit.edu/entities/publication/604a3f79-451d-4186-a94e-d12690ffe40a"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "Institutional abstract and published-version citation supply the retained evidence; the institution identifies an author final manuscript record. Exact publication day is not supplied. Operating-system-specific experiments are not asserted.", "pages": "183-192", "pdf_url": null, "publisher": null, "series": null, "title": "Sikuli: Using GUI screenshots for search and automation", "type": "conference-paper", "url": "https://doi.org/10.1145/1622176.1622213", "venue": "Proceedings of the 22nd Annual ACM Symposium on User Interface Software and Technology (UIST '09)", "volume": null}, "published_at": null, "record_sha256": "b611a5aaff2aedd8225da5177aa66db9de402324f973e7e34ce2d55cf728e100", "relations": [], "resource_id": "url:a4cdb2ee0d07d3980651", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/1622176.1622213", "topics": ["grounding", "automation", "screen-understanding"], "updated_at": null}, "text": "Sikuli: Using GUI screenshots for search and automation\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces screenshot-based GUI search and visual scripting, using image patterns to locate interface targets and direct mouse and keyboard events. It evaluates screenshot queries and demonstrates automation and interactive-help tasks.\nPlatforms: unspecified\nTopics: grounding, automation, screen-understanding\nRecord publication date (recorded precision): 2009-10\nAuthors: Tom Yeh, Tsung-Hsiang Chang, Robert C. Miller\n\nPublication edition: Sikuli: Using GUI screenshots for search and automation\nPublication authors: Tom Yeh, Tsung-Hsiang Chang, Robert C. Miller\nPublication edition date (recorded precision): 2009-10\nPublication venue: Proceedings of the 22nd Annual ACM Symposium on User Interface Software and Technology (UIST '09)\nPublication pages: 183-192\nPublication doi: 10.1145/1622176.1622213\nPublication date basis: Supplied month precision: 2009-10. Date Issued: October 2009\nPublication notes: Institutional abstract and published-version citation supply the retained evidence; the institution identifies an author final manuscript record. Exact publication day is not supplied. Operating-system-specific experiments are not asserted.\nPublication source: https://doi.org/10.1145/1622176.1622213\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. MIT identifies the retained content as an author final manuscript record. Its PDF download returned HTTP 429; no retry was made. The complete institutional abstract and bibliographic record were read.\n\nPrimary source: https://doi.org/10.1145/1622176.1622213\nCatalog record: https://awesomecua.github.io/resources/url-a4cdb2ee0d07d3980651/"}
{"id": "url:a4da77ee63f879cc8e63", "metadata": {"aliases": ["url:4338ec0f1a67bdb6a139", "microsoft/Fara-7B", "https://huggingface.co/microsoft/Fara-7B", "url:d416331cdf9730afd6b9", "microsoft/fara-7b", "https://huggingface.co/microsoft/fara-7b"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/microsoft/Fara-7b", "citation_url": "https://awesomecua.github.io/resources/url-a4da77ee63f879cc8e63/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/Fara-7b/tree/main", "gated": false, "hub_id": "microsoft/Fara-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "92bfa3c1f34e7e76a35f72389a3b3dd9fdab5fcff0569c9fc8aabe331c703f07", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "92bfa3c1f34e7e76a35f72389a3b3dd9fdab5fcff0569c9fc8aabe331c703f07", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:59Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/Fara-7B"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:00Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/Fara-7b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:01Z", "role": "primary", "sha256": "cd1f83cd2ebf30d45501d1a31ecbe3bcb2039928d580c750733d2edb49916398", "supports": ["aliases"], "url": "https://huggingface.co/api/models/microsoft/fara-7b"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/Fara-7b"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/Fara-7b"}], "name": "Fara 7B", "publication": null, "published_at": null, "record_sha256": "649ad0b61297736004424c65afa21d9b8070ffdbca63902564a016e5862a3092", "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/Fara-7b", "target_id": "github:microsoft/fara", "type": "variant_of"}], "resource_id": "url:a4da77ee63f879cc8e63", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/microsoft/Fara-7b", "topics": ["planning", "web-navigation"], "updated_at": null}, "text": "Fara 7B\nResource type: model\nReview status: curated\nScope: core\n\nMicrosoft's Fara checkpoint, a Qwen2.5-VL-architecture image-text-to-text model with about 8.3B BF16 parameters, English, MIT-licensed and linked to arXiv paper 2511.19663.\nRecorded aliases: url:4338ec0f1a67bdb6a139, microsoft/Fara-7B, https://huggingface.co/microsoft/Fara-7B, url:d416331cdf9730afd6b9, microsoft/fara-7b, https://huggingface.co/microsoft/fara-7b\nPlatforms: web\nTopics: planning, web-navigation\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/microsoft/Fara-7b/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/Fara-7B\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/microsoft/Fara-7b\nCatalog record: https://awesomecua.github.io/resources/url-a4da77ee63f879cc8e63/"}
{"id": "url:a6ee20fea67438d91e06", "metadata": {"aliases": ["url:2f67f5ad7954632dfc4e", "THUDM/webrl-llama-3.1-8b", "https://huggingface.co/THUDM/webrl-llama-3.1-8b"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b", "citation_url": "https://awesomecua.github.io/resources/url-a6ee20fea67438d91e06/", "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/tree/main", "gated": false, "hub_id": "zai-org/webrl-llama-3.1-8b"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "4712446382f106a115a317d0ec6273a975545386b9cedf56946a96a56479dc9f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "6523b7f198d8debec2f3074190812955c6251590b61f8a6ea1bbb681b8ea128f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/webrl-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "6523b7f198d8debec2f3074190812955c6251590b61f8a6ea1bbb681b8ea128f", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/webrl-llama-3.1-8b"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "4712446382f106a115a317d0ec6273a975545386b9cedf56946a96a56479dc9f", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:06Z", "role": "primary", "sha256": "96621f060dddec91453d8c2b451b3793a34b34810d82c7cd13ce4cb05ca90a76", "supports": ["aliases"], "url": "https://huggingface.co/api/models/THUDM/webrl-llama-3.1-8b"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:06Z", "role": "primary", "sha256": "96621f060dddec91453d8c2b451b3793a34b34810d82c7cd13ce4cb05ca90a76", "supports": ["aliases"], "url": "https://huggingface.co/api/models/zai-org/webrl-llama-3.1-8b"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:05:40Z", "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b"}], "name": "WebRL Llama 3.1 8B", "publication": null, "published_at": null, "record_sha256": "418e41cb8b0a81057b4d3ef809a3d85ffb655127ac717e5a9c63cb6cbd2050df", "relations": [{"evidence_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b/raw/main/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "resource_id": "url:a6ee20fea67438d91e06", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zai-org/webrl-llama-3.1-8b", "topics": ["web-navigation", "reinforcement-learning"], "updated_at": null}, "text": "WebRL Llama 3.1 8B\nResource type: model\nReview status: curated\nScope: core\n\nA WebRL browser-policy checkpoint for performing actions in the five website environments of WebArena.\nRecorded aliases: url:2f67f5ad7954632dfc4e, THUDM/webrl-llama-3.1-8b, https://huggingface.co/THUDM/webrl-llama-3.1-8b\nPlatforms: web\nTopics: web-navigation, reinforcement-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/zai-org/webrl-llama-3.1-8b/tree/main\", \"gated\": false, \"hub_id\": \"zai-org/webrl-llama-3.1-8b\"}\n\nThe official Hub API resolves the former THUDM namespace to zai-org. Seven safetensors shards and configuration/tokenizer files are public and ungated. The card marks the weight license as other without naming complete terms, so no permissive weight license is inferred. The card’s evaluation prose also names GLM-4 despite the Llama checkpoint identity; no score is imported.\n\nPrimary source: https://huggingface.co/zai-org/webrl-llama-3.1-8b\nCatalog record: https://awesomecua.github.io/resources/url-a6ee20fea67438d91e06/"}
{"id": "url:aa28e1329a52e341d285", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset", "citation_url": "https://awesomecua.github.io/resources/url-aa28e1329a52e341d285/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/tree/main", "gated": false, "hub_id": "osunlp/GUI-Drag-dataset", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "907c8a5169c180491ef8c26c9428b2c67bc4e26f472d6423ce3715cbe47f76ef", "supports": ["notes"], "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "cdb8be8c4692a8ddcf8a66c1f0be666a9214718445925fb3c88615a0d641bf3a", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/GUI-Drag-dataset"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset"}], "name": "GUI-Drag", "publication": null, "published_at": null, "record_sha256": "1223ad4dc2d3ec760e75f1af4356b8d082336b7ad20ecc2ac7244b6e0524fc28", "relations": [], "resource_id": "url:aa28e1329a52e341d285", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/osunlp/GUI-Drag-dataset", "topics": ["grounding"], "updated_at": null}, "text": "GUI-Drag\nResource type: dataset\nReview status: curated\nScope: core\n\nData for learning text-dragging interactions in graphical interfaces.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/osunlp/GUI-Drag-dataset/tree/main\", \"gated\": false, \"hub_id\": \"osunlp/GUI-Drag-dataset\", \"modalities\": [\"image\"]}\n\nThe card links its paper (arXiv 2601.06031) and code repository (OSU-NLP-Group/GUI-Drag). Neither is a reviewed catalog record yet, so no relation is recorded.\n\nPrimary source: https://huggingface.co/datasets/osunlp/GUI-Drag-dataset\nCatalog record: https://awesomecua.github.io/resources/url-aa28e1329a52e341d285/"}
{"id": "url:aa693ca3e8948623f31e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.deeplearning.ai/courses/building-ai-browser-agents", "citation_url": "https://awesomecua.github.io/resources/url-aa693ca3e8948623f31e/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-20T12:51:44.325799+00:00", "role": "primary", "sha256": "1f0315e8758fe0edd97adb2c5253955dcdbe6d7bd8be393fe14dcd093483705f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes"], "url": "https://www.deeplearning.ai/courses/building-ai-browser-agents"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:41:10.403317+00:00", "role": "primary", "sha256": "0915d862638a308da6200d9603d668ab4d11bcabfa713ff0ccb82c7b21a5a92f", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://learn.deeplearning.ai/courses/building-ai-browser-agents/lesson/wwewm/building-a-simple-web-agent"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-20T13:42:56.029732+00:00", "role": "primary", "sha256": "20d17ca47d1d16826f8e2736d6639d0cf11e826ef28c788187e16adcac9d7561", "supports": ["summary", "environments", "topics", "scope", "notes"], "url": "https://learn.deeplearning.ai/courses/building-ai-browser-agents/lesson/lot5j/building-an-autonomous-web-agent"}], "identifiers": {}, "kind": "learning", "last_verified_at": "2026-10-20T12:51:44.325799+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.deeplearning.ai/courses/building-ai-browser-agents"}], "name": "Building AI Browser Agents", "publication": null, "published_at": null, "record_sha256": "110617b06b5c2a6962f8288a8b1d0adb6d26c026dfd17e57bc9db4eac88ca55f", "relations": [], "resource_id": "url:aa693ca3e8948623f31e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.deeplearning.ai/courses/building-ai-browser-agents", "topics": ["web-navigation", "planning"], "updated_at": null}, "text": "Building AI Browser Agents\nResource type: learning\nReview status: curated\nScope: core\n\nBrowser-agent course with worked lessons on structured web extraction and multi-step browser interaction, followed by AgentQ search and self-correction concepts.\nPlatforms: web\nTopics: web-navigation, planning\n\nReview covers the public syllabus and complete published transcripts of the simple-agent and autonomous-agent lessons, not the whole course or its video frames. The first uses Playwright and structured HTML extraction; the second uses a MultiOn session API and shows missed steps, continuations and clarifications. Hosted notebooks and helper files were not retrieved, and current service access was not tested. No account, enrollment, terms acceptance, installation or execution occurred. Course and example-code reuse terms remain unknown.\n\nPrimary source: https://www.deeplearning.ai/courses/building-ai-browser-agents\nCatalog record: https://awesomecua.github.io/resources/url-aa693ca3e8948623f31e/"}
{"id": "url:ab26f9ef56d651cee587", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/alibabagroup/CMGUI", "citation_url": "https://awesomecua.github.io/resources/url-ab26f9ef56d651cee587/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/alibabagroup/CMGUI/tree/main", "gated": false, "hub_id": "alibabagroup/CMGUI", "languages": ["zh"], "modalities": [], "size_category_reported": ["100K<n<1M"]}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "aa1e063e73269acd0cb06adcef67d6230f0fe9498ed361d92234053dcf035685", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/alibabagroup/CMGUI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:26Z", "role": "primary", "sha256": "6f43b9e5f2c783f91cc3271f51db6a9a521c29f21954a2a30ad0cfa186e1c1b7", "supports": ["relations"], "url": "https://huggingface.co/datasets/alibabagroup/CMGUI/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "aa1e063e73269acd0cb06adcef67d6230f0fe9498ed361d92234053dcf035685", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/alibabagroup/CMGUI"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/alibabagroup/CMGUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/alibabagroup/CMGUI"}], "name": "CMGUI", "publication": null, "published_at": null, "record_sha256": "71ec9c63e3c287b67a938784c42be3239d595a6c640d1f2dc9d443960cc6443e", "relations": [{"evidence_url": "https://huggingface.co/datasets/alibabagroup/CMGUI/raw/main/README.md", "target_id": "arxiv:2603.08533", "type": "documented_by"}], "resource_id": "url:ab26f9ef56d651cee587", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/alibabagroup/CMGUI", "topics": ["data-generation"], "updated_at": null}, "text": "CMGUI\nResource type: dataset\nReview status: curated\nScope: core\n\n18k trajectories with 98k steps from more than 50 Chinese mobile apps in domains such as e-commerce and social media, for building GUI agents on Chinese apps.\nPlatforms: android\nTopics: data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/alibabagroup/CMGUI/tree/main\", \"gated\": false, \"hub_id\": \"alibabagroup/CMGUI\", \"languages\": [\"zh\"], \"size_category_reported\": [\"100K<n<1M\"]}\n\nPrimary source: https://huggingface.co/datasets/alibabagroup/CMGUI\nCatalog record: https://awesomecua.github.io/resources/url-ab26f9ef56d651cee587/"}
{"id": "url:ac6d0f80765f5a32255a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench", "citation_url": "https://awesomecua.github.io/resources/url-ac6d0f80765f5a32255a/", "details": {"benchmark_protocol": {"actions": ["candidate-decision preference"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "population": "Hub test split", "qualifier": "exact", "unit": "rows", "value": 776}], "evaluation": "Compare reward-model preferences with the provided chosen/rejected decision labels; no independent score was reproduced.", "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"], "focus": ["web process-reward evaluation"], "notes": "The current Hub row count and project case-count claim use unresolved counting units. Candidate lists mean a row must not be assumed to equal one task or one pair.", "observations": ["page screenshots", "text observations", "action history", "task checklists"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebRewardBench"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "0968f60eea10b95f79cd16ad3746b0e0b1acaeb0e97c45856f709d53a14dd73a", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "db06c66b8cf5687008b0ad0d9ae269fcc27936887d238cfe2bdeb52151e1683e", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebRewardBench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench"}], "name": "WebRewardBench", "publication": null, "published_at": null, "record_sha256": "910b7e662b89b81667c0d2bb931ef4cf6e423d0df61f8cdc81f72cfc71f5317d", "relations": [{"evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md", "target_id": "arxiv:2505.15277", "type": "documented_by"}], "resource_id": "url:ac6d0f80765f5a32255a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench", "topics": ["web-navigation", "reward-models", "evaluation"], "updated_at": null}, "text": "WebRewardBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nAn offline preference-evaluation set for web process reward models, pairing page and trajectory context with preferred and rejected candidate decisions.\nPlatforms: web\nTopics: web-navigation, reward-models, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"candidate-decision preference\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md\", \"population\": \"Hub test split\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 776}], \"evaluation\": \"Compare reward-model preferences with the provided chosen/rejected decision labels; no independent score was reproduced.\", \"evidence_urls\": [\"https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/raw/main/README.md\", \"https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md\"], \"focus\": [\"web process-reward evaluation\"], \"notes\": \"The current Hub row count and project case-count claim use unresolved counting units. Candidate lists mean a row must not be assumed to equal one task or one pair.\", \"observations\": [\"page screenshots\", \"text observations\", \"action history\", \"task checklists\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench/tree/main\", \"gated\": false, \"hub_id\": \"LangAGI-Lab/WebRewardBench\"}\n\nTwo public Parquet shards are present. Card metadata reports 776 test rows, whereas the project describes 1,000-plus test cases without reconciling that count to rows. Both observations are retained as a grain/coverage caveat rather than treating either as a confirmed number of independent tasks. No data license is declared.\n\nPrimary source: https://huggingface.co/datasets/LangAGI-Lab/WebRewardBench\nCatalog record: https://awesomecua.github.io/resources/url-ac6d0f80765f5a32255a/"}
{"id": "url:ad7f176848fd84de64e2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/xlangai/OpenCUA-32B", "citation_url": "https://awesomecua.github.io/resources/url-ad7f176848fd84de64e2/", "details": {"disabled": false, "download_url": "https://huggingface.co/xlangai/OpenCUA-32B/tree/main", "gated": false, "hub_id": "xlangai/OpenCUA-32B"}, "environments": ["desktop", "linux", "windows", "macos"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a3be2e1a6cd39ee89809baa9504813c4a6a9a03bbd916e82924cc0f89cb5da33", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/xlangai/OpenCUA-32B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "151217eb96e4ab22a47c98766052e5cfd76497a2a9197cc125c0237f3fbc59f6", "supports": ["licenses.weights"], "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xlangai/OpenCUA-32B"}], "name": "OpenCUA 32B", "publication": null, "published_at": null, "record_sha256": "e631bc718f8c2ab5114e36ae6f87793f91691b4336cccb6e4e437ccbe78661a6", "relations": [{"evidence_url": "https://huggingface.co/xlangai/OpenCUA-32B/raw/main/README.md", "target_id": "arxiv:2508.09123", "type": "documented_by"}], "resource_id": "url:ad7f176848fd84de64e2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/xlangai/OpenCUA-32B", "topics": ["grounding", "planning", "self-reflection"], "updated_at": null}, "text": "OpenCUA 32B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-based computer-use checkpoint that combines screenshot history with planning and grounded executable actions.\nPlatforms: desktop, linux, windows, macos\nTopics: grounding, planning, self-reflection\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/xlangai/OpenCUA-32B/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/OpenCUA-32B\"}\n\nThe public manifest contains 64 safetensors shards. The card declares MIT for this model and names its Qwen2.5-VL base; upstream terms remain separate. This is the 32B release, distinct from the existing 72B preview. Family-level or 72B benchmark scores are not transferred to it.\n\nPrimary source: https://huggingface.co/xlangai/OpenCUA-32B\nCatalog record: https://awesomecua.github.io/resources/url-ad7f176848fd84de64e2/"}
{"id": "url:ae8f5208eea9e7f1f606", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/mllmTeam/DroidCall", "citation_url": "https://awesomecua.github.io/resources/url-ae8f5208eea9e7f1f606/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/tree/main", "gated": false, "hub_id": "mllmTeam/DroidCall", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"], "reason": "The Hub dataset card declares Apache-2.0 in its metadata header (license: apache-2.0). The companion repository's MIT code license is a separate declaration."}}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "08a6c5cfa655e1ecd774a41a9a815e845ad0afd47b289378bc1d6ee66c81dda9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "c7804987d56b03314a4c48e3ddf2e5819fc0d9c83cc65ba54a981a2f4a584dd9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mllmTeam/DroidCall"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:02:28Z", "url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mllmTeam/DroidCall"}], "name": "DroidCall data", "publication": null, "published_at": null, "record_sha256": "ca005493f9193e4aa46a185e95eeb1e2e305e7a59ac06c9664267a6c8908f1c2", "relations": [{"evidence_url": "https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md", "target_id": "arxiv:2412.00402", "type": "documented_by"}], "resource_id": "url:ae8f5208eea9e7f1f606", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://huggingface.co/datasets/mllmTeam/DroidCall", "topics": ["mobile-navigation", "tool-integration", "data-generation"], "updated_at": null}, "text": "DroidCall data\nResource type: dataset\nReview status: curated\nScope: adjacent\n\nInstruction-to-function-call examples for learning Android intent invocation, with training, testing and API-annotation files for mobile task execution.\nPlatforms: android\nTopics: mobile-navigation, tool-integration, data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/mllmTeam/DroidCall/tree/main\", \"gated\": false, \"hub_id\": \"mllmTeam/DroidCall\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/mllmTeam/DroidCall/raw/main/README.md\"], \"reason\": \"The Hub dataset card declares Apache-2.0 in its metadata header (license: apache-2.0). The companion repository's MIT code license is a separate declaration.\"}}}\n\nThe release provides JSONL data and API annotations. This is Android intent-calling supervision rather than screenshot-and-click trajectories, so it is marked adjacent. The dataset card declares Apache-2.0; the companion repository’s MIT code license is a separate declaration.\n\nPrimary source: https://huggingface.co/datasets/mllmTeam/DroidCall\nCatalog record: https://awesomecua.github.io/resources/url-ae8f5208eea9e7f1f606/"}
{"id": "url:aeb32973694f91ad9920", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/mPLUG/GUI-Owl-7B", "citation_url": "https://awesomecua.github.io/resources/url-aeb32973694f91ad9920/", "details": {"disabled": false, "download_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/tree/main", "gated": false, "hub_id": "mPLUG/GUI-Owl-7B"}, "environments": ["android", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "4822eab7fd35e583d52db7369d7d929d69dbde66b27c9da3d370cad9ec06ec89", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/mPLUG/GUI-Owl-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "a0e6dd76aee6f138c15f244f4e2e8ba650b8829295397fdbfa839a8edf852b57", "supports": ["licenses.weights"], "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:41Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/mPLUG/GUI-Owl-7B"}], "name": "GUI-Owl 7B", "publication": null, "published_at": null, "record_sha256": "e9a8ceca08a7a64a2a38ad981fd41d7418192de6c554f263dffc43cc2090c6a0", "relations": [{"evidence_url": "https://huggingface.co/mPLUG/GUI-Owl-7B/raw/main/README.md", "target_id": "arxiv:2508.15144", "type": "documented_by"}], "resource_id": "url:aeb32973694f91ad9920", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/mPLUG/GUI-Owl-7B", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null}, "text": "GUI-Owl 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-based GUI checkpoint released with Mobile-Agent-v3 for grounded actions and specialized roles in its agent framework.\nPlatforms: android, desktop\nTopics: grounding, action-prediction, planning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/mPLUG/GUI-Owl-7B/tree/main\", \"gated\": false, \"hub_id\": \"mPLUG/GUI-Owl-7B\"}\n\nFive public safetensors shards are present. MIT is declared in this model card. This record identifies the original 7B release; other sizes and later GUI-Owl generations are not aliases or assumed to share its exact terms. The online demo is still marked forthcoming.\n\nPrimary source: https://huggingface.co/mPLUG/GUI-Owl-7B\nCatalog record: https://awesomecua.github.io/resources/url-aeb32973694f91ad9920/"}
{"id": "url:b074ccdb2d9ec0277b65", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://openai.com/index/introducing-chatgpt-agent/", "citation_url": "https://awesomecua.github.io/resources/url-b074ccdb2d9ec0277b65/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web", "desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "fb3b4dbc435158c256abae36cf092172284c4d6fd53fa556499cb7b0730a5bc6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-chatgpt-agent/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-chatgpt-agent/"}], "name": "ChatGPT Agent", "publication": null, "published_at": null, "record_sha256": "bdc37c995b1cbea659cd7d198d60b1d121df95ea34e167db6199157c3a05ee04", "relations": [{"evidence_url": "https://openai.com/index/introducing-chatgpt-agent/", "target_id": "organization:openai", "type": "published_by"}], "resource_id": "url:b074ccdb2d9ec0277b65", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openai.com/index/introducing-chatgpt-agent/", "topics": ["workflow"], "updated_at": null}, "text": "ChatGPT Agent\nResource type: product\nReview status: curated\nScope: core\n\nHosted agent combining browser interaction, tools, and computer-based tasks.\nPlatforms: web, desktop\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nThe launch page identifies itself as outdated. This entry records the product introduction, not current availability, limits, or pricing.\n\nPrimary source: https://openai.com/index/introducing-chatgpt-agent/\nCatalog record: https://awesomecua.github.io/resources/url-b074ccdb2d9ec0277b65/"}
{"id": "url:b53e03a6e4ca4226d126", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench", "citation_url": "https://awesomecua.github.io/resources/url-b53e03a6e4ca4226d126/", "details": {"availability": "Ungated v1.0 JSON task-definition file.", "benchmark_protocol": {"actions": ["mobile GUI actions through a separately configured agent"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md", "population": "MobBench tasks used in the paper", "qualifier": "exact", "unit": "tasks", "value": 50}], "evaluation": "The publisher reports a fifty-task real-device evaluation; the inspected release does not establish an automatic correctness judge.", "evidence_urls": ["https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"], "focus": ["mobile task completion"], "notes": "Task definitions are distinct from device images, agent execution and automated scoring.", "observations": ["Android device interface"], "setting": "interactive"}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"], "reason": "Dataset-card declaration for the MobBench task-definition release only."}}, "release_revision": "28941d10c49a16efb991d4157e512e14204bf3d6"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "231a8f483256c774840c99383ab6536520a0aed0f04d62b4d64f1f59477b906e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol", "licenses.data"], "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:32:07Z", "role": "primary", "sha256": "4e7225f956d7cee98fe671e7bfd91fc407d87dd805d8f0c11c3b0c839714ccf3", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/OpenDFM/MobA-MobBench"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T12:32:07Z", "licenses": {"code": null, "data": {"basis": "Dataset-card declaration for the MobBench task-definition release only.", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-17T12:32:07Z", "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench"}], "name": "MobBench task definitions", "publication": null, "published_at": null, "record_sha256": "00580f628852bc12d94aa4519c70a88f9a37554a8ffb71db49958f8075be1db2", "relations": [{"evidence_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md", "target_id": "arxiv:2410.13757", "type": "documented_by"}], "resource_id": "url:b53e03a6e4ca4226d126", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OpenDFM/MobA-MobBench", "topics": ["evaluation", "mobile-navigation"], "updated_at": null}, "text": "MobBench task definitions\nResource type: benchmark\nReview status: curated\nScope: core\n\nA public set of real-device mobile task definitions used to evaluate MobA, with Chinese and English task material packaged in a versioned JSON release.\nPlatforms: android\nTopics: evaluation, mobile-navigation\n\nRecorded specifications: {\"availability\": \"Ungated v1.0 JSON task-definition file.\", \"benchmark_protocol\": {\"actions\": [\"mobile GUI actions through a separately configured agent\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md\", \"population\": \"MobBench tasks used in the paper\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 50}], \"evaluation\": \"The publisher reports a fifty-task real-device evaluation; the inspected release does not establish an automatic correctness judge.\", \"evidence_urls\": [\"https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md\"], \"focus\": [\"mobile task completion\"], \"notes\": \"Task definitions are distinct from device images, agent execution and automated scoring.\", \"observations\": [\"Android device interface\"], \"setting\": \"interactive\"}, \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/OpenDFM/MobA-MobBench/raw/main/README.md\"], \"reason\": \"Dataset-card declaration for the MobBench task-definition release only.\"}}, \"release_revision\": \"28941d10c49a16efb991d4157e512e14204bf3d6\"}\n\nThe ungated manifest lists MobBench_tasks_release_v1.0.json and bilingual cards. The card reports fifty tasks; no sample records were downloaded to recount them. This release supplies task definitions and agent setup guidance, not an independently verified automatic judge or a bundled Android environment. The data card declares CC-BY-NC-SA-4.0; code and model licenses are separate.\n\nPrimary source: https://huggingface.co/datasets/OpenDFM/MobA-MobBench\nCatalog record: https://awesomecua.github.io/resources/url-b53e03a6e4ca4226d126/"}
{"id": "url:b560b5d6abaab98483f6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Biplab Deka", "Zifeng Huang", "Chad Franzen", "Joshua Hibschman", "Daniel Afergan", "Yang Li", "Jeffrey Nichols", "Ranjitha Kumar"], "canonical_source_url": "https://doi.org/10.1145/3126594.3126651", "citation_url": "https://awesomecua.github.io/resources/url-b560b5d6abaab98483f6/", "details": {"publication": {"authors": ["Biplab Deka", "Zifeng Huang", "Chad Franzen", "Joshua Hibschman", "Daniel Afergan", "Yang Li", "Jeffrey Nichols", "Ranjitha Kumar"], "date_basis": "Supplied day precision: 2017-10-20. PDF cover, Published: 20 October 2017 The 18 January 2026 download date and October 22-25 conference dates are not publication dates.", "date_parts": [2017, 10, 20], "doi": "10.1145/3126594.3126651", "event_dates": null, "evidence_urls": ["https://danafergan.com/publications/deka2017rico.pdf", "https://dev.interactionmining.org/archive/rico"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim. No volume number is supplied for this proceedings article.", "pages": "845-854", "pdf_url": "https://danafergan.com/publications/deka2017rico.pdf", "publisher": null, "series": null, "title": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications", "type": "conference-paper", "url": "https://doi.org/10.1145/3126594.3126651", "venue": "UIST '17: Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "fb00401fe531ca3aed078dccc57147bcaad33207dacf7cd4cfd4af14f1487ff5", "supports": ["details.publication.notes"], "url": "https://dev.interactionmining.org/archive/rico"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T10:01:32Z", "role": "primary", "sha256": "a1ec58c526b152c052c5c72f9a62725f1026b5f0e4f45c8fe85fce284295da84", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://danafergan.com/publications/deka2017rico.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "fb00401fe531ca3aed078dccc57147bcaad33207dacf7cd4cfd4af14f1487ff5", "supports": ["details.publication.notes"], "url": "https://dev.interactionmining.org/archive/rico"}], "identifiers": {"doi": "10.1145/3126594.3126651"}, "kind": "paper", "last_verified_at": "2026-10-17T10:01:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/3126594.3126651"}, {"rel": "publication_pdf", "url": "https://danafergan.com/publications/deka2017rico.pdf"}], "name": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications", "publication": {"authors": ["Biplab Deka", "Zifeng Huang", "Chad Franzen", "Joshua Hibschman", "Daniel Afergan", "Yang Li", "Jeffrey Nichols", "Ranjitha Kumar"], "date_basis": "Supplied day precision: 2017-10-20. PDF cover, Published: 20 October 2017 The 18 January 2026 download date and October 22-25 conference dates are not publication dates.", "date_parts": [2017, 10, 20], "doi": "10.1145/3126594.3126651", "event_dates": null, "evidence_urls": ["https://danafergan.com/publications/deka2017rico.pdf", "https://dev.interactionmining.org/archive/rico"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim. No volume number is supplied for this proceedings article.", "pages": "845-854", "pdf_url": "https://danafergan.com/publications/deka2017rico.pdf", "publisher": null, "series": null, "title": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications", "type": "conference-paper", "url": "https://doi.org/10.1145/3126594.3126651", "venue": "UIST '17: Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology", "volume": null}, "published_at": "2017-10-20", "record_sha256": "53785664754fa029cb7086983828b7613148a76964318f17c342e307bc7417fe", "relations": [], "resource_id": "url:b560b5d6abaab98483f6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/3126594.3126651", "topics": ["data-generation", "screen-understanding", "ui-design"], "updated_at": null}, "text": "Rico: A Mobile App Dataset for Building Data-Driven Design Applications\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Rico, collecting Android interface designs and interaction traces through crowdsourced use followed by automated exploration. The paper demonstrates UI-layout retrieval using learned representations and describes visual, textual, structural, and interaction data.\nPlatforms: android\nTopics: data-generation, screen-understanding, ui-design\nPublished: 2017-10-20\nAuthors: Biplab Deka, Zifeng Huang, Chad Franzen, Joshua Hibschman, Daniel Afergan, Yang Li, Jeffrey Nichols, Ranjitha Kumar\n\nPublication edition: Rico: A Mobile App Dataset for Building Data-Driven Design Applications\nPublication authors: Biplab Deka, Zifeng Huang, Chad Franzen, Joshua Hibschman, Daniel Afergan, Yang Li, Jeffrey Nichols, Ranjitha Kumar\nPublication edition date (recorded precision): 2017-10-20\nPublication venue: UIST '17: Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology\nPublication pages: 845-854\nPublication doi: 10.1145/3126594.3126651\nPublication date basis: Supplied day precision: 2017-10-20. PDF cover, Published: 20 October 2017 The 18 January 2026 download date and October 22-25 conference dates are not publication dates.\nPublication notes: The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim. No volume number is supplied for this proceedings article.\nPublication source: https://doi.org/10.1145/3126594.3126651\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The paper is a distinct resource from the already cataloged Rico dataset (url:0b134b082a7bd507fc6a). The publisher PDF describes 9.7k apps and 72k screens; the project page describes 9.3k apps and 66k screens. The annotation intentionally supplies no merged size claim.\n\nPrimary source: https://doi.org/10.1145/3126594.3126651\nCatalog record: https://awesomecua.github.io/resources/url-b560b5d6abaab98483f6/"}
{"id": "url:b5e2127e07f692d516b9", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2", "citation_url": "https://awesomecua.github.io/resources/url-b5e2127e07f692d516b9/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md", "https://huggingface.co/api/datasets/xlangai/aguvis-stage2", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md", "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage2.yaml"], "formats": ["JSON annotations", "ZIP image packages", "multipart TAR image packages"], "modalities": ["GUI screenshots", "natural-language instructions", "action trajectories"], "notes": "Stage 2 trajectory release, Hub revision bd6f1ed4, is distinct from Stage 1 grounding. The public manifest lists source-specific JSON levels and image ZIP/multipart TAR packages. The training recipe uses all rows for some inputs and random:33% for others; several l2 paths occur twice while the manifest also contains l3 files. Recipe rows/levels are neither independent datasets nor disjoint train/test splits. No example/action/trajectory count is inferred. Apache-2.0 is the card declaration, with separate underlying source terms.", "splits": ["training"], "supervision": ["GUI trajectory supervised fine-tuning", "planning and reasoning with inner monologue"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/tree/main", "gated": false, "hub_id": "xlangai/aguvis-stage2"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "550293f00905615e19e63ac896818f9875dad38fea0ffe196c6302a29719aa72", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "f030264a7a8c7485dc7cf15ca554d84450cc492d2ac90e5965810a0207ef370b", "supports": ["identity", "release_status", "details", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xlangai/aguvis-stage2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "2dd9f36b818c1b7b6bd772ba8575f6c0ebab6d9e0d307edcc2f196f3e4ba502e", "supports": ["environments", "summary", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:11Z", "role": "primary", "sha256": "8c1b39e9d16067018ce845fd4e996f34e2b47346c342b2ba26b030f4a5338b66", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage2.yaml"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "550293f00905615e19e63ac896818f9875dad38fea0ffe196c6302a29719aa72", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/aguvis-stage2"}], "name": "Aguvis stage 2 trajectory data", "publication": null, "published_at": null, "record_sha256": "d02b88a242f9cc58d774d5273c466915563a72b6a6df1bdfd8179bcbedf7cd19", "relations": [{"evidence_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md", "target_id": "arxiv:2412.04454", "type": "documented_by"}], "resource_id": "url:b5e2127e07f692d516b9", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xlangai/aguvis-stage2", "topics": ["planning", "data-generation"], "updated_at": null}, "text": "Aguvis stage 2 trajectory data\nResource type: dataset\nReview status: curated\nScope: core\n\nThe second-stage Aguvis collection of interface trajectories for learning action sequences, planning and reasoning after grounding pretraining.\nPlatforms: cross-platform\nTopics: planning, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/xlangai/aguvis-stage2/raw/main/README.md\", \"https://huggingface.co/api/datasets/xlangai/aguvis-stage2\", \"https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/README.md\", \"https://raw.githubusercontent.com/xlang-ai/aguvis/HEAD/data/stage2.yaml\"], \"formats\": [\"JSON annotations\", \"ZIP image packages\", \"multipart TAR image packages\"], \"modalities\": [\"GUI screenshots\", \"natural-language instructions\", \"action trajectories\"], \"notes\": \"Stage 2 trajectory release, Hub revision bd6f1ed4, is distinct from Stage 1 grounding. The public manifest lists source-specific JSON levels and image ZIP/multipart TAR packages. The training recipe uses all rows for some inputs and random:33% for others; several l2 paths occur twice while the manifest also contains l3 files. Recipe rows/levels are neither independent datasets nor disjoint train/test splits. No example/action/trajectory count is inferred. Apache-2.0 is the card declaration, with separate underlying source terms.\", \"splits\": [\"training\"], \"supervision\": [\"GUI trajectory supervised fine-tuning\", \"planning and reasoning with inner monologue\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xlangai/aguvis-stage2/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/aguvis-stage2\"}\n\nThe manifest contains trajectory annotations and archives associated with mobile, desktop and web sources. The card declares Apache-2.0; underlying source content may retain separate terms. This stage has a different training purpose and files from the grounding release. No archive or training sample was downloaded.\n\nPrimary source: https://huggingface.co/datasets/xlangai/aguvis-stage2\nCatalog record: https://awesomecua.github.io/resources/url-b5e2127e07f692d516b9/"}
{"id": "url:b80bfa0a4939a493a11f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xwk123/Mobile3M", "citation_url": "https://awesomecua.github.io/resources/url-b80bfa0a4939a493a11f/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md", "https://huggingface.co/api/datasets/xwk123/Mobile3M", "https://raw.githubusercontent.com/XiaoMi/mobilevlm/HEAD/README.md"], "formats": ["7z app-graph archives", "multipart 7z files"], "modalities": [], "notes": "Official README calls this a partial training release and says app data will be uploaded gradually. Its training-JSON and test-data links point to separate Hub repositories; those artifacts are not counted as files or splits of xwk123/Mobile3M. Public revision 3df26694 lists app-graph archives; their names do not establish a three-million-example downloadable inventory. Precise inner record/modalities schema is not inspected. The card YAML says Apache-2.0 while its prose and official README assign CC-BY-NC-SA-4.0 to data and Apache to code; the existing data-license value remains unknown with the conflict documented.", "splits": ["partial training release"], "supervision": ["mobile UI pretraining for MobileVLM"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xwk123/Mobile3M/tree/main", "gated": false, "hub_id": "xwk123/Mobile3M", "languages": [], "license_review": {"data": {"evidence_urls": ["https://huggingface.co/api/datasets/xwk123/Mobile3M"], "reason": "The retained first-party API has Apache-2.0 card metadata but its publisher description explicitly assigns Apache to source code and CC-BY-NC-SA-4.0 to the dataset. A single permissive dataset declaration is unresolved."}}, "modalities": ["image"], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:01Z", "role": "primary", "sha256": "8a097dbc7384d3743f262dc0164c176ed577bf720cbbc92ff8b3cd820d1aed4f", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:04Z", "role": "primary", "sha256": "e8a92a783b6b65cacedf105e5ede583668f8774f2a81a2005dc6897da6fa88ef", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/XiaoMi/mobilevlm/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "356a8dca28fa3dc46bc05b13eaf04a3db273e63c06be4fd15d0e87e0343c0824", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xwk123/Mobile3M"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xwk123/Mobile3M"}], "name": "Mobile3M", "publication": null, "published_at": null, "record_sha256": "00583d636ad1248e5136d0bf883bde6a344441c6c2c3f738ec8f6d388efbc63e", "relations": [{"evidence_url": "https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md", "target_id": "arxiv:2409.14818", "type": "documented_by"}], "resource_id": "url:b80bfa0a4939a493a11f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xwk123/Mobile3M", "topics": ["screen-understanding"], "updated_at": null}, "text": "Mobile3M\nResource type: dataset\nReview status: curated\nScope: core\n\nPre-training corpus for Xiaomi's MobileVLM, distributed as per-app graph archives for mobile apps such as Baidu Map, QQ Music, Ctrip and Vipshop.\nPlatforms: android\nTopics: screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/xwk123/Mobile3M/raw/3df26694ff9186604eb44a8e633894d3ece67ed9/README.md\", \"https://huggingface.co/api/datasets/xwk123/Mobile3M\", \"https://raw.githubusercontent.com/XiaoMi/mobilevlm/HEAD/README.md\"], \"formats\": [\"7z app-graph archives\", \"multipart 7z files\"], \"modalities\": [], \"notes\": \"Official README calls this a partial training release and says app data will be uploaded gradually. Its training-JSON and test-data links point to separate Hub repositories; those artifacts are not counted as files or splits of xwk123/Mobile3M. Public revision 3df26694 lists app-graph archives; their names do not establish a three-million-example downloadable inventory. Precise inner record/modalities schema is not inspected. The card YAML says Apache-2.0 while its prose and official README assign CC-BY-NC-SA-4.0 to data and Apache to code; the existing data-license value remains unknown with the conflict documented.\", \"splits\": [\"partial training release\"], \"supervision\": [\"mobile UI pretraining for MobileVLM\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xwk123/Mobile3M/tree/main\", \"gated\": false, \"hub_id\": \"xwk123/Mobile3M\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/api/datasets/xwk123/Mobile3M\"], \"reason\": \"The retained first-party API has Apache-2.0 card metadata but its publisher description explicitly assigns Apache to source code and CC-BY-NC-SA-4.0 to the dataset. A single permissive dataset declaration is unresolved.\"}}, \"modalities\": [\"image\"]}\n\nPrimary source: https://huggingface.co/datasets/xwk123/Mobile3M\nCatalog record: https://awesomecua.github.io/resources/url-b80bfa0a4939a493a11f/"}
{"id": "url:b854ff53e497da3a5ad8", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Hcompany/DragOn", "citation_url": "https://awesomecua.github.io/resources/url-b854ff53e497da3a5ad8/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Hcompany/DragOn/tree/main", "gated": false, "hub_id": "Hcompany/DragOn", "languages": ["en", "fr"], "modalities": [], "size_category_reported": ["1M<n<10M"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T16:10:54Z", "role": "primary", "sha256": "e7ef94cafcfd155961beec373f2f4e727715b7d517a819f52b5407fc2dec4009", "supports": ["notes"], "url": "https://huggingface.co/datasets/Hcompany/DragOn/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "08b5529ce01781143147d56e2278aec2e5bdd972ba7e2fae6ba7e65e62688d40", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Hcompany/DragOn"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Hcompany/DragOn"}], "name": "DragOn", "publication": null, "published_at": null, "record_sha256": "b14bb3be107df632ed073c62afd259306aed058e042c48a6b2ef8b6374da90b5", "relations": [], "resource_id": "url:b854ff53e497da3a5ad8", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Hcompany/DragOn", "topics": ["grounding"], "updated_at": null}, "text": "DragOn\nResource type: dataset\nReview status: curated\nScope: core\n\nDrag-grounding training and evaluation data pairing each screenshot instruction with start and end boxes, across text highlighting, spreadsheet cells, slide resizing and sliders.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/Hcompany/DragOn/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/DragOn\", \"languages\": [\"en\", \"fr\"], \"size_category_reported\": [\"1M<n<10M\"]}\n\nThe card cites a workshop paper titled \"DragOn: A Drag-Grounding Benchmark and Training Dataset for GUI Agents\" without an arXiv link. arXiv 2606.06322 has a different title, so no paper relation is recorded.\n\nPrimary source: https://huggingface.co/datasets/Hcompany/DragOn\nCatalog record: https://awesomecua.github.io/resources/url-b854ff53e497da3a5ad8/"}
{"id": "url:b940a884f200a916a61b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://steel.dev/", "citation_url": "https://awesomecua.github.io/resources/url-b940a884f200a916a61b/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "5eacfa15c116ddeb02f6fd731fbb2e9856a933d743870fe41d947ef3b01aeb6f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://steel.dev/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://steel.dev/"}], "name": "Steel Cloud", "publication": null, "published_at": null, "record_sha256": "c6ff52c3e8c194530ef7d234282d99ab35d5a86c6b7bdc990b7557b0be9377f5", "relations": [{"evidence_url": "https://steel.dev/", "target_id": "organization:steel", "type": "published_by"}], "resource_id": "url:b940a884f200a916a61b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://steel.dev/", "topics": ["sandbox"], "updated_at": null}, "text": "Steel Cloud\nResource type: product\nReview status: curated\nScope: core\n\nManaged browser infrastructure associated with Steel Browser.\nPlatforms: web\nTopics: sandbox\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://steel.dev/\nCatalog record: https://awesomecua.github.io/resources/url-b940a884f200a916a61b/"}
{"id": "url:b94e285c3f757e2d0f3a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/KDEGroup/UI-AGILE", "citation_url": "https://awesomecua.github.io/resources/url-b94e285c3f757e2d0f3a/", "details": {"disabled": false, "download_url": "https://huggingface.co/KDEGroup/UI-AGILE/tree/main", "gated": false, "hub_id": "KDEGroup/UI-AGILE"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "a52032bb185885d47466f0bea26c15ccefeb8d6ff89f3bacc5eb4192c35d45f7", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/KDEGroup/UI-AGILE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:43:19Z", "role": "primary", "sha256": "ce35413bc1ef28338bc6fd9f132fa271a4f5e0bf23cecd7346f54adf0a964fd7", "supports": ["environments"], "url": "https://huggingface.co/datasets/KDEGroup/UI-AGILE-Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:30Z", "role": "primary", "sha256": "75aed1ebff7dd54ab1dc0112fe524a118d9e8845c2e079080d2bb7954f7ab36e", "supports": ["licenses.weights"], "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:30Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:44:30Z", "url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/KDEGroup/UI-AGILE"}], "name": "UI-AGILE checkpoint", "publication": null, "published_at": null, "record_sha256": "24c280138a68e8ee1ede49afda59a7ef7cfe3b747aca56278d88d729e0fb2e9c", "relations": [{"evidence_url": "https://huggingface.co/KDEGroup/UI-AGILE/raw/main/README.md", "target_id": "arxiv:2507.22025", "type": "documented_by"}], "resource_id": "url:b94e285c3f757e2d0f3a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/KDEGroup/UI-AGILE", "topics": ["grounding", "reinforcement-learning", "efficiency"], "updated_at": null}, "text": "UI-AGILE checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nThe released UI-AGILE model for GUI grounding with reward shaping, crop-based resampling and decomposed selection during inference.\nPlatforms: android\nTopics: grounding, reinforcement-learning, efficiency\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/KDEGroup/UI-AGILE/tree/main\", \"gated\": false, \"hub_id\": \"KDEGroup/UI-AGILE\"}\n\nFour safetensors shards and model/tokenizer files are public. The card declares MIT for these weights; the companion dataset has no declared license and is separate. The short card does not state a parameter count, so this entry does not infer one from shard sizes.\n\nPrimary source: https://huggingface.co/KDEGroup/UI-AGILE\nCatalog record: https://awesomecua.github.io/resources/url-b94e285c3f757e2d0f3a/"}
{"id": "url:ba2c664c1b3c6f1b172f", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "canonical_source_url": "https://ieeexplore.ieee.org/document/8506506", "citation_url": "https://awesomecua.github.io/resources/url-ba2c664c1b3c6f1b172f/", "details": {"publication": {"authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "date_basis": "Supplied year precision: 2018. APPINITE publication citation: VL/HCC 2018", "date_parts": [2018], "doi": null, "event_dates": null, "evidence_urls": ["https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/publications/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations. Author-hosted manuscript and institutional publication lists supply the retained evidence. DOI, final page range, and exact publication day remain unknown.", "pages": null, "pdf_url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "publisher": null, "series": null, "title": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "type": "conference-paper", "url": "https://ieeexplore.ieee.org/document/8506506", "venue": "Proceedings of the 2018 IEEE Symposium on Visual Languages and Human-Centric Computing", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:54:46Z", "role": "primary", "sha256": "6e05fd08b187adba32148e0835ef6b9e20cd6a3ba5e0de6e2778eb18c02788ee", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.pdf_url", "details.publication.title", "environments", "name", "scope", "summary", "topics"], "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:22Z", "role": "primary", "sha256": "32ee05ff1a91ed4a22d4bc156a462872aa4e28f5c79abfbce5dbf3655f412033", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.notes", "details.publication.type", "details.publication.venue", "kind"], "url": "https://www.cs.cmu.edu/~NatProg/sugilite.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:20Z", "role": "primary", "sha256": "adfcec73830b612b181d4846ace3fbaab1768e906715c1d90fe657e23fe4d985", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://toby.li/publications/"}], "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:54:46Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://ieeexplore.ieee.org/document/8506506"}, {"rel": "publication_pdf", "url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf"}], "name": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "publication": {"authors": ["Toby Jia-Jun Li", "Igor Labutov", "Xiaohan Nancy Li", "Xiaoyi Zhang", "Wenze Shi", "Wanling Ding", "Tom M. Mitchell", "Brad A. Myers"], "date_basis": "Supplied year precision: 2018. APPINITE publication citation: VL/HCC 2018", "date_parts": [2018], "doi": null, "event_dates": null, "evidence_urls": ["https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "https://www.cs.cmu.edu/~NatProg/sugilite.html", "https://toby.li/publications/"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations. Author-hosted manuscript and institutional publication lists supply the retained evidence. DOI, final page range, and exact publication day remain unknown.", "pages": null, "pdf_url": "https://toby.li/files/TobyLi-VLHCC18-APPINITE.pdf", "publisher": null, "series": null, "title": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions", "type": "conference-paper", "url": "https://ieeexplore.ieee.org/document/8506506", "venue": "Proceedings of the 2018 IEEE Symposium on Visual Languages and Human-Centric Computing", "volume": null}, "published_at": null, "record_sha256": "8a550e9a45778b7b359277cd569934dc295899089a4374693848405a9422a24f", "relations": [], "resource_id": "url:ba2c664c1b3c6f1b172f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://ieeexplore.ieee.org/document/8506506", "topics": ["grounding", "human-in-the-loop", "mobile-navigation"], "updated_at": null}, "text": "APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces APPINITE, a conversational interface for specifying the intended targets of demonstrated Android actions. It constructs semantic UI representations and uses an on-screen overlay to preview and refine executable data descriptions.\nPlatforms: android\nTopics: grounding, human-in-the-loop, mobile-navigation\nRecord publication date (recorded precision): 2018\nAuthors: Toby Jia-Jun Li, Igor Labutov, Xiaohan Nancy Li, Xiaoyi Zhang, Wenze Shi, Wanling Ding, Tom M. Mitchell, Brad A. Myers\n\nPublication edition: APPINITE: A Multi-Modal Interface for Specifying Data Descriptions in Programming by Demonstration Using Natural Language Instructions\nPublication authors: Toby Jia-Jun Li, Igor Labutov, Xiaohan Nancy Li, Xiaoyi Zhang, Wenze Shi, Wanling Ding, Tom M. Mitchell, Brad A. Myers\nPublication edition date (recorded precision): 2018\nPublication venue: Proceedings of the 2018 IEEE Symposium on Visual Languages and Human-Centric Computing\nPublication date basis: Supplied year precision: 2018. APPINITE publication citation: VL/HCC 2018\nPublication notes: The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations. Author-hosted manuscript and institutional publication lists supply the retained evidence. DOI, final page range, and exact publication day remain unknown.\nPublication source: https://ieeexplore.ieee.org/document/8506506\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The base SUGILITE framework and later PUMICE work are distinct publications; this paper focuses on resolving data-description ambiguity in demonstrations.\n\nPrimary source: https://ieeexplore.ieee.org/document/8506506\nCatalog record: https://awesomecua.github.io/resources/url-ba2c664c1b3c6f1b172f/"}
{"id": "url:ba6453cfdae0eb049ced", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data", "citation_url": "https://awesomecua.github.io/resources/url-ba6453cfdae0eb049ced/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md", "https://huggingface.co/api/datasets/xTRam1/plan-and-act-data"], "formats": ["JSON", "JSONL"], "modalities": ["text", "plans", "actions"], "notes": "The release contains planner and actor files for two web environments. No sample count is inferred from the package-file manifest. Card loading examples have repository/path inconsistencies.", "splits": ["train (QwQ-32B files)", "validation (QwQ-32B files)"], "supervision": ["high-level plans", "action execution"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/tree/main", "gated": false, "hub_id": "xTRam1/plan-and-act-data"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "2ec8f4a03b8331bdd2ba8d6f6800d6dea9d955cfd9e7c1d2168ecb64deb25448", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:28Z", "role": "primary", "sha256": "a98f98b7d4ceb81b00f217983ff271257f2f1e7d2c1a09836b872b2350d79ae9", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xTRam1/plan-and-act-data"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "2ec8f4a03b8331bdd2ba8d6f6800d6dea9d955cfd9e7c1d2168ecb64deb25448", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:02:29Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:02:29Z", "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data"}], "name": "Plan-and-Act training data", "publication": null, "published_at": null, "record_sha256": "92e9c55389eceb8af181911816c2c37af5bc4f20ce1a6e259a3b6a0389a92b6b", "relations": [{"evidence_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md", "target_id": "arxiv:2503.09572", "type": "documented_by"}], "resource_id": "url:ba6453cfdae0eb049ced", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xTRam1/plan-and-act-data", "topics": ["planning", "web-navigation", "data-generation"], "updated_at": null}, "text": "Plan-and-Act training data\nResource type: dataset\nReview status: curated\nScope: core\n\nPlanner and action-executor training examples for WebArena-Lite and WebVoyager, with model-specific processed files and raw trajectory annotations.\nPlatforms: web\nTopics: planning, web-navigation, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/xTRam1/plan-and-act-data/raw/main/README.md\", \"https://huggingface.co/api/datasets/xTRam1/plan-and-act-data\"], \"formats\": [\"JSON\", \"JSONL\"], \"modalities\": [\"text\", \"plans\", \"actions\"], \"notes\": \"The release contains planner and actor files for two web environments. No sample count is inferred from the package-file manifest. Card loading examples have repository/path inconsistencies.\", \"splits\": [\"train (QwQ-32B files)\", \"validation (QwQ-32B files)\"], \"supervision\": [\"high-level plans\", \"action execution\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xTRam1/plan-and-act-data/tree/main\", \"gated\": false, \"hub_id\": \"xTRam1/plan-and-act-data\"}\n\nThe public manifest exposes JSON/JSONL data for planner and actor roles, including train and validation variants. The card declares MIT. Its loading examples use a different Hub repository identifier and omit the data/ prefix for one file; use the inspected xTRam1 manifest rather than copying those examples verbatim. Third-party site content retains its own terms.\n\nPrimary source: https://huggingface.co/datasets/xTRam1/plan-and-act-data\nCatalog record: https://awesomecua.github.io/resources/url-ba6453cfdae0eb049ced/"}
{"id": "url:bc3deaf8eeecf569e73b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/GD-ML/Code2World", "citation_url": "https://awesomecua.github.io/resources/url-bc3deaf8eeecf569e73b/", "details": {"disabled": false, "download_url": "https://huggingface.co/GD-ML/Code2World/tree/3e48194a8228570164ecd510f9d83ac839e9e931", "gated": false, "hub_id": "GD-ML/Code2World", "hub_revision": "3e48194a8228570164ecd510f9d83ac839e9e931", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/AMAP-ML/Code2World"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:15Z", "role": "primary", "sha256": "befc89b09d099b86bda891141bca32b4a13b650c64d9b74ca18a9431e76f6c06", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "4c91d95f7533cb5550447c95bfb12d6064fd510ac01cba45ebb8d7fca305fdc9", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/GD-ML/Code2World"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "26c14a14166d3b747454d1924ecdf6d96652ac857ae6ad488630dc843036a777", "supports": ["environments", "details.repository_url"], "url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:15Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:15Z", "url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/GD-ML/Code2World"}], "name": "Code2World-8B checkpoint", "publication": null, "published_at": null, "record_sha256": "a26a1bdcabb921bf9bc9c4e542be3608aa1e244fa27ae40b8e1cd8bc5ca91b64", "relations": [{"evidence_url": "https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md", "target_id": "arxiv:2602.09856", "type": "documented_by"}], "resource_id": "url:bc3deaf8eeecf569e73b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/GD-ML/Code2World", "topics": ["world-models", "screen-understanding"], "updated_at": null}, "text": "Code2World-8B checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nReleased GUI world-model checkpoint that turns a screenshot and proposed action into HTML for rendering a predicted next screen.\nPlatforms: android\nTopics: world-models, screen-understanding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/GD-ML/Code2World/tree/3e48194a8228570164ecd510f9d83ac839e9e931\", \"gated\": false, \"hub_id\": \"GD-ML/Code2World\", \"hub_revision\": \"3e48194a8228570164ecd510f9d83ac839e9e931\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/GD-ML/Code2World/raw/3e48194a8228570164ecd510f9d83ac839e9e931/README.md\"], \"reason\": \"Publisher model-card declaration for the checkpoint; code and training-data terms are separate.\"}}, \"repository_url\": \"https://github.com/AMAP-ML/Code2World\"}\n\nCard identifies Code2World-8B and Qwen3-VL-8B-Instruct. Public files include four Safetensors shards, an index and rendering/prompt helper scripts. The GitHub adapter targets AndroidWorld; model-generated HTML rendering and model loading were not executed. MIT is recorded as the weight declaration, without transferring it to the separate GitHub repository’s code or all training sources.\n\nPrimary source: https://huggingface.co/GD-ML/Code2World\nCatalog record: https://awesomecua.github.io/resources/url-bc3deaf8eeecf569e73b/"}
{"id": "url:bd7214de6d98917c2350", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2", "citation_url": "https://awesomecua.github.io/resources/url-bd7214de6d98917c2350/", "details": {"benchmark_protocol": {"actions": ["click coordinates"], "counts": [], "evaluation": "The inspected Jedi evaluation adapter reads the v2 mobile/desktop/web annotation files, scores a point against the target box, and reports overall, platform, and text/icon accuracy. This is a named compatible evaluator implementation; the OS-Atlas dataset card itself supplies no evaluator specification.", "evidence_urls": ["https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2", "https://osatlas.github.io/", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py", "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"], "focus": ["single-step GUI grounding on the ScreenSpot-v2 release"], "notes": "The public Hub revision 5efbb1f1 contains three platform JSON annotation files and an image ZIP; its card only declares Apache-2.0. No sample total or equivalence to the original ScreenSpot test count is asserted. Keep v2 separate from v1 and retain evaluator/version and coordinate conversion when comparing results. Files were listed, not downloaded as data.", "observations": ["screenshot", "instruction"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2/tree/main", "gated": false, "hub_id": "OS-Copilot/ScreenSpot-v2", "languages": [], "modalities": ["image"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.benchmark_protocol"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:22:05Z", "role": "primary", "sha256": "94547cb88756cf6a36bfb86d5fe9bdca1e40d7edb9e452c33cb462bbe760fed0", "supports": ["details.benchmark_protocol"], "url": "https://osatlas.github.io/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "0e8772bc8fd5a25c50d3c2f6b0caee05ed2ab4659d46d99d60adf2b5ab29e997", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:21:09Z", "role": "primary", "sha256": "b85f5685ce6c200ea648471666177859693d7f40cf50c12d08765ae68580de4a", "supports": ["details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "7feddadc306e384f8df6a3d9c8afc0c776cb3734c4c7fe3eb842eab85537f408", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2"}], "name": "ScreenSpot v2", "publication": null, "published_at": null, "record_sha256": "b7dbf8e0231817d5e8d0067c471af9a24611cafea7d01e59a5c054a2962bc8b9", "relations": [], "resource_id": "url:bd7214de6d98917c2350", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2", "topics": ["grounding", "evaluation"], "updated_at": null}, "text": "ScreenSpot v2\nResource type: benchmark\nReview status: curated\nScope: core\n\nSecond version of the ScreenSpot grounding test set, published by OS-Copilot as separate desktop, mobile and web annotation files with a shared screenshot archive.\nPlatforms: cross-platform\nTopics: grounding, evaluation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"click coordinates\"], \"counts\": [], \"evaluation\": \"The inspected Jedi evaluation adapter reads the v2 mobile/desktop/web annotation files, scores a point against the target box, and reports overall, platform, and text/icon accuracy. This is a named compatible evaluator implementation; the OS-Atlas dataset card itself supplies no evaluator specification.\", \"evidence_urls\": [\"https://huggingface.co/api/datasets/OS-Copilot/ScreenSpot-v2\", \"https://osatlas.github.io/\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/qwen25_vllm_screenspot_v2_jedi.py\", \"https://raw.githubusercontent.com/xlang-ai/OSWorld-G/HEAD/evaluation/eval.py\"], \"focus\": [\"single-step GUI grounding on the ScreenSpot-v2 release\"], \"notes\": \"The public Hub revision 5efbb1f1 contains three platform JSON annotation files and an image ZIP; its card only declares Apache-2.0. No sample total or equivalence to the original ScreenSpot test count is asserted. Keep v2 separate from v1 and retain evaluator/version and coordinate conversion when comparing results. Files were listed, not downloaded as data.\", \"observations\": [\"screenshot\", \"instruction\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/ScreenSpot-v2\", \"modalities\": [\"image\"]}\n\nPrimary source: https://huggingface.co/datasets/OS-Copilot/ScreenSpot-v2\nCatalog record: https://awesomecua.github.io/resources/url-bd7214de6d98917c2350/"}
{"id": "url:bdfb50330deb97893ce4", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web", "citation_url": "https://awesomecua.github.io/resources/url-bdfb50330deb97893ce4/", "details": {"availability": "Ungated JSON conversation shards and ranking-file manifest.", "dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "train", "qualifier": "exact", "unit": "conversations", "value": 600}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-task", "qualifier": "exact", "unit": "conversations", "value": 34}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-website", "qualifier": "exact", "unit": "conversations", "value": 42}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-subdomain", "qualifier": "exact", "unit": "conversations", "value": 44}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "train", "qualifier": "exact", "unit": "dialogue turns", "value": 2896}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-task", "qualifier": "exact", "unit": "dialogue turns", "value": 191}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-website", "qualifier": "exact", "unit": "dialogue turns", "value": 218}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md", "population": "test-subdomain", "qualifier": "exact", "unit": "dialogue turns", "value": 216}], "evidence_urls": ["https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"], "formats": ["JSON", "auxiliary PKL"], "modalities": ["dialogue text", "HTML", "element candidates", "actions"], "notes": "Conversation, dialogue-turn and action-step units differ. The released file manifest is inspected; underlying examples were not downloaded.", "splits": ["train", "test_task", "test_website", "test_subdomain"], "supervision": ["multi-turn web instruction following", "element/action prediction"]}, "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"], "reason": "Publisher dataset-card declaration for the multi-turn annotation release."}}, "release_revision": "5ae9a3bedf431f54c2218fbb8ee474e5efbda3a8"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "8685e0a26a06b3339b5c078632a7b0ade1ead40e6b586ef1ff3676414699a4bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "details.dataset_profile", "licenses.data"], "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T12:34:10Z", "role": "primary", "sha256": "4455e6daf1151694c03b2a1ebd4a7af40b5b079b0568706a1783a0222383bb1d", "supports": ["identity", "release_status", "details.release_revision"], "url": "https://huggingface.co/api/datasets/magicgh/MT-Mind2Web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:58Z", "role": "primary", "sha256": "c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T12:34:10Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for the multi-turn annotation release.", "declared": "cc-by-4.0", "retrieved_at": "2026-10-17T12:34:10Z", "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web"}], "name": "MT-Mind2Web", "publication": null, "published_at": null, "record_sha256": "8ff3feb5e9fcebe934f52da693aae616ba57f3d12810458e50e4e80ea1d12287", "relations": [{"evidence_url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=800&max_results=100&sortBy=submittedDate&sortOrder=descending", "target_id": "arxiv:2402.15057", "type": "documented_by"}], "resource_id": "url:bdfb50330deb97893ce4", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/magicgh/MT-Mind2Web", "topics": ["web-navigation", "imitation-learning", "memory"], "updated_at": null}, "text": "MT-Mind2Web\nResource type: dataset\nReview status: curated\nScope: core\n\nMulti-turn web-navigation conversations constructed from Mind2Web interactions, with dialogue turns, HTML snapshots, element candidates and click/type/select action sequences.\nPlatforms: web\nTopics: web-navigation, imitation-learning, memory\n\nRecorded specifications: {\"availability\": \"Ungated JSON conversation shards and ranking-file manifest.\", \"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"train\", \"qualifier\": \"exact\", \"unit\": \"conversations\", \"value\": 600}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-task\", \"qualifier\": \"exact\", \"unit\": \"conversations\", \"value\": 34}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-website\", \"qualifier\": \"exact\", \"unit\": \"conversations\", \"value\": 42}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-subdomain\", \"qualifier\": \"exact\", \"unit\": \"conversations\", \"value\": 44}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"train\", \"qualifier\": \"exact\", \"unit\": \"dialogue turns\", \"value\": 2896}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-task\", \"qualifier\": \"exact\", \"unit\": \"dialogue turns\", \"value\": 191}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-website\", \"qualifier\": \"exact\", \"unit\": \"dialogue turns\", \"value\": 218}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\", \"population\": \"test-subdomain\", \"qualifier\": \"exact\", \"unit\": \"dialogue turns\", \"value\": 216}], \"evidence_urls\": [\"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\"], \"formats\": [\"JSON\", \"auxiliary PKL\"], \"modalities\": [\"dialogue text\", \"HTML\", \"element candidates\", \"actions\"], \"notes\": \"Conversation, dialogue-turn and action-step units differ. The released file manifest is inspected; underlying examples were not downloaded.\", \"splits\": [\"train\", \"test_task\", \"test_website\", \"test_subdomain\"], \"supervision\": [\"multi-turn web instruction following\", \"element/action prediction\"]}, \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/magicgh/MT-Mind2Web/raw/main/README.md\"], \"reason\": \"Publisher dataset-card declaration for the multi-turn annotation release.\"}}, \"release_revision\": \"5ae9a3bedf431f54c2218fbb8ee474e5efbda3a8\"}\n\nThe ungated Hub manifest contains train, test-task, test-website and test-subdomain JSON shards plus a candidate-ranking pickle file. Counts distinguish conversation sessions from dialogue turns; neither is a screenshot or action-step count. No pickle or sample data was loaded. The card declares CC-BY-4.0 and identifies the original Mind2Web interactions as its source.\n\nPrimary source: https://huggingface.co/datasets/magicgh/MT-Mind2Web\nCatalog record: https://awesomecua.github.io/resources/url-bdfb50330deb97893ce4/"}
{"id": "url:beb8f1fd6b945ffc9d93", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Shangzhan Zhang", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Yuchen Yan", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang"], "canonical_source_url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "citation_url": "https://awesomecua.github.io/resources/url-beb8f1fd6b945ffc9d93/", "details": {"publication": {"authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Shangzhan Zhang", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Yuchen Yan", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang"], "date_basis": "Explicit publisher metadata: month = {June}, year = {2026}; precision=month.", "date_parts": [2026, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"], "identity_basis": "No title/DOI/canonical URL match in the current paper catalog or raw archive; closest substantive same-team works were reread and have different contributions.", "issue": null, "notes": "CVF open-access accepted version; PDF states that the final proceedings version is on IEEE Xplore.", "pages": "13007-13016", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf", "publisher": null, "series": null, "title": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)", "volume": null}}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "1d5ed4e24cf9763cc43f0a8ce14f0e98c01504c73578e5632cea573dfb7da647", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "244f40d77a40dfd184e78cf8e1e2dd9e469d3e594a8ce214279c2ce17b8f1012", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:48+00:00", "role": "primary", "sha256": "1d5ed4e24cf9763cc43f0a8ce14f0e98c01504c73578e5632cea573dfb7da647", "supports": ["details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "scope", "summary", "topics"], "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:56:04+00:00", "role": "primary", "sha256": "244f40d77a40dfd184e78cf8e1e2dd9e469d3e594a8ce214279c2ce17b8f1012", "supports": ["details.publication.authors", "details.publication.identity_basis", "details.publication.notes", "details.publication.title", "environments", "name", "scope", "summary", "topics"], "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}], "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:53:48+00:00", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html"}, {"rel": "publication_pdf", "url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"}], "name": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning", "publication": {"authors": ["Fei Tang", "Zhangxuan Gu", "Zhengxi Lu", "Shangzhan Zhang", "Zhengwen Zeng", "Shuheng Shen", "Changhua Meng", "Yuchen Yan", "Wenqi Zhang", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang"], "date_basis": "Explicit publisher metadata: month = {June}, year = {2026}; precision=month.", "date_parts": [2026, 6], "doi": null, "event_dates": null, "evidence_urls": ["https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf"], "identity_basis": "No title/DOI/canonical URL match in the current paper catalog or raw archive; closest substantive same-team works were reread and have different contributions.", "issue": null, "notes": "CVF open-access accepted version; PDF states that the final proceedings version is on IEEE Xplore.", "pages": "13007-13016", "pdf_url": "https://openaccess.thecvf.com/content/CVPR2026/papers/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.pdf", "publisher": null, "series": null, "title": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning", "type": "conference-paper", "url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "venue": "Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)", "volume": null}, "published_at": null, "record_sha256": "e26ac97c8b5016bec0ef47fa2ef9fd50980f1044d2f9395514fafcefe2f391d1", "relations": [], "resource_id": "url:beb8f1fd6b945ffc9d93", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html", "topics": ["reinforcement-learning", "data-generation", "evaluation", "robustness"], "updated_at": null}, "text": "GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning\nResource type: paper\nReview status: curated\nScope: core\n\nGUI-SAGE generates reasoning conditioned on known actions and weights policy updates using confidence and reward signals. Its experiments use AndroidControl and GUIOdyssey data and separately evaluate interactive AndroidWorld tasks.\nPlatforms: android\nTopics: reinforcement-learning, data-generation, evaluation, robustness\nRecord publication date (recorded precision): 2026-06\nAuthors: Fei Tang, Zhangxuan Gu, Zhengxi Lu, Shangzhan Zhang, Zhengwen Zeng, Shuheng Shen, Changhua Meng, Yuchen Yan, Wenqi Zhang, Yongliang Shen, Weiming Lu, Yueting Zhuang\n\nPublication edition: GUI-SAGE: Enhancing GUI Automation with Self-Explanatory Learning\nPublication authors: Fei Tang, Zhangxuan Gu, Zhengxi Lu, Shangzhan Zhang, Zhengwen Zeng, Shuheng Shen, Changhua Meng, Yuchen Yan, Wenqi Zhang, Yongliang Shen, Weiming Lu, Yueting Zhuang\nPublication edition date (recorded precision): 2026-06\nPublication venue: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)\nPublication pages: 13007-13016\nPublication date basis: Explicit publisher metadata: month = {June}, year = {2026}; precision=month.\nPublication notes: CVF open-access accepted version; PDF states that the final proceedings version is on IEEE Xplore.\nPublication source: https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. New work proposal from the official CVF accepted open-access version. No invented day, DOI, or arXiv identifier; original annotations describe the reviewed sources and not independent reproduction.\n\nPrimary source: https://openaccess.thecvf.com/content/CVPR2026/html/Tang_GUI-SAGE_Enhancing_GUI_Automation_with_Self-Explanatory_Learning_CVPR_2026_paper.html\nCatalog record: https://awesomecua.github.io/resources/url-beb8f1fd6b945ffc9d93/"}
{"id": "url:bf0c11018e49dff6aeec", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Allen Cypher"], "canonical_source_url": "https://acypher.com/Publications/CHI91/EagerCHI.html", "citation_url": "https://awesomecua.github.io/resources/url-bf0c11018e49dff6aeec/", "details": {"publication": {"authors": ["Allen Cypher"], "date_basis": "Supplied year precision: 1991. Opening CHI 1991 citation and copyright", "date_parts": [1991], "doi": null, "event_dates": null, "evidence_urls": ["https://acypher.com/Eager/", "https://acypher.com/Publications/CHI91/EagerCHI.html"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS. DOI not verified in a retained first-party source. Exact publication month/day is not inferred from conference dates.", "pages": "33-39", "pdf_url": null, "publisher": null, "series": null, "title": "Eager: Programming Repetitive Tasks by Example", "type": "conference-paper", "url": "https://acypher.com/Publications/CHI91/EagerCHI.html", "venue": "Proceedings of CHI '91", "volume": null}}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:27Z", "role": "primary", "sha256": "64cffd3fca934465e1ea28bdf62e087bc9abdf50b9b72d04c2e57df0bcf41adb", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title"], "url": "https://acypher.com/Eager/"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "876d4c8dfac48443336a5ee97b463642db013631a2b97cacd090e61afd95ad59", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.type", "details.publication.url", "details.publication.venue"], "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "876d4c8dfac48443336a5ee97b463642db013631a2b97cacd090e61afd95ad59", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.type", "details.publication.url", "details.publication.venue", "environments", "kind", "scope", "summary", "topics"], "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:27Z", "role": "primary", "sha256": "64cffd3fca934465e1ea28bdf62e087bc9abdf50b9b72d04c2e57df0bcf41adb", "supports": ["details.publication.identity_basis", "details.publication.notes", "details.publication.title", "name"], "url": "https://acypher.com/Eager/"}], "identifiers": {}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://acypher.com/Publications/CHI91/EagerCHI.html"}], "name": "Eager: Programming Repetitive Tasks by Example", "publication": {"authors": ["Allen Cypher"], "date_basis": "Supplied year precision: 1991. Opening CHI 1991 citation and copyright", "date_parts": [1991], "doi": null, "event_dates": null, "evidence_urls": ["https://acypher.com/Eager/", "https://acypher.com/Publications/CHI91/EagerCHI.html"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS. DOI not verified in a retained first-party source. Exact publication month/day is not inferred from conference dates.", "pages": "33-39", "pdf_url": null, "publisher": null, "series": null, "title": "Eager: Programming Repetitive Tasks by Example", "type": "conference-paper", "url": "https://acypher.com/Publications/CHI91/EagerCHI.html", "venue": "Proceedings of CHI '91", "volume": null}, "published_at": null, "record_sha256": "01b367ee77e10cb5175b7d0fd07cfa8cb312704cd94e87bf0a6018fb72004e5a", "relations": [], "resource_id": "url:bf0c11018e49dff6aeec", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://acypher.com/Publications/CHI91/EagerCHI.html", "topics": ["automation", "human-in-the-loop", "interaction-history"], "updated_at": null}, "text": "Eager: Programming Repetitive Tasks by Example\nResource type: paper\nReview status: curated\nScope: core\n\nDescribes Eager, which detects repeated HyperCard interactions, generalizes them into loops, and previews predicted actions by highlighting interface objects. Users can then ask it to execute the inferred repetitive task.\nPlatforms: desktop\nTopics: automation, human-in-the-loop, interaction-history\nRecord publication date (recorded precision): 1991\nAuthors: Allen Cypher\n\nPublication edition: Eager: Programming Repetitive Tasks by Example\nPublication authors: Allen Cypher\nPublication edition date (recorded precision): 1991\nPublication venue: Proceedings of CHI '91\nPublication pages: 33-39\nPublication date basis: Supplied year precision: 1991. Opening CHI 1991 citation and copyright\nPublication notes: The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS. DOI not verified in a retained first-party source. Exact publication month/day is not inferred from conference dates.\nPublication source: https://acypher.com/Publications/CHI91/EagerCHI.html\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. The 1991 conference paper is kept distinct from the 1991 video and 1993/1995 reprints listed on the author project page. Macintosh/HyperCard is tagged desktop; it is not represented as support for contemporary macOS.\n\nPrimary source: https://acypher.com/Publications/CHI91/EagerCHI.html\nCatalog record: https://awesomecua.github.io/resources/url-bf0c11018e49dff6aeec/"}
{"id": "url:c0bd07d82ce25ad1d99d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://hcompany.ai/research/models", "citation_url": "https://awesomecua.github.io/resources/url-c0bd07d82ce25ad1d99d/", "details": {}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "84ce0fd682367a6dac5cc5abc39d5c877556024bbf6ffdad30f76fc2ec08523f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://hcompany.ai/research/models"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://hcompany.ai/research/models"}], "name": "Holo Model Family", "publication": null, "published_at": null, "record_sha256": "73c7aee784e746754211e94c14ab27de4d0662825109dd4dc516d26168187a1b", "relations": [{"evidence_url": "https://hcompany.ai/research/models", "target_id": "organization:h", "type": "published_by"}], "resource_id": "url:c0bd07d82ce25ad1d99d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://hcompany.ai/research/models", "topics": ["planning", "grounding"], "updated_at": null}, "text": "Holo Model Family\nResource type: model\nReview status: curated\nScope: core\n\nH computer-use model family and its release-specific access options.\nPlatforms: cross-platform\nTopics: planning, grounding\n\nPrimary source: https://hcompany.ai/research/models\nCatalog record: https://awesomecua.github.io/resources/url-c0bd07d82ce25ad1d99d/"}
{"id": "url:c0f39f7cf2cd9c18e8bd", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/showlab/ShowUI-2B", "citation_url": "https://awesomecua.github.io/resources/url-c0f39f7cf2cd9c18e8bd/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-2B/tree/main", "gated": false, "hub_id": "showlab/ShowUI-2B", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "bb6d120d4795492a197e0a9e9df5c56b7baa8ac5c08182ce0f20f36c8b65783d", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-2B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/showlab/ShowUI-2B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-2B"}], "name": "ShowUI 2B", "publication": null, "published_at": null, "record_sha256": "2f1e5035d9fd1f558a0fa88a059db2a811b56439b35f8a8061c84b5e14b2f213", "relations": [{"evidence_url": "https://huggingface.co/api/models/showlab/ShowUI-2B", "target_id": "github:showlab/showui", "type": "variant_of"}], "resource_id": "url:c0f39f7cf2cd9c18e8bd", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/showlab/ShowUI-2B", "topics": ["grounding", "efficiency"], "updated_at": null}, "text": "ShowUI 2B\nResource type: model\nReview status: curated\nScope: core\n\nShowUI vision-language-action checkpoint fine-tuned from Qwen2-VL-2B-Instruct, published as a single PyTorch weight file under the MIT license and tagged for GUI agents and computer use.\nPlatforms: cross-platform\nTopics: grounding, efficiency\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/showlab/ShowUI-2B/tree/main\", \"gated\": false, \"hub_id\": \"showlab/ShowUI-2B\"}\n\nPrimary source: https://huggingface.co/showlab/ShowUI-2B\nCatalog record: https://awesomecua.github.io/resources/url-c0f39f7cf2cd9c18e8bd/"}
{"id": "url:c1db2ba79e6d6a3db2b9", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations", "citation_url": "https://awesomecua.github.io/resources/url-c1db2ba79e6d6a3db2b9/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "full recipe", "qualifier": "exact", "unit": "annotation rows", "value": 633369}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "population": "point recipe", "qualifier": "exact", "unit": "annotation rows", "value": 610894}], "evidence_urls": ["https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"], "formats": ["JSONL", "JSON task recipes", "TSV manifests"], "modalities": ["text", "coordinates", "image references"], "notes": "The recipes reuse source images and are not additive datasets. This download contains no screenshots. Counts are publisher-reported; rows named for trajectories are independent decision examples, not full trajectories.", "splits": ["full/train", "point/train"], "supervision": ["instruction-to-click grounding", "bounding boxes", "absent-target abstention"]}, "license_name": "restricted-research-only"}, "environments": ["desktop", "web", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "96a87222645c3f823e6f67e50708792e2917ddeb8eea5aacdf2de8c5bbff165f", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations"}], "name": "ScreenRef Annotations", "publication": null, "published_at": null, "record_sha256": "a3a42097ecf9be6869c02e071064f567e2c02023d2d241097134e071e7c6658f", "relations": [{"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "component_of"}], "resource_id": "url:c1db2ba79e6d6a3db2b9", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "ScreenRef Annotations\nResource type: dataset\nReview status: curated\nScope: core\n\nAnnotation files, task recipes, and image manifests for ScreenRef; screenshots reside in the separately gated image repository.\nPlatforms: desktop, web, android, linux\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\", \"population\": \"full recipe\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 633369}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\", \"population\": \"point recipe\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 610894}], \"evidence_urls\": [\"https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations/raw/main/README.md\"], \"formats\": [\"JSONL\", \"JSON task recipes\", \"TSV manifests\"], \"modalities\": [\"text\", \"coordinates\", \"image references\"], \"notes\": \"The recipes reuse source images and are not additive datasets. This download contains no screenshots. Counts are publisher-reported; rows named for trajectories are independent decision examples, not full trajectories.\", \"splits\": [\"full/train\", \"point/train\"], \"supervision\": [\"instruction-to-click grounding\", \"bounding boxes\", \"absent-target abstention\"]}, \"license_name\": \"restricted-research-only\"}\n\nThe card declares restricted academic research use and no redistribution. Annotation coordinates use a normalized 1000 by 1000 space. Some source instructions name gestures other than the single-click target; the publisher recommends treating those as location supervision.\n\nPrimary source: https://huggingface.co/datasets/PrentisAI/ScreenRef-Annotations\nCatalog record: https://awesomecua.github.io/resources/url-c1db2ba79e6d6a3db2b9/"}
{"id": "url:c4898ad1bfc049926a63", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data", "citation_url": "https://awesomecua.github.io/resources/url-c4898ad1bfc049926a63/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/api/datasets/osunlp/UGround-V1-Data", "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"], "formats": ["Parquet shards"], "modalities": ["GUI screenshots", "natural-language grounding descriptions"], "notes": "Public API revision 33325877 lists Parquet assets but gated=auto; the pinned raw card returned HTTP 401 and no access was requested. Official README identifies this as V1 training data; the API excerpt also announces a Web-Hybrid bounding-box variant with coordinates normalized to [0,999]. That does not establish every shard schema or a release-wide point/bbox count. No train/test partition, example total, or unrestricted downloadability is claimed. Preserve the recorded CC-BY-NC-SA-4.0 data declaration separately from model/code terms.", "splits": [], "supervision": ["GUI visual-grounding training for the UGround-V1 family"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data/tree/main", "gated": "auto", "hub_id": "osunlp/UGround-V1-Data", "languages": [], "modalities": ["tabular", "text"], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "72ca3f23114c3a08fefb724573acf10c9f0a1b3d4a50d811c98356264ecf6ffc", "supports": ["details.dataset_profile", "relations"], "url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "0342862bd7857dd21e389ee5d351375dfea722cad4725c8e33271efc93e009f6", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/osunlp/UGround-V1-Data"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data"}], "name": "UGround V1 Data", "publication": null, "published_at": null, "record_sha256": "ab729b110346cde5f4ca960a4d7a84d251b30dc3311480a95417d7665e307da3", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md", "target_id": "github:osu-nlp-group/uground", "type": "component_of"}], "resource_id": "url:c4898ad1bfc049926a63", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/osunlp/UGround-V1-Data", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "UGround V1 Data\nResource type: dataset\nReview status: curated\nScope: core\n\nGated web grounding training data for UGround in Web-Hybrid and Web-Direct subsets, with a bounding-box version whose coordinates are normalized to 0-999.\nPlatforms: cross-platform\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/api/datasets/osunlp/UGround-V1-Data\", \"https://raw.githubusercontent.com/OSU-NLP-Group/UGround/HEAD/README.md\"], \"formats\": [\"Parquet shards\"], \"modalities\": [\"GUI screenshots\", \"natural-language grounding descriptions\"], \"notes\": \"Public API revision 33325877 lists Parquet assets but gated=auto; the pinned raw card returned HTTP 401 and no access was requested. Official README identifies this as V1 training data; the API excerpt also announces a Web-Hybrid bounding-box variant with coordinates normalized to [0,999]. That does not establish every shard schema or a release-wide point/bbox count. No train/test partition, example total, or unrestricted downloadability is claimed. Preserve the recorded CC-BY-NC-SA-4.0 data declaration separately from model/code terms.\", \"splits\": [], \"supervision\": [\"GUI visual-grounding training for the UGround-V1 family\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/osunlp/UGround-V1-Data/tree/main\", \"gated\": \"auto\", \"hub_id\": \"osunlp/UGround-V1-Data\", \"modalities\": [\"tabular\", \"text\"]}\n\nPrimary source: https://huggingface.co/datasets/osunlp/UGround-V1-Data\nCatalog record: https://awesomecua.github.io/resources/url-c4898ad1bfc049926a63/"}
{"id": "url:c5eb975c0d66f8a2954f", "metadata": {"aliases": ["url:4a309ef39b33a7d1e543", "THUDM/cogagent-9b-20241220", "https://huggingface.co/THUDM/cogagent-9b-20241220"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zai-org/cogagent-9b-20241220", "citation_url": "https://awesomecua.github.io/resources/url-c5eb975c0d66f8a2954f/", "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/tree/main", "gated": false, "hub_id": "zai-org/cogagent-9b-20241220"}, "environments": ["desktop", "windows", "macos", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "ae72fbc40f9c8afd52d1688b2233e5c915571de33618cfd1ffcb7f3fadfb4402", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "c1ad1d99a8d46a00610db35d97c787d10b7e81661e48a1e389ae6cbb51a45630", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/LICENSE"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "de2c8e7a9792c3d129451ec813d458f83263460a5808cf87c45b04b260a376c9", "supports": ["licenses.weights"], "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:13Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/THUDM/cogagent-9b-20241220"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:15:14Z", "role": "primary", "sha256": "35bbe72f95711e042dab7a94bf2711e30a3da50190c7315178c37de5725cc1d9", "supports": ["aliases"], "url": "https://huggingface.co/api/models/zai-org/cogagent-9b-20241220"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/cogagent-9b-20241220"}], "name": "CogAgent 9B 20241220", "publication": null, "published_at": null, "record_sha256": "2245a7691c852fc25b61c2201ded90168a896bc1118a9dfb5d51cf8baa6cbe30", "relations": [{"evidence_url": "https://huggingface.co/zai-org/cogagent-9b-20241220/raw/main/README.md", "target_id": "github:thudm/cogagent", "type": "variant_of"}], "resource_id": "url:c5eb975c0d66f8a2954f", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zai-org/cogagent-9b-20241220", "topics": ["grounding", "action-prediction", "screen-understanding"], "updated_at": null}, "text": "CogAgent 9B 20241220\nResource type: model\nReview status: curated\nScope: core\n\nA bilingual GLM-4V-based execution model that uses screenshots, task instructions and action history to predict GUI operations.\nRecorded aliases: url:4a309ef39b33a7d1e543, THUDM/cogagent-9b-20241220, https://huggingface.co/THUDM/cogagent-9b-20241220\nPlatforms: desktop, windows, macos, cross-platform\nTopics: grounding, action-prediction, screen-understanding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/zai-org/cogagent-9b-20241220/tree/main\", \"gated\": false, \"hub_id\": \"zai-org/cogagent-9b-20241220\"}\n\nThe former THUDM model URL resolves to zai-org. Six public safetensors shards are present. Weights use the custom CogAgent License, including a registration requirement for commercial use, rather than a blanket permissive license. The card cites the 2023 CogAgent paper as previous work and links a later technical report; no claim that the old paper documents this 2024 checkpoint is asserted.\n\nPrimary source: https://huggingface.co/zai-org/cogagent-9b-20241220\nCatalog record: https://awesomecua.github.io/resources/url-c5eb975c0d66f8a2954f/"}
{"id": "url:c73ba0618d1b3c505887", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench", "citation_url": "https://awesomecua.github.io/resources/url-c73ba0618d1b3c505887/", "details": {"availability": "Ungated benchmark JSONL/image manifest and public evaluation scripts; training pipeline/checkpoints are pending.", "benchmark_protocol": {"actions": ["critic-model responses"], "counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md", "population": "mobile/desktop/web configurations in the Hub card", "qualifier": "exact", "unit": "platform JSONL splits", "value": 3}], "evaluation": "The official repository runs model inference and computes accuracy with cal_acc.py.", "evidence_urls": ["https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md", "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"], "focus": ["GUI critic prediction"], "notes": "No sample count is inferred from the n<1K Hub size band. Evaluation code availability does not establish the release of OS-Oracle training data or weights.", "observations": ["benchmark JSONL", "UI screenshots"], "setting": "offline"}, "gated": false, "hub_id": "OS-Copilot/OS-Critic-Bench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"], "reason": "Publisher dataset-card license declaration for OS-Critic-Bench; not an evaluation-code or model-weight license."}}, "release_revision": "8fe774ef6a422c0e555f90205eb392f9132e272b"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:18Z", "role": "primary", "sha256": "58305b8ebe1af5dfbddd636484d63722d4e21d28ed3e22bb8c0068e654f33479", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.benchmark_protocol"], "url": "https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:17Z", "role": "primary", "sha256": "17bcf851f39d06e7e357871df1dc1ddf697bcbea31842c16348227bbf4fbb224", "supports": ["identity"], "url": "https://api.github.com/repos/numbmelon/OS-Oracle"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:22:55Z", "role": "primary", "sha256": "516beea465382b54548e4b8bfd12ae510080a49451b30262ad1c769ed5a3398e", "supports": ["release_status", "licenses.code"], "url": "https://api.github.com/repos/numbmelon/OS-Oracle/git/trees/main?recursive=1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:06Z", "role": "primary", "sha256": "ef311b49437e246ffef6ad13c090277a0bdacc4a124e5b3cfe8d2643ced8f843", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Critic-Bench"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "4139e25e8911ca8450f6db9acb8dc3a98c126c3c43514c72bcefa4ed33b8ffa8", "supports": ["licenses.data", "details.benchmark_protocol"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T11:18:18Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card license declaration for OS-Critic-Bench; not an evaluation-code or model-weight license.", "declared": "apache-2.0", "retrieved_at": "2026-10-17T11:23:07Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench"}], "name": "OS-Critic Bench", "publication": null, "published_at": null, "record_sha256": "9cfe8c8bf50073f99cb4f00e846f18f121a28157184f1127b5abd12fa5520156", "relations": [{"evidence_url": "https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md", "target_id": "arxiv:2512.16295", "type": "documented_by"}], "resource_id": "url:c73ba0618d1b3c505887", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench", "topics": ["evaluation", "reward-models", "screen-understanding"], "updated_at": null}, "text": "OS-Critic Bench\nResource type: benchmark\nReview status: curated\nScope: core\n\nA cross-platform GUI-critic benchmark with mobile, desktop and web JSONL splits and screenshot assets, accompanied by public inference and accuracy-scoring scripts.\nPlatforms: cross-platform, desktop, web\nTopics: evaluation, reward-models, screen-understanding\n\nRecorded specifications: {\"availability\": \"Ungated benchmark JSONL/image manifest and public evaluation scripts; training pipeline/checkpoints are pending.\", \"benchmark_protocol\": {\"actions\": [\"critic-model responses\"], \"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md\", \"population\": \"mobile/desktop/web configurations in the Hub card\", \"qualifier\": \"exact\", \"unit\": \"platform JSONL splits\", \"value\": 3}], \"evaluation\": \"The official repository runs model inference and computes accuracy with cal_acc.py.\", \"evidence_urls\": [\"https://raw.githubusercontent.com/numbmelon/OS-Oracle/HEAD/README.md\", \"https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md\"], \"focus\": [\"GUI critic prediction\"], \"notes\": \"No sample count is inferred from the n<1K Hub size band. Evaluation code availability does not establish the release of OS-Oracle training data or weights.\", \"observations\": [\"benchmark JSONL\", \"UI screenshots\"], \"setting\": \"offline\"}, \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Critic-Bench\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench/raw/main/README.md\"], \"reason\": \"Publisher dataset-card license declaration for OS-Critic-Bench; not an evaluation-code or model-weight license.\"}}, \"release_revision\": \"8fe774ef6a422c0e555f90205eb392f9132e272b\"}\n\nThe Hub card provides split/format/license metadata, while the official OS-Oracle README establishes critic-evaluation purpose and platform scope. The repository currently releases only benchmark inference/scoring components; data-synthesis code, training datasets and model checkpoints remain unchecked TODOs and are not counted as released assets. Apache-2.0 is recorded for benchmark data only; the evaluation code has no inspected license declaration.\n\nPrimary source: https://huggingface.co/datasets/OS-Copilot/OS-Critic-Bench\nCatalog record: https://awesomecua.github.io/resources/url-c73ba0618d1b3c505887/"}
{"id": "url:c79e152e6f7af1a9efed", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro", "citation_url": "https://awesomecua.github.io/resources/url-c79e152e6f7af1a9efed/", "details": {"dataset_profile": {"counts": [{"basis": "computed", "evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "population": "sum of icon/text counts in the published card table", "qualifier": "exact", "unit": "target annotations", "value": 1581}], "evidence_urls": ["https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"], "formats": ["JSON annotations", "screen images"], "modalities": ["high-resolution screenshots", "instructions", "bounding boxes"], "notes": "Counts describe target annotations, not unique screenshots or expanded language/style variants. The evaluation code can reuse a target across configurations.", "splits": [], "supervision": ["GUI grounding"]}, "download_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/tree/main"}, "environments": ["desktop", "windows", "macos", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:38:05Z", "role": "primary", "sha256": "a77c859775b33591b5619e8520c94702a53a1f9a29ada93e66744dbe23afe9dc", "supports": ["details.dataset_profile"], "url": "https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:54Z", "role": "primary", "sha256": "4953a8e0f029016aef668bce98cec4d2873a7ae49c1fe33c4311fa5eb8d24089", "supports": ["relations"], "url": "https://export.arxiv.org/api/query?search_query=%28ti%3A%22computer+use%22+OR+ti%3A%22computer-use%22+OR+ti%3A%22GUI+agent%22+OR+ti%3A%22GUI+agents%22+OR+ti%3A%22GUI+grounding%22+OR+ti%3A%22web+agents%22+OR+ti%3A%22browser+agent%22%29+AND+submittedDate%3A%5B200001010000+TO+202610160636%5D&start=700&max_results=100&sortBy=submittedDate&sortOrder=descending"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:10Z", "role": "primary", "sha256": "2f7f635dee3c593b40efc61246c082ab1d76f7ce78f84a1d4b6391d30fa4e8fd", "supports": ["relations"], "url": "https://gui-agent.github.io/grounding-leaderboard/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:35:55Z", "role": "primary", "sha256": "76ae2231564fe84af029f8625c02d9cd7fb6cf85c5f8af1b2a2593e2107719d9", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:35:55Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-16T11:35:55Z", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro"}], "name": "ScreenSpot-Pro data", "publication": null, "published_at": null, "record_sha256": "be643387092c7bf17b35a6a5faad519021a4941be22efea6539fcfa3b60b5cf6", "relations": [{"evidence_url": "https://gui-agent.github.io/grounding-leaderboard/", "target_id": "arxiv:2504.07981", "type": "documented_by"}, {"evidence_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md", "target_id": "github:likaixin2000/screenspot-pro-gui-grounding", "type": "component_of"}], "resource_id": "url:c79e152e6f7af1a9efed", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/likaixin/ScreenSpot-Pro", "topics": ["grounding", "evaluation"], "updated_at": null}, "text": "ScreenSpot-Pro data\nResource type: dataset\nReview status: curated\nScope: core\n\nHigh-resolution professional interface screenshots and target annotations for evaluating instruction-based GUI localization.\nPlatforms: desktop, windows, macos, linux\nTopics: grounding, evaluation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"computed\", \"evidence_url\": \"https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md\", \"population\": \"sum of icon/text counts in the published card table\", \"qualifier\": \"exact\", \"unit\": \"target annotations\", \"value\": 1581}], \"evidence_urls\": [\"https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/raw/main/README.md\", \"https://raw.githubusercontent.com/likaixin2000/ScreenSpot-Pro-GUI-Grounding/HEAD/eval_screenspot_pro.py\"], \"formats\": [\"JSON annotations\", \"screen images\"], \"modalities\": [\"high-resolution screenshots\", \"instructions\", \"bounding boxes\"], \"notes\": \"Counts describe target annotations, not unique screenshots or expanded language/style variants. The evaluation code can reuse a target across configurations.\", \"splits\": [], \"supervision\": [\"GUI grounding\"]}, \"download_url\": \"https://huggingface.co/datasets/likaixin/ScreenSpot-Pro/tree/main\"}\n\nThis is the benchmark data asset, distinct from its evaluation-code repository. Target counts are not unique image counts.\n\nPrimary source: https://huggingface.co/datasets/likaixin/ScreenSpot-Pro\nCatalog record: https://awesomecua.github.io/resources/url-c79e152e6f7af1a9efed/"}
{"id": "url:c8c046847f829fea4821", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zonghanHZH/ZonUI-3B", "citation_url": "https://awesomecua.github.io/resources/url-c8c046847f829fea4821/", "details": {"disabled": false, "download_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/tree/07212ea7724a69400422571e0f6e053212e66fe6", "gated": false, "hub_id": "zonghanHZH/ZonUI-3B", "hub_revision": "07212ea7724a69400422571e0f6e053212e66fe6", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"], "reason": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate."}}, "repository_url": "https://github.com/Han1018/ZonUI-3B"}, "environments": ["cross-platform", "desktop", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "0dcd557508c07c9569e642a0319e71162b50c7ac6a0449564760ac2d6fffb089", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.weights"], "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "c419027b878ed7ebebd91f89a7395506d2f2aa8e1745918833cf1bac9e0a054e", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/zonghanHZH/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "9918231748808ea1af93b559b0a511f1f9687689766a09b43c7726f14f27248b", "supports": ["release_status", "details.repository_url"], "url": "https://raw.githubusercontent.com/Han1018/ZonUI-3B/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:02Z", "role": "primary", "sha256": "7e3e63c1f2b38daaf4bb222b51fa5ae6594a0185e618723371e044bee6f05f7d", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:25:20Z", "role": "primary", "sha256": "d4c1500dbaceba738565d766e7b7453abbb6c0b12edaa34e5600d351214bfe4f", "supports": ["notes"], "url": "https://api.github.com/repos/Han1018/ZonUI-3B/git/trees/main?recursive=1"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher model-card declaration for the checkpoint; code and training-data terms are separate.", "declared": "Apache-2.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zonghanHZH/ZonUI-3B"}], "name": "ZonUI-3B checkpoint", "publication": null, "published_at": null, "record_sha256": "08c6c1f40b589a92c61863a86b6cd9b56a377aebe6adc3ae27d6520eaa9c9f90", "relations": [{"evidence_url": "https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md", "target_id": "arxiv:2506.23491", "type": "documented_by"}], "resource_id": "url:c8c046847f829fea4821", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zonghanHZH/ZonUI-3B", "topics": ["grounding", "screen-understanding", "efficiency"], "updated_at": null}, "text": "ZonUI-3B checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nQwen2.5-VL-3B-based GUI grounding checkpoint that predicts click locations from screenshots and text instructions across desktop, mobile and web screens.\nPlatforms: cross-platform, desktop, web\nTopics: grounding, screen-understanding, efficiency\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/zonghanHZH/ZonUI-3B/tree/07212ea7724a69400422571e0f6e053212e66fe6\", \"gated\": false, \"hub_id\": \"zonghanHZH/ZonUI-3B\", \"hub_revision\": \"07212ea7724a69400422571e0f6e053212e66fe6\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/zonghanHZH/ZonUI-3B/raw/07212ea7724a69400422571e0f6e053212e66fe6/README.md\"], \"reason\": \"Publisher model-card declaration for the checkpoint; code and training-data terms are separate.\"}}, \"repository_url\": \"https://github.com/Han1018/ZonUI-3B\"}\n\nThe Hub manifest lists four-part PyTorch and Safetensors files as alternative representations of one checkpoint, plus configuration/tokenizer files. The card’s example rescales predicted locations through the resized image dimensions. No checkpoint was loaded and no benchmark score was reproduced. Apache-2.0 is the model-card weight declaration; the separate code repository has no project-level license in the inspected tree.\n\nPrimary source: https://huggingface.co/zonghanHZH/ZonUI-3B\nCatalog record: https://awesomecua.github.io/resources/url-c8c046847f829fea4821/"}
{"id": "url:c918d5c828a3c6cfbc5d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B", "citation_url": "https://awesomecua.github.io/resources/url-c918d5c828a3c6cfbc5d/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B/tree/main", "gated": false, "hub_id": "zai-org/AutoGLM-Phone-9B", "languages": ["zh"], "modalities": [], "size_category_reported": null}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "94b02fa7104849cedee8ddd1fa33e22d3039e05ae9dd03033805a000e3fed847", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "94b02fa7104849cedee8ddd1fa33e22d3039e05ae9dd03033805a000e3fed847", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B"}], "name": "AutoGLM Phone 9B", "publication": null, "published_at": null, "record_sha256": "e602710f37806139cc2259b32a841e63cb151fb2d80d208cd79bcf2b21299513", "relations": [{"evidence_url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B", "target_id": "github:zai-org/open-autoglm", "type": "variant_of"}], "resource_id": "url:c918d5c828a3c6cfbc5d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B", "topics": ["mobile-navigation"], "updated_at": null}, "text": "AutoGLM Phone 9B\nResource type: model\nReview status: curated\nScope: core\n\nAutoGLM phone-agent checkpoint for Open-AutoGLM, fine-tuned from GLM-4.1V-9B-Base with about 10.3B BF16 parameters and tagged for Chinese.\nPlatforms: android, ios\nTopics: mobile-navigation\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/zai-org/AutoGLM-Phone-9B/tree/main\", \"gated\": false, \"hub_id\": \"zai-org/AutoGLM-Phone-9B\", \"languages\": [\"zh\"]}\n\nPrimary source: https://huggingface.co/zai-org/AutoGLM-Phone-9B\nCatalog record: https://awesomecua.github.io/resources/url-c918d5c828a3c6cfbc5d/"}
{"id": "url:cc4aa8caeee0cff58e73", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/mllmTeam/MobileViews", "citation_url": "https://awesomecua.github.io/resources/url-cc4aa8caeee0cff58e73/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "population": "publisher-reported MobileViews-600K first release", "qualifier": "more-than", "unit": "screenshot–view-hierarchy pairs", "value": 600000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "population": "publisher-reported MobileViews-600K first release", "qualifier": "more-than", "unit": "Android apps", "value": 20000}], "evidence_urls": ["https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md", "https://huggingface.co/api/datasets/mllmTeam/MobileViews"], "formats": ["JPEG", "JSON", "XML", "Parquet", "CSV", "ZIP packages", "HTML/JavaScript trace visualization"], "modalities": ["screenshots", "view hierarchies", "action traces", "app metadata"], "notes": "Counts describe the card’s first-release claim, not an independently counted archive inventory. The card distinguishes globally image-hash-deduplicated screenshot/VH pairs from the November 2024 complete-app traces; trace state counts must not be added as new unique screens. JPEG/JSON and Parquet are alternative representations. Listed filename ranges are examples, not a complete count proof. No train/validation/test split is documented. Hub revision c734715e is ungated; MIT is its data declaration, without resolving third-party app content rights.", "splits": [], "supervision": ["paired screenshot/view-hierarchy structure", "recorded app states, actions and transitions"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/mllmTeam/MobileViews/tree/main", "gated": false, "hub_id": "mllmTeam/MobileViews", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "d36d62283deecd3c5ccf1ac1af49aa7ba779a1d69cdcaea39a1bed37ff5d66b9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T14:10:01Z", "role": "primary", "sha256": "94a4f04e180dbd597c271b0d7470c907e6526b95a6420f01b3183feb1ee24d6e", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "d36d62283deecd3c5ccf1ac1af49aa7ba779a1d69cdcaea39a1bed37ff5d66b9", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:38:26Z", "url": "https://huggingface.co/datasets/mllmTeam/MobileViews"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/mllmTeam/MobileViews"}], "name": "MobileViews", "publication": null, "published_at": null, "record_sha256": "58b45717de9e394cf19fb584d24920e0d914d3b5aedec0183274add1969fb45d", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/mllmTeam/MobileViews", "target_id": "arxiv:2409.14337", "type": "documented_by"}], "resource_id": "url:cc4aa8caeee0cff58e73", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/mllmTeam/MobileViews", "topics": ["screen-understanding"], "updated_at": null}, "text": "MobileViews\nResource type: dataset\nReview status: curated\nScope: core\n\nOver 600,000 Android screenshot and view-hierarchy pairs from more than 20,000 Google Play apps, collected with an optimized DroidBot, with complete per-app traces.\nPlatforms: android\nTopics: screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md\", \"population\": \"publisher-reported MobileViews-600K first release\", \"qualifier\": \"more-than\", \"unit\": \"screenshot–view-hierarchy pairs\", \"value\": 600000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md\", \"population\": \"publisher-reported MobileViews-600K first release\", \"qualifier\": \"more-than\", \"unit\": \"Android apps\", \"value\": 20000}], \"evidence_urls\": [\"https://huggingface.co/datasets/mllmTeam/MobileViews/raw/c734715eb90a8f3ac0c80973f93ee63950e5a8c8/README.md\", \"https://huggingface.co/api/datasets/mllmTeam/MobileViews\"], \"formats\": [\"JPEG\", \"JSON\", \"XML\", \"Parquet\", \"CSV\", \"ZIP packages\", \"HTML/JavaScript trace visualization\"], \"modalities\": [\"screenshots\", \"view hierarchies\", \"action traces\", \"app metadata\"], \"notes\": \"Counts describe the card’s first-release claim, not an independently counted archive inventory. The card distinguishes globally image-hash-deduplicated screenshot/VH pairs from the November 2024 complete-app traces; trace state counts must not be added as new unique screens. JPEG/JSON and Parquet are alternative representations. Listed filename ranges are examples, not a complete count proof. No train/validation/test split is documented. Hub revision c734715e is ungated; MIT is its data declaration, without resolving third-party app content rights.\", \"splits\": [], \"supervision\": [\"paired screenshot/view-hierarchy structure\", \"recorded app states, actions and transitions\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/mllmTeam/MobileViews/tree/main\", \"gated\": false, \"hub_id\": \"mllmTeam/MobileViews\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/datasets/mllmTeam/MobileViews\nCatalog record: https://awesomecua.github.io/resources/url-cc4aa8caeee0cff58e73/"}
{"id": "url:cc931199bda2bdf1f5f7", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/tencent/POINTS-GUI-G", "citation_url": "https://awesomecua.github.io/resources/url-cc931199bda2bdf1f5f7/", "details": {"disabled": false, "download_url": "https://huggingface.co/tencent/POINTS-GUI-G/tree/main", "gated": false, "hub_id": "tencent/POINTS-GUI-G"}, "environments": ["web", "desktop", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "3b97405c0a7f50298cf31a4933d31da356363c2ba5ecfaf73144cedeebc44db2", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/tencent/POINTS-GUI-G"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "9f851237932688c15d261c016d51ff5ae80c43c5d0b4eaccac0b48d941118d80", "supports": ["environments"], "url": "https://raw.githubusercontent.com/Tencent/POINTS-GUI/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "ac679d3cfdda8754e945d127935c9ed9030e6c9d0c9872f6224909ab2abe5c3f", "supports": ["licenses.weights"], "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "other", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/tencent/POINTS-GUI-G"}], "name": "POINTS-GUI-G 8B", "publication": null, "published_at": null, "record_sha256": "e2909651e13a1b60517cf488c72f01a440cefcb7065aa5a510e708eb23fc4365", "relations": [{"evidence_url": "https://huggingface.co/tencent/POINTS-GUI-G/raw/main/README.md", "target_id": "arxiv:2602.06391", "type": "documented_by"}], "resource_id": "url:cc931199bda2bdf1f5f7", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/tencent/POINTS-GUI-G", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "POINTS-GUI-G 8B\nResource type: model\nReview status: curated\nScope: core\n\nA specialized interface-grounding checkpoint that returns normalized point coordinates for a requested element in a screenshot.\nPlatforms: web, desktop, cross-platform\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/tencent/POINTS-GUI-G/tree/main\", \"gated\": false, \"hub_id\": \"tencent/POINTS-GUI-G\"}\n\nFour safetensors shards and model code/configuration are public and ungated. The card declares license other without specifying a named weight license. This release is the grounding model; the same card says the end-to-end GUI agent model is still under development. No end-to-end agent capability is inferred from grounding scores.\n\nPrimary source: https://huggingface.co/tencent/POINTS-GUI-G\nCatalog record: https://awesomecua.github.io/resources/url-cc931199bda2bdf1f5f7/"}
{"id": "url:cccd967aa993404d74d0", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": ["Greg Little", "Tessa A. Lau", "Allen Cypher", "James Lin", "Eben M. Haber", "Eser Kandogan"], "canonical_source_url": "https://doi.org/10.1145/1240624.1240767", "citation_url": "https://awesomecua.github.io/resources/url-cccd967aa993404d74d0/", "details": {"publication": {"authors": ["Greg Little", "Tessa A. Lau", "Allen Cypher", "James Lin", "Eben M. Haber", "Eser Kandogan"], "date_basis": "Supplied year precision: 2007. CHI 2007 proceedings heading and copyright year The IBM page has an unlabeled Date of 22 Oct 2007, while the paper gives April/May event dates. Their roles do not establish a reliable day-level publication date.", "date_parts": [2007], "doi": "10.1145/1240624.1240767", "event_dates": null, "evidence_urls": ["https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web", "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record. Exact publication date is unresolved; only the clearly established year is proposed.", "pages": "943-946", "pdf_url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf", "publisher": null, "series": null, "title": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web", "type": "conference-paper", "url": "https://doi.org/10.1145/1240624.1240767", "venue": "CHI 2007 Proceedings", "volume": null}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "ae61a94cca68e058858244b6ff54ed9409973a5cecaf25c0e325b27adaf0e3dd", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "944b6f74a6cdb88fc253009f364dcc71cc42186b4907e04720da128c5292e357", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.venue"], "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:55:21Z", "role": "primary", "sha256": "944b6f74a6cdb88fc253009f364dcc71cc42186b4907e04720da128c5292e357", "supports": ["details.publication.authors", "details.publication.date_basis", "details.publication.date_parts", "details.publication.identity_basis", "details.publication.notes", "details.publication.pages", "details.publication.pdf_url", "details.publication.title", "details.publication.type", "details.publication.venue", "environments", "kind", "name", "scope", "summary", "topics"], "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T09:53:23Z", "role": "primary", "sha256": "ae61a94cca68e058858244b6ff54ed9409973a5cecaf25c0e325b27adaf0e3dd", "supports": ["details.publication.doi", "details.publication.identity_basis", "details.publication.notes", "details.publication.url"], "url": "https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web"}], "identifiers": {"doi": "10.1145/1240624.1240767"}, "kind": "paper", "last_verified_at": "2026-10-17T09:55:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://doi.org/10.1145/1240624.1240767"}, {"rel": "publication_pdf", "url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"}], "name": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web", "publication": {"authors": ["Greg Little", "Tessa A. Lau", "Allen Cypher", "James Lin", "Eben M. Haber", "Eser Kandogan"], "date_basis": "Supplied year precision: 2007. CHI 2007 proceedings heading and copyright year The IBM page has an unlabeled Date of 22 Oct 2007, while the paper gives April/May event dates. Their roles do not establish a reliable day-level publication date.", "date_parts": [2007], "doi": "10.1145/1240624.1240767", "event_dates": null, "evidence_urls": ["https://research.ibm.com/publications/koala-capture-share-automate-personalize-business-processes-on-the-web", "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf"], "identity_basis": "No same-work record located in the full active catalog. Similarity screening was not used as automatic identity approval. The raw archive and exclusions are additionally checked during normalization.", "issue": null, "notes": "This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record. Exact publication date is unresolved; only the clearly established year is proposed.", "pages": "943-946", "pdf_url": "https://plunk.org/eben/PublishedPapers/KoalaCHI.pdf", "publisher": null, "series": null, "title": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web", "type": "conference-paper", "url": "https://doi.org/10.1145/1240624.1240767", "venue": "CHI 2007 Proceedings", "volume": null}, "published_at": null, "record_sha256": "1c2bd6da124f938b166e6fa744e1884261656262fa15ce161d0e51e979aa01b4", "relations": [], "resource_id": "url:cccd967aa993404d74d0", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://doi.org/10.1145/1240624.1240767", "topics": ["web-navigation", "workflow", "recording"], "updated_at": null}, "text": "Koala: Capture, Share, Automate, Personalize Business Processes on the Web\nResource type: paper\nReview status: curated\nScope: core\n\nIntroduces Koala, a collaborative web automation system that records browser interactions as editable pseudo-natural-language scripts. Scripts are shared through a wiki and instantiated with personal or corporate data for reuse by other users.\nPlatforms: web\nTopics: web-navigation, workflow, recording\nRecord publication date (recorded precision): 2007\nAuthors: Greg Little, Tessa A. Lau, Allen Cypher, James Lin, Eben M. Haber, Eser Kandogan\n\nPublication edition: Koala: Capture, Share, Automate, Personalize Business Processes on the Web\nPublication authors: Greg Little, Tessa A. Lau, Allen Cypher, James Lin, Eben M. Haber, Eser Kandogan\nPublication edition date (recorded precision): 2007\nPublication venue: CHI 2007 Proceedings\nPublication pages: 943-946\nPublication doi: 10.1145/1240624.1240767\nPublication date basis: Supplied year precision: 2007. CHI 2007 proceedings heading and copyright year The IBM page has an unlabeled Date of 22 Oct 2007, while the paper gives April/May event dates. Their roles do not establish a reliable day-level publication date.\nPublication notes: This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record. Exact publication date is unresolved; only the clearly established year is proposed.\nPublication source: https://doi.org/10.1145/1240624.1240767\n\nThe annotation was reviewed against retained primary publication sources. Experiments have not been independently reproduced. This proposal is the specific Koala paper. It does not automatically merge every later CoScripter publication or project record.\n\nPrimary source: https://doi.org/10.1145/1240624.1240767\nCatalog record: https://awesomecua.github.io/resources/url-cccd967aa993404d74d0/"}
{"id": "url:cee45b940d5fbff3ccca", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/microsoft/Phi-Ground", "citation_url": "https://awesomecua.github.io/resources/url-cee45b940d5fbff3ccca/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/Phi-Ground/tree/main", "gated": false, "hub_id": "microsoft/Phi-Ground", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "ce8e0ebc3d4c2642bfb6a1d807176eae4c14b52fbc6f221b3c234a9d55e0d362", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/Phi-Ground"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/Phi-Ground"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/Phi-Ground"}], "name": "Phi-Ground checkpoint", "publication": null, "published_at": null, "record_sha256": "470036d975247f214d7071368ea551bff92e43ea85d236770a8d4ebcc9ababf9", "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/Phi-Ground", "target_id": "github:microsoft/phi-ground", "type": "variant_of"}], "resource_id": "url:cee45b940d5fbff3ccca", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/microsoft/Phi-Ground", "topics": ["grounding"], "updated_at": null}, "text": "Phi-Ground checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nMicrosoft Phi-Ground GUI grounding checkpoint fine-tuned from Phi-3.5-vision-instruct, shipped as two PyTorch weight shards with custom Phi-3-V configuration and processing code under MIT.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/microsoft/Phi-Ground/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/Phi-Ground\"}\n\nPrimary source: https://huggingface.co/microsoft/Phi-Ground\nCatalog record: https://awesomecua.github.io/resources/url-cee45b940d5fbff3ccca/"}
{"id": "url:cf4d2d45a7314cb16acb", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/GD-ML/AndroidCode", "citation_url": "https://awesomecua.github.io/resources/url-cf4d2d45a7314cb16acb/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "population": "AndroidCode corpus described by Code2World README; released JSON rows not recounted", "qualifier": "more-than", "unit": "screen–action pairs", "value": 80000}], "evidence_urls": ["https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "https://huggingface.co/api/datasets/GD-ML/AndroidCode"], "formats": ["HTML", "JSON", "Python extraction/loading scripts"], "modalities": ["GUI screenshots (described input)", "GUI actions", "HTML"], "notes": "The manifest lists 58,953 HTML files, which are not converted into transition/example counts. The reported >80k pairs and the HTML-file inventory have different units. The card itself has no sample schema or count; archive/image completeness and JSON contents were not audited.", "splits": ["Android-Control train file", "Android-Control test file"], "supervision": ["action-conditioned next-state code prediction", "generated HTML supervision"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/GD-ML/AndroidCode/tree/af316a6be79bbe08c4fc09ea588b5f80f177ed25", "gated": false, "hub_id": "GD-ML/AndroidCode", "hub_revision": "af316a6be79bbe08c4fc09ea588b5f80f177ed25", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/AMAP-ML/Code2World"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "26c14a14166d3b747454d1924ecdf6d96652ac857ae6ad488630dc843036a777", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:22Z", "role": "primary", "sha256": "5774a3c096264fbba58b7a16587ea65f35628edb9ca39afe8a32c74ec7fd1ce7", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/GD-ML/AndroidCode"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "d8d7a46d41a1a37fe4f0a5f637bf55c649310185329127d8a2204632e480be17", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:27:14Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/GD-ML/AndroidCode"}], "name": "AndroidCode", "publication": null, "published_at": null, "record_sha256": "a471e5d82586b3949d32bacd90784bab31770ffec177417b0a0592cb0d11e19b", "relations": [{"evidence_url": "https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md", "target_id": "arxiv:2602.09856", "type": "documented_by"}], "resource_id": "url:cf4d2d45a7314cb16acb", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/GD-ML/AndroidCode", "topics": ["world-models", "data-generation", "screen-understanding"], "updated_at": null}, "text": "AndroidCode\nResource type: dataset\nReview status: curated\nScope: core\n\nHTML and annotation release used for Code2World’s action-conditioned GUI-state prediction training, constructed from mobile GUI trajectories.\nPlatforms: android\nTopics: world-models, data-generation, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md\", \"population\": \"AndroidCode corpus described by Code2World README; released JSON rows not recounted\", \"qualifier\": \"more-than\", \"unit\": \"screen–action pairs\", \"value\": 80000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/AMAP-ML/Code2World/HEAD/readme.md\", \"https://huggingface.co/api/datasets/GD-ML/AndroidCode\"], \"formats\": [\"HTML\", \"JSON\", \"Python extraction/loading scripts\"], \"modalities\": [\"GUI screenshots (described input)\", \"GUI actions\", \"HTML\"], \"notes\": \"The manifest lists 58,953 HTML files, which are not converted into transition/example counts. The reported >80k pairs and the HTML-file inventory have different units. The card itself has no sample schema or count; archive/image completeness and JSON contents were not audited.\", \"splits\": [\"Android-Control train file\", \"Android-Control test file\"], \"supervision\": [\"action-conditioned next-state code prediction\", \"generated HTML supervision\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/GD-ML/AndroidCode/tree/af316a6be79bbe08c4fc09ea588b5f80f177ed25\", \"gated\": false, \"hub_id\": \"GD-ML/AndroidCode\", \"hub_revision\": \"af316a6be79bbe08c4fc09ea588b5f80f177ed25\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/GD-ML/AndroidCode/raw/af316a6be79bbe08c4fc09ea588b5f80f177ed25/README.md\"], \"reason\": \"Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.\"}}, \"repository_url\": \"https://github.com/AMAP-ML/Code2World\"}\n\nThe Hub card is license-only; semantic description and reported scale come from the official Code2World README. The public manifest contains HTML files, Android-Control train/test JSON files and extraction/loading scripts. It does not list standalone screenshot image files, and embedded/reference image packaging was not inspected. MIT is the data-card declaration, not a grant for every upstream source dataset.\n\nPrimary source: https://huggingface.co/datasets/GD-ML/AndroidCode\nCatalog record: https://awesomecua.github.io/resources/url-cf4d2d45a7314cb16acb/"}
{"id": "url:cfea8a0949472a637990", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/cckevinn/SeeClick", "citation_url": "https://awesomecua.github.io/resources/url-cfea8a0949472a637990/", "details": {"disabled": false, "download_url": "https://huggingface.co/cckevinn/SeeClick/tree/main", "gated": false, "hub_id": "cckevinn/SeeClick", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web", "desktop", "android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "c4112415bec58fc77f9a1e8fa98f8d602c5d382747579dfda9f866900c55995c", "supports": ["summary", "environments", "relations", "licenses"], "url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:05:52Z", "role": "primary", "sha256": "bc9f7c36f3e7c4515fa4982b0dd921a9038ff082cb6b283572d6390ef651c2a9", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/cckevinn/SeeClick"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:52Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/cckevinn/SeeClick"}], "name": "SeeClick checkpoint", "publication": null, "published_at": null, "record_sha256": "73aa2c1a9061df774bc64918dc7403fab01d152089a3a709912251fc2bb5e000", "relations": [{"evidence_url": "https://raw.githubusercontent.com/njucckevin/SeeClick/HEAD/readme.md", "target_id": "arxiv:2401.10935", "type": "documented_by"}], "resource_id": "url:cfea8a0949472a637990", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/cckevinn/SeeClick", "topics": ["grounding"], "updated_at": null}, "text": "SeeClick checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nThe released SeeClick visual grounding checkpoint for mapping an interface screenshot and referring instruction to a point or bounding box.\nPlatforms: web, desktop, android, ios\nTopics: grounding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/cckevinn/SeeClick/tree/main\", \"gated\": false, \"hub_id\": \"cckevinn/SeeClick\"}\n\nThe public ungated Hub manifest contains ten safetensors shards and model configuration/code but no README model card or explicit weight-license field. The official repository identifies the checkpoint and says incorporated checkpoints follow their original licenses, without resolving a specific standalone weight license. No source or model code was executed.\n\nPrimary source: https://huggingface.co/cckevinn/SeeClick\nCatalog record: https://awesomecua.github.io/resources/url-cfea8a0949472a637990/"}
{"id": "url:d0300db5ba754b1a64fa", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/wuuuuuz/GUI-CIDER", "citation_url": "https://awesomecua.github.io/resources/url-d0300db5ba754b1a64fa/", "details": {"availability": "Ungated JSONL artifact manifest; no dataset card or license declaration.", "dataset_profile": {"counts": [], "evidence_urls": ["https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md", "https://huggingface.co/api/datasets/wuuuuuz/GUI-CIDER"], "formats": ["JSONL"], "modalities": ["text"], "notes": "Purpose is established by the official project, not by API tags. No sample records were downloaded; count and data-license fields remain unknown.", "splits": [], "supervision": ["synthesized static planning knowledge", "synthesized dynamic causal knowledge"]}, "gated": false, "hub_id": "wuuuuuz/GUI-CIDER", "release_revision": "47ba5a55ed9ad993ad9562904ea6e978c4524d17"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:18:20Z", "role": "primary", "sha256": "9a46741bbde1373ed4b87887cdadde98e05ea8795d34efe39fc25d69b4d50213", "supports": ["name", "kind", "summary", "topics", "scope", "notes", "relations", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:23:07Z", "role": "primary", "sha256": "aec6da0e872092d737953f8b4db227867024da10e127cd32d15d05ff95552d45", "supports": ["identity", "release_status", "details.hub_id", "details.gated", "details.release_revision"], "url": "https://huggingface.co/api/datasets/wuuuuuz/GUI-CIDER"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T11:24:47Z", "role": "primary", "sha256": "bffca456e649a08239f6a1a3be10a7d7c21bcb427e783ba2c304e4b8410d385a", "supports": ["environments", "details.dataset_profile"], "url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T11:18:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/wuuuuuz/GUI-CIDER"}], "name": "GUI-CIDER causal knowledge data", "publication": null, "published_at": null, "record_sha256": "bdcbd0be7e8ae8ab1de338d02bef15b8f11fb8f8c9eb3d6e59d9db5c6d7e10d7", "relations": [{"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "github:wuzheng02/gui-cider", "type": "component_of"}, {"evidence_url": "https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md", "target_id": "arxiv:2605.28534", "type": "documented_by"}], "resource_id": "url:d0300db5ba754b1a64fa", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/wuuuuuz/GUI-CIDER", "topics": ["data-generation", "imitation-learning", "planning"], "updated_at": null}, "text": "GUI-CIDER causal knowledge data\nResource type: dataset\nReview status: curated\nScope: core\n\nA published JSONL corpus of synthesized planning and causal knowledge for GUI-agent mid-training, linked directly from the official GUI-CIDER project.\nPlatforms: android\nTopics: data-generation, imitation-learning, planning\n\nRecorded specifications: {\"availability\": \"Ungated JSONL artifact manifest; no dataset card or license declaration.\", \"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/README.md\", \"https://raw.githubusercontent.com/Wuzheng02/GUI-CIDER/HEAD/data/dataset.md\", \"https://huggingface.co/api/datasets/wuuuuuz/GUI-CIDER\"], \"formats\": [\"JSONL\"], \"modalities\": [\"text\"], \"notes\": \"Purpose is established by the official project, not by API tags. No sample records were downloaded; count and data-license fields remain unknown.\", \"splits\": [], \"supervision\": [\"synthesized static planning knowledge\", \"synthesized dynamic causal knowledge\"]}, \"gated\": false, \"hub_id\": \"wuuuuuz/GUI-CIDER\", \"release_revision\": \"47ba5a55ed9ad993ad9562904ea6e978c4524d17\"}\n\nThe official project explicitly links this corpus, and the ungated Hub API lists GUI-CIDER.jsonl. The repository has no README card at the Hub (404); purpose and Android context are therefore supported by the project README and dataset guide, while the API supports only identity/file availability. No sample count, split inventory, or data license was established. The project README’s MIT badge is not transferred to this data artifact.\n\nPrimary source: https://huggingface.co/datasets/wuuuuuz/GUI-CIDER\nCatalog record: https://awesomecua.github.io/resources/url-d0300db5ba754b1a64fa/"}
{"id": "url:d1830de02226155ee16d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI", "citation_url": "https://awesomecua.github.io/resources/url-d1830de02226155ee16d/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/tree/main", "gated": false, "hub_id": "Miaosen/Grounding-Anything-GUI", "languages": ["en"], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "31b964cfd198f99491f0f480024b2322fbdf3a237f7806847e83978109bc5856", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "a0609394a736dc495d077cd57a2945b018123e5b3f9514d7d5bafb9e690aeae2", "supports": ["summary", "release_status", "relations"], "url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "c64f72eae3dd50d28996f99ab744a5093896e8bb29caa6f7b33a2f60a0b530a5", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/Miaosen/Grounding-Anything-GUI"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI"}], "name": "Grounding-Anything-GUI", "publication": null, "published_at": null, "record_sha256": "a4edcb092549b01649dfd3d1111dc19fa2bb51cf4da7f098a377858b380f7ff0", "relations": [{"evidence_url": "https://raw.githubusercontent.com/microsoft/Phi-Ground/HEAD/README.md", "target_id": "github:microsoft/phi-ground", "type": "component_of"}], "resource_id": "url:d1830de02226155ee16d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "Grounding-Anything-GUI\nResource type: dataset\nReview status: curated\nScope: core\n\nThe public training archive linked by the Phi-Ground project for its Phi-Ground-Any model.\nPlatforms: unspecified\nTopics: grounding, data-generation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI/tree/main\", \"gated\": false, \"hub_id\": \"Miaosen/Grounding-Anything-GUI\", \"languages\": [\"en\"], \"modalities\": [\"image\", \"text\"]}\n\nThe Hub exposes train.zip and a card containing only license/language metadata. The official repository marks this training-data release complete, but its Phi-Ground-Any paper is still pending. No binding to the earlier Phi-Ground paper, sample count, split structure or operating-system coverage is inferred.\n\nPrimary source: https://huggingface.co/datasets/Miaosen/Grounding-Anything-GUI\nCatalog record: https://awesomecua.github.io/resources/url-d1830de02226155ee16d/"}
{"id": "url:d1e85aeb6d96859c109d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b", "citation_url": "https://awesomecua.github.io/resources/url-d1e85aeb6d96859c109d/", "details": {"disabled": false, "download_url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b/tree/main", "gated": false, "hub_id": "zai-org/webrl-orm-llama-3.1-8b", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:29Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["identity"], "url": "https://huggingface.co/api/models/THUDM/webrl-orm-llama-3.1-8b"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:06Z", "role": "primary", "sha256": "eb217b96c54c3c2ccfd3a27255c6fd740df40c2d6743ae811858339bf7f740b0", "supports": ["summary", "environments", "relations"], "url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T02:05:40Z", "role": "primary", "sha256": "8bafcdf63fe484b6005ca7082da1310459f8c87d4c011e84293a3cc8fca6ae9e", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/webrl-orm-llama-3.1-8b"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:05:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b"}], "name": "WebRL outcome reward model 8B", "publication": null, "published_at": null, "record_sha256": "b2d081e1f0f1f66479b8361b22dbd3a1d93a439bba572c6165dc1c2070ef0be0", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/WebRL/HEAD/README.md", "target_id": "arxiv:2411.02337", "type": "documented_by"}], "resource_id": "url:d1e85aeb6d96859c109d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b", "topics": ["reward-models", "reinforcement-learning", "web-navigation"], "updated_at": null}, "text": "WebRL outcome reward model 8B\nResource type: model\nReview status: curated\nScope: core\n\nThe outcome-supervision checkpoint used by WebRL to score browser trajectories during policy training.\nPlatforms: web\nTopics: reward-models, reinforcement-learning, web-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b/tree/main\", \"gated\": false, \"hub_id\": \"zai-org/webrl-orm-llama-3.1-8b\"}\n\nThe official API resolves the former THUDM namespace to zai-org and lists four safetensors shards. No model card or weight-license declaration is present in the inspected manifest. The official WebRL README identifies this as its outcome-supervised reward model; it is a training component rather than an action policy.\n\nPrimary source: https://huggingface.co/zai-org/webrl-orm-llama-3.1-8b\nCatalog record: https://awesomecua.github.io/resources/url-d1e85aeb6d96859c109d/"}
{"id": "url:d236e7ec7a2b0028950a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual", "citation_url": "https://awesomecua.github.io/resources/url-d236e7ec7a2b0028950a/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual/tree/main", "gated": false, "hub_id": "zai-org/AutoGLM-Phone-9B-Multilingual", "languages": ["zh"], "modalities": [], "size_category_reported": null}, "environments": ["android", "ios"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "d66bc86d9fb61f88d1456b36cb0c2c9ab5e178b55bc9d34f307b35733f9b90bc", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "d66bc86d9fb61f88d1456b36cb0c2c9ab5e178b55bc9d34f307b35733f9b90bc", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual"}], "name": "AutoGLM Phone 9B Multilingual", "publication": null, "published_at": null, "record_sha256": "4f4fc948f5d9a63bea2cc8116293e025c4c81ac7080e1d26af01ace053cdd6a8", "relations": [{"evidence_url": "https://huggingface.co/api/models/zai-org/AutoGLM-Phone-9B-Multilingual", "target_id": "github:zai-org/open-autoglm", "type": "variant_of"}], "resource_id": "url:d236e7ec7a2b0028950a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual", "topics": ["mobile-navigation"], "updated_at": null}, "text": "AutoGLM Phone 9B Multilingual\nResource type: model\nReview status: curated\nScope: core\n\nMultilingual variant of the AutoGLM phone-agent checkpoint for Open-AutoGLM, likewise fine-tuned from GLM-4.1V-9B-Base with about 10.3B BF16 parameters.\nPlatforms: android, ios\nTopics: mobile-navigation\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual/tree/main\", \"gated\": false, \"hub_id\": \"zai-org/AutoGLM-Phone-9B-Multilingual\", \"languages\": [\"zh\"]}\n\nPrimary source: https://huggingface.co/zai-org/AutoGLM-Phone-9B-Multilingual\nCatalog record: https://awesomecua.github.io/resources/url-d236e7ec7a2b0028950a/"}
{"id": "url:d3414794302313051ff0", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://manus.im/", "citation_url": "https://awesomecua.github.io/resources/url-d3414794302313051ff0/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "832b84c366eaadc69e328f207f2659cf07a0901c47662244831c0f1695d02efd", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://manus.im/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://manus.im/"}], "name": "Manus", "publication": null, "published_at": null, "record_sha256": "9be357e03e9f58b36f872ab1410fb3f106ac73f52e52898ee28834782e408eb4", "relations": [{"evidence_url": "https://manus.im/", "target_id": "organization:manus", "type": "published_by"}], "resource_id": "url:d3414794302313051ff0", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://manus.im/", "topics": ["workflow"], "updated_at": null}, "text": "Manus\nResource type: product\nReview status: curated\nScope: core\n\nGeneral agent platform with browser-based task execution.\nPlatforms: web\nTopics: workflow\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://manus.im/\nCatalog record: https://awesomecua.github.io/resources/url-d3414794302313051ff0/"}
{"id": "url:d35224f787e7ec21ce36", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B", "citation_url": "https://awesomecua.github.io/resources/url-d35224f787e7ec21ce36/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-9B", "gated": false, "hub_revision": "68ebbac4d5bdf725103a22d076122b8aa883a0f2", "listed_weight_files": ["model.safetensors"]}, "environments": ["android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "09c59e6b35a5fe1eea528595f326546269ea8a2a75aa404f1f7950b9de5f30d1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "1aab6befd309e54ba71676670039b4098362974ce0ce89b25872f137d50c103f", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/OpenMobile-2/OpenMobile-2-9B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:46:06Z", "role": "primary", "sha256": "9b385e9f6aa2f7169f616a76887df3b76df605cb2bea071d99ee7ca3f9b659cc", "supports": ["environments"], "url": "https://os-copilot.github.io/OpenMobile2-Home/"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:51Z", "role": "primary", "sha256": "09c59e6b35a5fe1eea528595f326546269ea8a2a75aa404f1f7950b9de5f30d1", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:51Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:51Z", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B"}], "name": "OpenMobile-2 9B", "publication": null, "published_at": null, "record_sha256": "85f38b39e6600b20e8aaee1930ae7337f1460c890f674c37cabff7cfd39b284c", "relations": [{"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "github:openmobile2/openmobile2-code", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "url:546e6c95825d1318c98d", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B/raw/main/README.md", "target_id": "arxiv:2610.14157", "type": "documented_by"}], "resource_id": "url:d35224f787e7ec21ce36", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B", "topics": ["mobile-navigation", "tool-integration"], "updated_at": null}, "text": "OpenMobile-2 9B\nResource type: model\nReview status: curated\nScope: core\n\nMobile-agent checkpoint trained on OpenMobile-Data-v2 for screenshot actions and app-native tool calls.\nPlatforms: android, web\nTopics: mobile-navigation, tool-integration\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.5-9B\", \"gated\": false, \"hub_revision\": \"68ebbac4d5bdf725103a22d076122b8aa883a0f2\", \"listed_weight_files\": [\"model.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.\n\nPrimary source: https://huggingface.co/OpenMobile-2/OpenMobile-2-9B\nCatalog record: https://awesomecua.github.io/resources/url-d35224f787e7ec21ce36/"}
{"id": "url:d3ec5122a2b52fd5ae4b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data", "citation_url": "https://awesomecua.github.io/resources/url-d3ec5122a2b52fd5ae4b/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md", "https://huggingface.co/api/datasets/OS-Copilot/OS-Genesis-mobile-data", "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"], "formats": ["JSONL", "ZIP"], "modalities": ["screenshots", "text", "actions"], "notes": "AC and AW refer to distinct target training sets; no sample count is inferred from the two JSONL files or ZIP packaging.", "splits": ["AndroidControl training", "AndroidWorld training"], "supervision": ["mobile action trajectories"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/tree/main", "gated": false, "hub_id": "OS-Copilot/OS-Genesis-mobile-data"}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "de9de63e21e60e882277a6268e8153abb6c14aaf3e9a7f83e7f98ac14bf37170", "supports": ["name", "kind", "relations", "licenses.data"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "6a5e0c2baecc2dc8bf4466903798f5ecda07cdac917042bbd4cfbe54d895c11e", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/OS-Copilot/OS-Genesis-mobile-data"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "430e495ab250395d666854bd18b0fc29b819cd784cd7730215ef65e461834d0d", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T04:01:34Z", "role": "primary", "sha256": "de9de63e21e60e882277a6268e8153abb6c14aaf3e9a7f83e7f98ac14bf37170", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:34Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T04:01:34Z", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data"}], "name": "OS-Genesis mobile training data", "publication": null, "published_at": null, "record_sha256": "1b5cdf7c6ac1caee95c050edf9b6b94fd34e1fab93d0732885cb21d4df740a82", "relations": [{"evidence_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md", "target_id": "arxiv:2412.19723", "type": "documented_by"}], "resource_id": "url:d3ec5122a2b52fd5ae4b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data", "topics": ["data-generation", "mobile-navigation", "imitation-learning"], "updated_at": null}, "text": "OS-Genesis mobile training data\nResource type: dataset\nReview status: curated\nScope: core\n\nReverse-task-synthesis mobile trajectories released in separate AndroidControl and AndroidWorld training annotation files.\nPlatforms: android\nTopics: data-generation, mobile-navigation, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/raw/main/README.md\", \"https://huggingface.co/api/datasets/OS-Copilot/OS-Genesis-mobile-data\", \"https://raw.githubusercontent.com/OS-Copilot/OS-Genesis/HEAD/README.md\"], \"formats\": [\"JSONL\", \"ZIP\"], \"modalities\": [\"screenshots\", \"text\", \"actions\"], \"notes\": \"AC and AW refer to distinct target training sets; no sample count is inferred from the two JSONL files or ZIP packaging.\", \"splits\": [\"AndroidControl training\", \"AndroidWorld training\"], \"supervision\": [\"mobile action trajectories\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data/tree/main\", \"gated\": false, \"hub_id\": \"OS-Copilot/OS-Genesis-mobile-data\"}\n\nThe public manifest contains AC/AW JSONL training annotations and a screenshot archive. MIT is declared for the data in its own card, separately from Apache-2.0 declarations on some model checkpoints. The gated web-data release is not included in this asset.\n\nPrimary source: https://huggingface.co/datasets/OS-Copilot/OS-Genesis-mobile-data\nCatalog record: https://awesomecua.github.io/resources/url-d3ec5122a2b52fd5ae4b/"}
{"id": "url:d5140d438466c1e1348a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench", "citation_url": "https://awesomecua.github.io/resources/url-d5140d438466c1e1348a/", "details": {"benchmark_protocol": {"actions": ["success verdict", "process score"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "fara7b_om2w_browserbase split", "qualifier": "exact", "unit": "trajectories", "value": 106}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "internal split", "qualifier": "exact", "unit": "trajectories", "value": 154}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "fara7b_om2w_browserbase split", "qualifier": "exact", "unit": "annotation rows", "value": 215}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "population": "internal split", "qualifier": "exact", "unit": "annotation rows", "value": 154}], "evaluation": "Agreement metrics including Cohen’s kappa, accuracy and F1 against human annotations; review stages must be kept distinct.", "evidence_urls": ["https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"], "focus": ["verifier-human agreement", "process quality"], "notes": "Trajectory rows and reviewer rows have different grains. Internal task labels do not make the released Parquet files private. No payload data was downloaded.", "observations": ["screenshots", "action logs", "final answers"], "setting": "offline"}, "disabled": false, "download_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/tree/main", "gated": false, "hub_id": "microsoft/CUAVerifierBench", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"], "reason": "The benchmark's Hub dataset card declares MIT in its metadata header (license: mit)."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "44d7190b72ae489297f59d6fadff635ced6dd067c845c77b04f187a7b2d7bec7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "fb1fe8a9fd6e45fc82157c65c08d5c682d08bead5196308776200ef3c39c1e41", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/microsoft/CUAVerifierBench"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench"}], "name": "CUAVerifierBench", "publication": null, "published_at": null, "record_sha256": "0c70997de5f42f8afa6c9eb2f90fedcdd3abf02c9fc69ba374e1e66b99f01e38", "relations": [{"evidence_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md", "target_id": "arxiv:2604.06240", "type": "documented_by"}], "resource_id": "url:d5140d438466c1e1348a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/microsoft/CUAVerifierBench", "topics": ["evaluation", "reward-models", "web-navigation"], "updated_at": null}, "text": "CUAVerifierBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nHuman-annotated browser-agent trajectories for comparing automated task-success and process-quality verifiers against reviewer judgments.\nPlatforms: web\nTopics: evaluation, reward-models, web-navigation\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"success verdict\", \"process score\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\", \"population\": \"fara7b_om2w_browserbase split\", \"qualifier\": \"exact\", \"unit\": \"trajectories\", \"value\": 106}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\", \"population\": \"internal split\", \"qualifier\": \"exact\", \"unit\": \"trajectories\", \"value\": 154}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\", \"population\": \"fara7b_om2w_browserbase split\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 215}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\", \"population\": \"internal split\", \"qualifier\": \"exact\", \"unit\": \"annotation rows\", \"value\": 154}], \"evaluation\": \"Agreement metrics including Cohen’s kappa, accuracy and F1 against human annotations; review stages must be kept distinct.\", \"evidence_urls\": [\"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\"], \"focus\": [\"verifier-human agreement\", \"process quality\"], \"notes\": \"Trajectory rows and reviewer rows have different grains. Internal task labels do not make the released Parquet files private. No payload data was downloaded.\", \"observations\": [\"screenshots\", \"action logs\", \"final answers\"], \"setting\": \"offline\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/microsoft/CUAVerifierBench/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/CUAVerifierBench\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/microsoft/CUAVerifierBench/raw/main/README.md\"], \"reason\": \"The benchmark's Hub dataset card declares MIT in its metadata header (license: mit).\"}}}\n\nPublic Parquet configs separate trajectories from per-reviewer annotations. The card declares MIT. The two splits have different review stages: Online-Mind2Web includes blind and verifier-informed annotations, whereas the internal split is blind-only. This measures verifier agreement rather than agent task success.\n\nPrimary source: https://huggingface.co/datasets/microsoft/CUAVerifierBench\nCatalog record: https://awesomecua.github.io/resources/url-d5140d438466c1e1348a/"}
{"id": "url:d5667c854845c77b9f5b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/xuyifan/MobileRL-9B", "citation_url": "https://awesomecua.github.io/resources/url-d5667c854845c77b9f5b/", "details": {"disabled": false, "download_url": "https://huggingface.co/xuyifan/MobileRL-9B/tree/main", "gated": false, "hub_id": "xuyifan/MobileRL-9B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "a00f13f5a86f2de7eb9c866cc4d08b958d21b2ac30dc3b9e1662e4bb336b7cb3", "supports": ["name", "kind", "relations", "licenses.weights", "identity", "release_status", "details"], "url": "https://huggingface.co/api/models/xuyifan/MobileRL-9B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "5c2d5eb247958f1af0646cbac831ecfbc7f839baa236a0d733075592da81f58e", "supports": ["licenses.weights"], "url": "https://huggingface.co/xuyifan/MobileRL-9B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "752fad8ebd8b8eda03648329a6fc2585e5bcb2b96875b8aa2ce7d9f3fff78153", "supports": ["summary", "environments", "relations", "topics", "scope"], "url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:03:40Z", "role": "primary", "sha256": "a00f13f5a86f2de7eb9c866cc4d08b958d21b2ac30dc3b9e1662e4bb336b7cb3", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/xuyifan/MobileRL-9B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T04:03:40Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-17T04:03:40Z", "url": "https://huggingface.co/xuyifan/MobileRL-9B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/xuyifan/MobileRL-9B"}], "name": "MobileRL 9B checkpoint", "publication": null, "published_at": null, "record_sha256": "8b79391ee1ca0bceb0343da1e74d7eea63605128cfae7c928fd1aa6e11438eda", "relations": [{"evidence_url": "https://raw.githubusercontent.com/THUDM/MobileRL/HEAD/README.md", "target_id": "arxiv:2509.18119", "type": "documented_by"}], "resource_id": "url:d5667c854845c77b9f5b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/xuyifan/MobileRL-9B", "topics": ["reinforcement-learning", "mobile-navigation"], "updated_at": null}, "text": "MobileRL 9B checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nA GLM-4.1V-based mobile action checkpoint released by the MobileRL project for Android interaction.\nPlatforms: android\nTopics: reinforcement-learning, mobile-navigation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/xuyifan/MobileRL-9B/tree/main\", \"gated\": false, \"hub_id\": \"xuyifan/MobileRL-9B\", \"languages\": [\"en\"]}\n\nFive public safetensors shards are present. The short card declares Apache-2.0 and a GLM-4.1V-9B base. The official project confirms the checkpoint and Android targets, and distinguishes MobileRL bounding-box actions from coordinate-format alternatives.\n\nPrimary source: https://huggingface.co/xuyifan/MobileRL-9B\nCatalog record: https://awesomecua.github.io/resources/url-d5667c854845c77b9f5b/"}
{"id": "url:d70a755e3b2102f25f34", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B", "citation_url": "https://awesomecua.github.io/resources/url-d70a755e3b2102f25f34/", "details": {"disabled": false, "download_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/tree/main", "gated": false, "hub_id": "InfiX-ai/InfiGUI-G1-7B"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "10278eb0f1600557e432a0f8dde2c40cfffe12330ddb949b6a0fc02bfbc6c9ee", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/InfiX-ai/InfiGUI-G1-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "4b68c82a2dd0063a450bde34c8bffbd931565a91a2982a5b47be7e711815b284", "supports": ["environments"], "url": "https://raw.githubusercontent.com/InfiXAI/InfiGUI-G1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:29Z", "role": "primary", "sha256": "253055f0f77266a64d2db5de8fd04c4202f4cda3be733dd1b00d8565116dd240", "supports": ["licenses.weights"], "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:29Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:29Z", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B"}], "name": "InfiGUI-G1 7B", "publication": null, "published_at": null, "record_sha256": "c9ca5cc38d0212184d8ee266727fc5ae01639b5ae29e41bc4056ba9a56266600", "relations": [{"evidence_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/raw/main/README.md", "target_id": "arxiv:2508.05731", "type": "documented_by"}], "resource_id": "url:d70a755e3b2102f25f34", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/InfiX-ai/InfiGUI-G1-7B", "topics": ["grounding", "reinforcement-learning"], "updated_at": null}, "text": "InfiGUI-G1 7B\nResource type: model\nReview status: curated\nScope: core\n\nA GUI-grounding checkpoint trained with adaptive exploration rewards and multiple candidate answers to improve instruction-to-element alignment.\nPlatforms: cross-platform\nTopics: grounding, reinforcement-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/InfiX-ai/InfiGUI-G1-7B/tree/main\", \"gated\": false, \"hub_id\": \"InfiX-ai/InfiGUI-G1-7B\"}\n\nFour safetensors shards are public. The model card declares Apache-2.0 and identifies Qwen2.5-VL-7B-Instruct as its base. It is the G1 grounding checkpoint, distinct from the previously cataloged InfiGUI-R1 reasoning checkpoint; benchmark scores remain publisher claims.\n\nPrimary source: https://huggingface.co/InfiX-ai/InfiGUI-G1-7B\nCatalog record: https://awesomecua.github.io/resources/url-d70a755e3b2102f25f34/"}
{"id": "url:d7a4f34a6a5202af5f64", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench", "citation_url": "https://awesomecua.github.io/resources/url-d7a4f34a6a5202af5f64/", "details": {"benchmark_protocol": {"actions": ["mobile interface operations", "clarification requests"], "counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "population": "MobileGUI-VBench release", "qualifier": "exact", "unit": "tasks", "value": 150}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "population": "MobileGUI-VBench release", "qualifier": "exact", "unit": "mobile apps", "value": 40}], "evaluation": "Published task definitions include preconditions and clarification requirements; the metadata review does not independently validate the reported agent scores.", "evidence_urls": ["https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md"], "focus": ["mobile task completion", "clarification behavior", "cross-app tasks"], "notes": "The card describes 140 single-app and 10 cross-app tasks. Task and app counts are different units.", "observations": ["mobile app interfaces"], "setting": "interactive"}, "disabled": false, "download_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/tree/main", "gated": false, "hub_id": "vivo-ai/MobileGUI-VBench"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "422ff6a557a1aabd8f6c6ef26e5fd05b992f4acf0e5128ab8bc437021fcc5ea7", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "5850fc58f54c877cffe328dba2a0b5e3e98767637aae6d6efd9a037cf06eabce", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/vivo-ai/MobileGUI-VBench"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench"}], "name": "MobileGUI-VBench", "publication": null, "published_at": null, "record_sha256": "940b58772e7665f1510b33999f28976bf122abbcb58aa0960b1a841bd454274e", "relations": [{"evidence_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md", "target_id": "arxiv:2609.12394", "type": "documented_by"}], "resource_id": "url:d7a4f34a6a5202af5f64", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench", "topics": ["evaluation", "mobile-navigation", "cross-app"], "updated_at": null}, "text": "MobileGUI-VBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nMobile-app task definitions covering everyday scenarios, with annotated preconditions and clarification requirements for evaluating agent behavior.\nPlatforms: unspecified\nTopics: evaluation, mobile-navigation, cross-app\n\nRecorded specifications: {\"benchmark_protocol\": {\"actions\": [\"mobile interface operations\", \"clarification requests\"], \"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md\", \"population\": \"MobileGUI-VBench release\", \"qualifier\": \"exact\", \"unit\": \"tasks\", \"value\": 150}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md\", \"population\": \"MobileGUI-VBench release\", \"qualifier\": \"exact\", \"unit\": \"mobile apps\", \"value\": 40}], \"evaluation\": \"Published task definitions include preconditions and clarification requirements; the metadata review does not independently validate the reported agent scores.\", \"evidence_urls\": [\"https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/raw/main/README.md\"], \"focus\": [\"mobile task completion\", \"clarification behavior\", \"cross-app tasks\"], \"notes\": \"The card describes 140 single-app and 10 cross-app tasks. Task and app counts are different units.\", \"observations\": [\"mobile app interfaces\"], \"setting\": \"interactive\"}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench/tree/main\", \"gated\": false, \"hub_id\": \"vivo-ai/MobileGUI-VBench\"}\n\nPublic task files are available as JSONL and XLSX with documentation. The card explicitly declares Apache-2.0 for the dataset, despite having no license field in its YAML metadata. It identifies mobile apps without fixing an operating-system taxonomy; model scores are not imported.\n\nPrimary source: https://huggingface.co/datasets/vivo-ai/MobileGUI-VBench\nCatalog record: https://awesomecua.github.io/resources/url-d7a4f34a6a5202af5f64/"}
{"id": "url:d7ad447ecd712c6d51b8", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview", "citation_url": "https://awesomecua.github.io/resources/url-d7ad447ecd712c6d51b8/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview/tree/main", "gated": false, "hub_id": "stepfun-ai/GELab-Zero-4B-preview", "languages": ["en", "zh"], "modalities": [], "size_category_reported": null}, "environments": ["android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "cd1efa947316f13967165ac906400170df6d28446804ba0763a49fc21a731e7e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "cd1efa947316f13967165ac906400170df6d28446804ba0763a49fc21a731e7e", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview"}], "name": "GELab Zero 4B Preview", "publication": null, "published_at": null, "record_sha256": "596ca627ba5e4f686dd82b69e6974708eb0bbf69f5fbc4e4278414948a481af1", "relations": [{"evidence_url": "https://huggingface.co/api/models/stepfun-ai/GELab-Zero-4B-preview", "target_id": "github:stepfun-ai/gelab-zero", "type": "variant_of"}], "resource_id": "url:d7ad447ecd712c6d51b8", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview", "topics": ["mobile-navigation"], "updated_at": null}, "text": "GELab Zero 4B Preview\nResource type: model\nReview status: curated\nScope: core\n\nPreview GELab-Zero phone-use agent checkpoint fine-tuned from Qwen3-VL-4B-Instruct, with about 4.4B BF16 parameters, tagged for Android in English and Chinese, under Apache-2.0.\nPlatforms: android\nTopics: mobile-navigation\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview/tree/main\", \"gated\": false, \"hub_id\": \"stepfun-ai/GELab-Zero-4B-preview\", \"languages\": [\"en\", \"zh\"]}\n\nPrimary source: https://huggingface.co/stepfun-ai/GELab-Zero-4B-preview\nCatalog record: https://awesomecua.github.io/resources/url-d7ad447ecd712c6d51b8/"}
{"id": "url:d92596cfe618e5322ec2", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://www.tinyfish.ai/", "citation_url": "https://awesomecua.github.io/resources/url-d92596cfe618e5322ec2/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "461de0060fae1ded024dfa724624b001665b463b14f454afafebd0c9a2dce71d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://www.tinyfish.ai/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://www.tinyfish.ai/"}], "name": "TinyFish", "publication": null, "published_at": null, "record_sha256": "bc5aa1a7d69995222280f8a38318184e10c9b5fd5e96e92d17844bd19bed953c", "relations": [{"evidence_url": "https://www.tinyfish.ai/", "target_id": "organization:tinyfish", "type": "published_by"}], "resource_id": "url:d92596cfe618e5322ec2", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://www.tinyfish.ai/", "topics": ["web-navigation"], "updated_at": null}, "text": "TinyFish\nResource type: product\nReview status: curated\nScope: core\n\nBrowser-agent API for executing web tasks and extracting structured results.\nPlatforms: web\nTopics: web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://www.tinyfish.ai/\nCatalog record: https://awesomecua.github.io/resources/url-d92596cfe618e5322ec2/"}
{"id": "url:d9d8e68f19ef0956d189", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Qwen/RecreationBench", "citation_url": "https://awesomecua.github.io/resources/url-d9d8e68f19ef0956d189/", "details": {"leaderboard_url": "https://recreation-bench.cc/#leaderboard"}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "79deaa899d085ece77a127d53b12ddb9edea8a796943f35b75e9acec67bcb976", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/Qwen/RecreationBench/raw/main/README.md"}], "identifiers": {}, "kind": "benchmark", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Qwen/RecreationBench"}], "name": "RecreationBench", "publication": null, "published_at": null, "record_sha256": "41d01dd92025347a4421abe1a7135833fafcc37a27047dc28c0d2b1a66041e04", "relations": [], "resource_id": "url:d9d8e68f19ef0956d189", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Qwen/RecreationBench", "topics": ["evaluation", "tool-integration"], "updated_at": null}, "text": "RecreationBench\nResource type: benchmark\nReview status: curated\nScope: core\n\nTasks for recreating reference applications with hybrid computer interaction.\nPlatforms: cross-platform\nTopics: evaluation, tool-integration\n\nRecorded specifications: {\"leaderboard_url\": \"https://recreation-bench.cc/#leaderboard\"}\n\nDataset card declares multiple upstream licenses; do not treat the package as having one blanket permissive license.\n\nPrimary source: https://huggingface.co/datasets/Qwen/RecreationBench\nCatalog record: https://awesomecua.github.io/resources/url-d9d8e68f19ef0956d189/"}
{"id": "url:db1b7b7efe3755ae5693", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B", "citation_url": "https://awesomecua.github.io/resources/url-db1b7b7efe3755ae5693/", "details": {"disabled": false, "download_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/tree/main", "gated": false, "hub_id": "apurvaga/go-browse-wa-qwen-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "470ef6f975990518f9622b259a6b8e5761b26dcd22985c248f1df2beb4aa2760", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "d083bf53b20c6c6124f1df29a83ce783799d46a1ab5600d19d658729fc6d67bb", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/apurvaga/go-browse-wa-qwen-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "0d4b28c993ccb977c7910ff3fecb5de29db364a7a7ad55523c8e052c9bae2566", "supports": ["summary", "environments"], "url": "https://raw.githubusercontent.com/ApGa/Go-Browse/HEAD/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B"}], "name": "Go-Browse WebArena Qwen 7B", "publication": null, "published_at": null, "record_sha256": "b851f05c451bfa814b9fe4881bf6d372395903f06b7cde2cdd41c201330ef004", "relations": [{"evidence_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/raw/main/README.md", "target_id": "arxiv:2506.03533", "type": "documented_by"}], "resource_id": "url:db1b7b7efe3755ae5693", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/apurvaga/go-browse-wa-qwen-7B", "topics": ["web-navigation", "imitation-learning"], "updated_at": null}, "text": "Go-Browse WebArena Qwen 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen-2.5-7B-Instruct checkpoint fine-tuned on Go-Browse WebArena trajectories for website task execution.\nPlatforms: web\nTopics: web-navigation, imitation-learning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/apurvaga/go-browse-wa-qwen-7B/tree/main\", \"gated\": false, \"hub_id\": \"apurvaga/go-browse-wa-qwen-7B\"}\n\nSeven public safetensors shards and model/tokenizer files are present. The card links the official code and paper but declares no weight license. The code repository’s MIT license is not inherited by this checkpoint, and the separate NNetNav-trained comparison model is not treated as an alias.\n\nPrimary source: https://huggingface.co/apurvaga/go-browse-wa-qwen-7B\nCatalog record: https://awesomecua.github.io/resources/url-db1b7b7efe3755ae5693/"}
{"id": "url:db6fc5f29f186d13991c", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B", "citation_url": "https://awesomecua.github.io/resources/url-db6fc5f29f186d13991c/", "details": {"disabled": false, "download_url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/tree/2c5604d86c529843ef346698edcf33b170c1e817", "gated": false, "hub_id": "Fanbin/ARPO_UITARS1.5_7B", "hub_revision": "2c5604d86c529843ef346698edcf33b170c1e817", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/raw/2c5604d86c529843ef346698edcf33b170c1e817/README.md"], "reason": "The inspected model card contains no checkpoint-license declaration. The code repository Apache license is not transferred to weights."}}, "repository_url": "https://github.com/JIA-Lab-research/ARPO"}, "environments": ["desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:12Z", "role": "primary", "sha256": "839b2de0e0c614fe8d2df4782d0f6a6e7d33f5c60d8020c9d4a01c61d533f39e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations"], "url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:21Z", "role": "primary", "sha256": "6e19a5339c4e52ca3753cce26ac2034883b96f5cb26383a4b41bc87a5fbf5bf6", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url"], "url": "https://huggingface.co/api/models/Fanbin/ARPO_UITARS1.5_7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "b6957fad37a7995e9dde4c7d4a4e0f6e2153b85354b2d234587c2a94ff95d0bd", "supports": ["summary", "notes", "licenses.weights"], "url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/raw/2c5604d86c529843ef346698edcf33b170c1e817/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-18T04:27:12Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B"}], "name": "ARPO UI-TARS-1.5-7B checkpoint", "publication": null, "published_at": null, "record_sha256": "308e4ba3ee7ea6333f0f3360d4d20f73697e527334f63cb5b248d8fc503de5ad", "relations": [{"evidence_url": "https://raw.githubusercontent.com/JIA-Lab-research/ARPO/HEAD/README.md", "target_id": "arxiv:2505.16282", "type": "documented_by"}], "resource_id": "url:db6fc5f29f186d13991c", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B", "topics": ["reinforcement-learning", "action-prediction"], "updated_at": null}, "text": "ARPO UI-TARS-1.5-7B checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nUI-TARS-1.5-7B-derived checkpoint trained with ARPO on OSWorld desktop tasks.\nPlatforms: desktop, linux\nTopics: reinforcement-learning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/tree/2c5604d86c529843ef346698edcf33b170c1e817\", \"gated\": false, \"hub_id\": \"Fanbin/ARPO_UITARS1.5_7B\", \"hub_revision\": \"2c5604d86c529843ef346698edcf33b170c1e817\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B/raw/2c5604d86c529843ef346698edcf33b170c1e817/README.md\"], \"reason\": \"The inspected model card contains no checkpoint-license declaration. The code repository Apache license is not transferred to weights.\"}}, \"repository_url\": \"https://github.com/JIA-Lab-research/ARPO\"}\n\nThe public Hub manifest lists four Safetensors shards, an index and configuration/tokenizer files. The model card reports a 15-step evaluation setting; no performance result is independently reproduced. No weight license is declared, despite Apache-2.0 being established for the separate code repository.\n\nPrimary source: https://huggingface.co/Fanbin/ARPO_UITARS1.5_7B\nCatalog record: https://awesomecua.github.io/resources/url-db6fc5f29f186d13991c/"}
{"id": "url:de31b07ff24ab15d9bdf", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/PrentisAI/GroundGUI-8B", "citation_url": "https://awesomecua.github.io/resources/url-de31b07ff24ab15d9bdf/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3-VL-8B-Instruct", "gated": false, "hub_revision": "9db378189fddaabb5f6e347b9053858a392022ac", "listed_weight_files": ["model.safetensors"]}, "environments": ["desktop", "web", "cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "8b680b9a63e654fe0b43754201d30c91714859578f294f4781e035280e862657", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/PrentisAI/GroundGUI-8B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "73a4690fab49d3b08ea9b0b1e44d62d5ec1802614f5cf22c37ed92dca49aa520", "supports": ["licenses.weights"], "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/PrentisAI/GroundGUI-8B"}], "name": "GroundGUI-8B", "publication": null, "published_at": null, "record_sha256": "73cd66adb71a0a379b1131229cec334096cc8628873853c7250a55ca189e1676", "relations": [{"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "github:prentisai/groundgui", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "url:56bda04bd8d3a438deb3", "type": "trained_on"}, {"evidence_url": "https://huggingface.co/PrentisAI/GroundGUI-8B/raw/main/README.md", "target_id": "arxiv:2610.19100", "type": "documented_by"}], "resource_id": "url:de31b07ff24ab15d9bdf", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/PrentisAI/GroundGUI-8B", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "GroundGUI-8B\nResource type: model\nReview status: curated\nScope: core\n\nScreenshot grounding checkpoint that emits a single click in a normalized coordinate space.\nPlatforms: desktop, web, cross-platform\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3-VL-8B-Instruct\", \"gated\": false, \"hub_revision\": \"9db378189fddaabb5f6e347b9053858a392022ac\", \"listed_weight_files\": [\"model.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data.\n\nPrimary source: https://huggingface.co/PrentisAI/GroundGUI-8B\nCatalog record: https://awesomecua.github.io/resources/url-de31b07ff24ab15d9bdf/"}
{"id": "url:dec156cbf0cb1181f7b6", "metadata": {"aliases": ["Nova Act"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://aws.amazon.com/nova/act/", "citation_url": "https://awesomecua.github.io/resources/url-dec156cbf0cb1181f7b6/", "details": {}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T17:18:21Z", "role": "primary", "sha256": "4e4e60092351a85d0b99186a745d6295b170958be59092a786eda369923bd691", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations"], "url": "https://aws.amazon.com/nova/act/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-17T17:18:21Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://aws.amazon.com/nova/act/"}], "name": "Amazon Nova Act", "publication": null, "published_at": null, "record_sha256": "a3226534985d7547f22d7cf2605633813c0b622155be78467ade060c87d62534", "relations": [], "resource_id": "url:dec156cbf0cb1181f7b6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://aws.amazon.com/nova/act/", "topics": ["automation", "web-navigation"], "updated_at": null}, "text": "Amazon Nova Act\nResource type: product\nReview status: curated\nScope: core\n\nAWS service for building and deploying agents that automate browser-based UI workflows from natural language and Python, powered by a custom Nova 2 Lite model.\nRecorded aliases: Nova Act\nPlatforms: web\nTopics: automation, web-navigation\n\nPrimary source: https://aws.amazon.com/nova/act/\nCatalog record: https://awesomecua.github.io/resources/url-dec156cbf0cb1181f7b6/"}
{"id": "url:df680385f18b01a825be", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/ritzzai/GUI-R1", "citation_url": "https://awesomecua.github.io/resources/url-df680385f18b01a825be/", "details": {"disabled": false, "download_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/tree/main", "gated": false, "hub_id": "ritzzai/GUI-R1"}, "environments": ["desktop", "windows", "linux", "macos", "android", "web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "f31ae9547d94967dd91f5cde5e5c66430d7d59ee3a525c0914bc600a79999af7", "supports": ["identity", "release_status", "details", "licenses.data"], "url": "https://huggingface.co/api/datasets/ritzzai/GUI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:43:11Z", "role": "primary", "sha256": "94c7c78bede11dde727940f56592a04594c1c59438812dff47540bf53c1bed7d", "supports": ["summary", "environments", "release_status"], "url": "https://raw.githubusercontent.com/ritzz-ai/GUI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:42:32Z", "role": "primary", "sha256": "7b14996d1a99675990cb2f3aeada2b2f9898f9daee46d3680988e16f6365abe3", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T02:42:32Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:42:32Z", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/ritzzai/GUI-R1"}], "name": "GUI-R1 data bundle", "publication": null, "published_at": null, "record_sha256": "87bea7e845716ce0a38163a6b6e10c00eaf42ba0e854676d9b5e6119af5ed2fd", "relations": [{"evidence_url": "https://huggingface.co/datasets/ritzzai/GUI-R1/raw/main/README.md", "target_id": "arxiv:2504.10458", "type": "documented_by"}], "resource_id": "url:df680385f18b01a825be", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/ritzzai/GUI-R1", "topics": ["grounding", "reinforcement-learning", "evaluation"], "updated_at": null}, "text": "GUI-R1 data bundle\nResource type: dataset\nReview status: curated\nScope: core\n\nTraining and evaluation Parquet bundles accompanying GUI-R1, including GUI grounding and mobile/web action-prediction data.\nPlatforms: desktop, windows, linux, macos, android, web\nTopics: grounding, reinforcement-learning, evaluation\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/ritzzai/GUI-R1/tree/main\", \"gated\": false, \"hub_id\": \"ritzzai/GUI-R1\"}\n\nThe Hub is a dataset repository with Parquet payloads, although the short card mistakenly calls it a model. The official project identifies it as training/evaluation data and describes subsequent expanded mixtures. MIT is declared in this dataset card; upstream dataset terms remain separate. No 3K/800K aggregate count is inferred for the mutable bundle.\n\nPrimary source: https://huggingface.co/datasets/ritzzai/GUI-R1\nCatalog record: https://awesomecua.github.io/resources/url-df680385f18b01a825be/"}
{"id": "url:e062cf1445eaee0f3ade", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/xlangai/Jedi", "citation_url": "https://awesomecua.github.io/resources/url-e062cf1445eaee0f3ade/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"], "formats": [], "modalities": ["images", "text", "coordinates"], "notes": "Coordinates were resized for a specified Qwen2.5-VL image processor and must remain aligned with image resizing. AGUVIS++ material requires images from upstream AGUVIS and OS-Atlas releases; this repository is not a self-contained copy of every referenced image collection.", "splits": [], "supervision": ["icon grounding", "component grounding", "layout grounding", "refusal"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/xlangai/Jedi/tree/main", "gated": false, "hub_id": "xlangai/Jedi", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:38Z", "role": "primary", "sha256": "1d4d15c0cd27837602d55f0fda1ed91d9e8a98b33b3e3c4ba64decae42399558", "supports": ["details.dataset_profile"], "url": "https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T09:33:18Z", "role": "primary", "sha256": "63ff72bea0a2cb9218093d7a5a420b39c01861aef66141433d7c56c1e0ad05ad", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/xlangai/Jedi"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T09:33:18Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T09:33:18Z", "url": "https://huggingface.co/datasets/xlangai/Jedi"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/xlangai/Jedi"}], "name": "Jedi", "publication": null, "published_at": null, "record_sha256": "2f5d2ee22b49447d5201f9d158e413371aa9e9d191870b0627339e0e067f0bda", "relations": [{"evidence_url": "https://huggingface.co/api/datasets/xlangai/Jedi", "target_id": "arxiv:2505.13227", "type": "documented_by"}], "resource_id": "url:e062cf1445eaee0f3ade", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/xlangai/Jedi", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "Jedi\nResource type: dataset\nReview status: curated\nScope: core\n\nSynthetic GUI grounding data associated with UI decomposition research and the OSWorld-G benchmark.\nPlatforms: desktop\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/xlangai/Jedi/raw/main/README.md\"], \"formats\": [], \"modalities\": [\"images\", \"text\", \"coordinates\"], \"notes\": \"Coordinates were resized for a specified Qwen2.5-VL image processor and must remain aligned with image resizing. AGUVIS++ material requires images from upstream AGUVIS and OS-Atlas releases; this repository is not a self-contained copy of every referenced image collection.\", \"splits\": [], \"supervision\": [\"icon grounding\", \"component grounding\", \"layout grounding\", \"refusal\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/xlangai/Jedi/tree/main\", \"gated\": false, \"hub_id\": \"xlangai/Jedi\"}\n\nThe dataset card requires coordinate scaling to stay aligned with the image processor. Follow the publisher recipe before using the grounding labels.\n\nPrimary source: https://huggingface.co/datasets/xlangai/Jedi\nCatalog record: https://awesomecua.github.io/resources/url-e062cf1445eaee0f3ade/"}
{"id": "url:e0ba328ae1a643b0a5aa", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0", "citation_url": "https://awesomecua.github.io/resources/url-e0ba328ae1a643b0a5aa/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "question–answer examples", "value": 307315}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 train", "qualifier": "exact", "unit": "web pages", "value": 4549}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "question–answer examples", "value": 52826}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 dev", "qualifier": "exact", "unit": "web pages", "value": 913}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "question–answer examples", "value": 40357}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "population": "WebSRC v1.0 test", "qualifier": "exact", "unit": "web pages", "value": 985}], "evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0", "https://x-lance.github.io/WebSRC/"], "formats": ["CSV", "HTML", "JSON", "ZIP packages"], "modalities": ["questions and answers", "HTML", "screenshots", "layout metadata"], "notes": "Each dataset.csv row is a QA example; web pages and QA examples are different units. Use the v1.0 split table rather than merging the older project headline into it. The card retains email-evaluation instructions, while the manifest now includes test answers; no submission was made. Payload archives were not downloaded.", "splits": ["train", "dev", "test"], "supervision": ["extractive or yes/no answers", "answer DOM-element references"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/tree/7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "gated": false, "hub_id": "X-LANCE/WebSRC_v1.0", "hub_revision": "7aa0bc6efc7ef43f68c192e2091108541acbaf1a", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:16Z", "role": "primary", "sha256": "061e5bcaa49e819f0133afb912ad5ce20a8c2b6c28ca60319a4635d6ce835ddb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "a7976dd931c3fcf11de2456614f95eb071978599bd70fb89c0ea009d31da4300", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T15:37:04Z", "role": "primary", "sha256": "3c2c601d8b4ad2f379cac9b97a2cbb901883d9bc6c1a682359f9b4135e6dd4bc", "supports": ["notes", "details.dataset_profile"], "url": "https://x-lance.github.io/WebSRC/"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:16Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "CC-BY-4.0", "retrieved_at": "2026-10-18T04:35:16Z", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0"}], "name": "WebSRC v1.0", "publication": null, "published_at": null, "record_sha256": "03568fb97933ba5bac66f51cb33d8c144be25d39237c2acaf8ead35e9caecc31", "relations": [{"evidence_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md", "target_id": "arxiv:2101.09465", "type": "documented_by"}], "resource_id": "url:e0ba328ae1a643b0a5aa", "review_status": "curated", "schema_version": 1, "scope": "adjacent", "source_url": "https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0", "topics": ["screen-understanding", "grounding"], "updated_at": null}, "text": "WebSRC v1.0\nResource type: dataset\nReview status: curated\nScope: adjacent\n\nWebpage segments, questions, answers, HTML, screenshots and layout metadata for structural reading comprehension.\nPlatforms: web\nTopics: screen-understanding, grounding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 train\", \"qualifier\": \"exact\", \"unit\": \"question–answer examples\", \"value\": 307315}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 train\", \"qualifier\": \"exact\", \"unit\": \"web pages\", \"value\": 4549}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 dev\", \"qualifier\": \"exact\", \"unit\": \"question–answer examples\", \"value\": 52826}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 dev\", \"qualifier\": \"exact\", \"unit\": \"web pages\", \"value\": 913}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 test\", \"qualifier\": \"exact\", \"unit\": \"question–answer examples\", \"value\": 40357}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"population\": \"WebSRC v1.0 test\", \"qualifier\": \"exact\", \"unit\": \"web pages\", \"value\": 985}], \"evidence_urls\": [\"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\", \"https://huggingface.co/api/datasets/X-LANCE/WebSRC_v1.0\", \"https://x-lance.github.io/WebSRC/\"], \"formats\": [\"CSV\", \"HTML\", \"JSON\", \"ZIP packages\"], \"modalities\": [\"questions and answers\", \"HTML\", \"screenshots\", \"layout metadata\"], \"notes\": \"Each dataset.csv row is a QA example; web pages and QA examples are different units. Use the v1.0 split table rather than merging the older project headline into it. The card retains email-evaluation instructions, while the manifest now includes test answers; no submission was made. Payload archives were not downloaded.\", \"splits\": [\"train\", \"dev\", \"test\"], \"supervision\": [\"extractive or yes/no answers\", \"answer DOM-element references\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/tree/7aa0bc6efc7ef43f68c192e2091108541acbaf1a\", \"gated\": false, \"hub_id\": \"X-LANCE/WebSRC_v1.0\", \"hub_revision\": \"7aa0bc6efc7ef43f68c192e2091108541acbaf1a\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0/raw/7aa0bc6efc7ef43f68c192e2091108541acbaf1a/README.md\"], \"reason\": \"Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.\"}}}\n\nPublic release includes train+dev and test ZIP packages plus a test-answer JSON file. The v1.0 card’s exact split table differs from the older project-page headline of 0.44M questions/6.5K pages; counts are kept with their sources. This is static webpage understanding, not an interactive browser-agent trajectory dataset. CC-BY-4.0 is the data-card declaration.\n\nPrimary source: https://huggingface.co/datasets/X-LANCE/WebSRC_v1.0\nCatalog record: https://awesomecua.github.io/resources/url-e0ba328ae1a643b0a5aa/"}
{"id": "url:e126bbd9729ebdaf05ab", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver", "citation_url": "https://awesomecua.github.io/resources/url-e126bbd9729ebdaf05ab/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md", "https://huggingface.co/api/datasets/CognitiveKernel/WebEvolver"], "formats": ["JSONL"], "modalities": [], "notes": "Two purpose-specific files are released: webevolver_policy_sft.jsonl and webevolver_world_model_sft.jsonl. They are not assumed disjoint train/test splits, and no modality or example count is inferred from filenames alone.", "splits": [], "supervision": ["agent-policy supervised fine-tuning", "web-world-model supervised fine-tuning"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/tree/13f4f45b359e138a45b1c89bd4e913298c586ca9", "gated": false, "hub_id": "CognitiveKernel/WebEvolver", "hub_revision": "13f4f45b359e138a45b1c89bd4e913298c586ca9", "license_review": {"data": {"evidence_urls": ["https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"], "reason": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content."}}, "repository_url": "https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:35:17Z", "role": "primary", "sha256": "36ccdce12c1b1512ad4b21326c5d0c0379c73e85f0b8150779d43240b54686de", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "notes", "relations", "licenses.data", "details.dataset_profile"], "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:33:29Z", "role": "primary", "sha256": "cbef3f8037405d311ed9fe8aa620ec3734f91c77ca8f4209fbe98167fb33bee8", "supports": ["identity", "release_status", "details.hub_id", "details.hub_revision", "details.gated", "details.disabled", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/api/datasets/CognitiveKernel/WebEvolver"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-18T04:27:14Z", "role": "primary", "sha256": "447dfa1e52789f7f3f111e9e0d453fb1f9fd8bf2d7b858f28fbe21a2239b4b2e", "supports": ["details.repository_url"], "url": "https://raw.githubusercontent.com/Tencent/SelfEvolvingAgent/HEAD/WebEvolver/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-18T04:35:17Z", "licenses": {"code": null, "data": {"basis": "Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.", "declared": "MIT", "retrieved_at": "2026-10-18T04:35:17Z", "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver"}], "name": "WebEvolver SFT data", "publication": null, "published_at": null, "record_sha256": "b1090fd953e745b44a5fdb253543b1a301f9b97000912ff456080ea85d8b1a2e", "relations": [{"evidence_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md", "target_id": "arxiv:2504.21024", "type": "documented_by"}], "resource_id": "url:e126bbd9729ebdaf05ab", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/CognitiveKernel/WebEvolver", "topics": ["web-navigation", "world-models", "imitation-learning"], "updated_at": null}, "text": "WebEvolver SFT data\nResource type: dataset\nReview status: curated\nScope: core\n\nTwo JSONL collections for supervised fine-tuning the WebEvolver agent policy and web world model.\nPlatforms: web\nTopics: web-navigation, world-models, imitation-learning\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md\", \"https://huggingface.co/api/datasets/CognitiveKernel/WebEvolver\"], \"formats\": [\"JSONL\"], \"modalities\": [], \"notes\": \"Two purpose-specific files are released: webevolver_policy_sft.jsonl and webevolver_world_model_sft.jsonl. They are not assumed disjoint train/test splits, and no modality or example count is inferred from filenames alone.\", \"splits\": [], \"supervision\": [\"agent-policy supervised fine-tuning\", \"web-world-model supervised fine-tuning\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/CognitiveKernel/WebEvolver/tree/13f4f45b359e138a45b1c89bd4e913298c586ca9\", \"gated\": false, \"hub_id\": \"CognitiveKernel/WebEvolver\", \"hub_revision\": \"13f4f45b359e138a45b1c89bd4e913298c586ca9\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://huggingface.co/datasets/CognitiveKernel/WebEvolver/raw/13f4f45b359e138a45b1c89bd4e913298c586ca9/README.md\"], \"reason\": \"Publisher dataset-card declaration for this release; it does not establish rights to all upstream screenshots, applications or source content.\"}}, \"repository_url\": \"https://github.com/Tencent/SelfEvolvingAgent/tree/main/WebEvolver\"}\n\nThe card and public manifest identify separate policy-model and world-model JSONL files. No row total or complete record schema is documented in the inspected card; the files were not downloaded as training samples. MIT is the publisher’s data declaration.\n\nPrimary source: https://huggingface.co/datasets/CognitiveKernel/WebEvolver\nCatalog record: https://awesomecua.github.io/resources/url-e126bbd9729ebdaf05ab/"}
{"id": "url:e21ca7c859544dbf5fb6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/microsoft/OmniParser-v2.0", "citation_url": "https://awesomecua.github.io/resources/url-e21ca7c859544dbf5fb6/", "details": {"disabled": false, "download_url": "https://huggingface.co/microsoft/OmniParser-v2.0/tree/main", "gated": false, "hub_id": "microsoft/OmniParser-v2.0", "license_review": {"weights": {"evidence_urls": ["https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE", "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"], "reason": "The top-level Hub metadata says MIT, but the same card and folder licenses distinguish the original detector from the new detector and captioner. Preserve one model-bundle identity with a mixed license declaration."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:04Z", "role": "primary", "sha256": "f4ebaafcc0ab0b9e9cc55cb587678e5c83a5bbd64469fb2cc52bc63843c5cc6d", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://raw.githubusercontent.com/microsoft/OmniParser/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "1ea8af0cfd9cd716de8af8b78b2451470e1e1f22233a7286c5f40e1f496dc4be", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/OmniParser-v2.0"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:03:41Z", "role": "primary", "sha256": "6ae154d65bfc2755703440be02acf63a37b741a1101b7c42b666da93a4ed2310", "supports": ["summary", "licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "4df3c306dddaaf4baffdff5ca820cc679ac8cd6dc263c6a74517783e42fa7a3b", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:06:57Z", "role": "primary", "sha256": "646f8936b8ddcd14e13e578ff6857e368780b0d1a4f6066bee89211923a373e2", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:09:13Z", "role": "primary", "sha256": "d9a1b1e30d633d5732ea18e3cba9538d293ebc53e1a9e4e96ab739e0c5c4f1cb", "supports": ["licenses.weights"], "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:36:04Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "Publisher per-folder model license declarations and retained LICENSE files; this is a component map, not one SPDX license for the full bundle.", "declared": "Mixed by component: icon_detect = AGPL-3.0; icon_detect_v3 = MIT; icon_caption = MIT", "retrieved_at": "2026-10-17T04:03:41Z", "url": "https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/OmniParser-v2.0"}], "name": "OmniParser v2 model bundle", "publication": null, "published_at": null, "record_sha256": "192b82102b2eea6385553e82fb286c48ccbbe88ecf285f89f0a1fc3514728ce5", "relations": [], "resource_id": "url:e21ca7c859544dbf5fb6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/microsoft/OmniParser-v2.0", "topics": ["screen-understanding", "grounding"], "updated_at": null}, "text": "OmniParser v2 model bundle\nResource type: model\nReview status: curated\nScope: core\n\nA screen-parsing weight bundle containing icon-region detectors and an icon captioner for converting screenshots into agent-readable interface elements.\nPlatforms: unspecified\nTopics: screen-understanding, grounding\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/microsoft/OmniParser-v2.0/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/OmniParser-v2.0\", \"license_review\": {\"weights\": {\"evidence_urls\": [\"https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/README.md\", \"https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect/LICENSE\", \"https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_detect_v3/LICENSE\", \"https://huggingface.co/microsoft/OmniParser-v2.0/raw/main/icon_caption/LICENSE\"], \"reason\": \"The top-level Hub metadata says MIT, but the same card and folder licenses distinguish the original detector from the new detector and captioner. Preserve one model-bundle identity with a mixed license declaration.\"}}}\n\nThe manifest contains icon_detect/model.pt, icon_detect_v3/model.pt and icon_caption/model.safetensors. Licenses differ by folder: original YOLOv8 detector AGPL-3.0; YOLOv9-E detector and Florence-2 captioner MIT. The card’s top-level MIT metadata must not cover the entire bundle. Components remain one resource; no binaries were executed.\n\nPrimary source: https://huggingface.co/microsoft/OmniParser-v2.0\nCatalog record: https://awesomecua.github.io/resources/url-e21ca7c859544dbf5fb6/"}
{"id": "url:e3c1a2a2d556e4728953", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/osunlp/Dreamer-7B", "citation_url": "https://awesomecua.github.io/resources/url-e3c1a2a2d556e4728953/", "details": {"base_model": "Qwen/Qwen2-VL-7B", "download_url": "https://huggingface.co/osunlp/Dreamer-7B/tree/main", "languages": ["en"]}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:10:45Z", "role": "primary", "sha256": "f15e05cc2718948e344bd1fb943402797ef2e3f2130a9ce2fd94d419210c0574", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:10:45Z", "role": "primary", "sha256": "f15e05cc2718948e344bd1fb943402797ef2e3f2130a9ce2fd94d419210c0574", "supports": ["licenses.weights"], "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T11:10:45Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-16T11:10:45Z", "url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/Dreamer-7B"}], "name": "Dreamer-7B", "publication": null, "published_at": null, "record_sha256": "d896407b608cf2fb112ef886eee4ec6cec49cc318f18efbc771f2567d6a27852", "relations": [{"evidence_url": "https://huggingface.co/osunlp/Dreamer-7B/raw/main/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "resource_id": "url:e3c1a2a2d556e4728953", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/osunlp/Dreamer-7B", "topics": ["world-models", "planning"], "updated_at": null}, "text": "Dreamer-7B\nResource type: model\nReview status: curated\nScope: core\n\nWebDreamer model checkpoint that predicts the consequences of web actions from screenshots and textual descriptions.\nPlatforms: web\nTopics: world-models, planning\n\nRecorded specifications: {\"base_model\": \"Qwen/Qwen2-VL-7B\", \"download_url\": \"https://huggingface.co/osunlp/Dreamer-7B/tree/main\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/osunlp/Dreamer-7B\nCatalog record: https://awesomecua.github.io/resources/url-e3c1a2a2d556e4728953/"}
{"id": "url:e3e5dd82a3102689c4ea", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL", "citation_url": "https://awesomecua.github.io/resources/url-e3e5dd82a3102689c4ea/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL/tree/main", "gated": false, "hub_id": "microsoft/GUI-Actor-7B-Qwen2-VL", "languages": [], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "c76d9811dc6152d1b33bec7f19e808c1882e86cca79bd3fea90075ac9df5c7e6", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "mit", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL"}], "name": "GUI-Actor 7B Qwen2-VL", "publication": null, "published_at": null, "record_sha256": "cdc03ee4d90c9866c8efc90a6098265920c99b1e1956a3cc0d0e10fc81ed1833", "relations": [{"evidence_url": "https://huggingface.co/api/models/microsoft/GUI-Actor-7B-Qwen2-VL", "target_id": "github:microsoft/gui-actor", "type": "variant_of"}], "resource_id": "url:e3e5dd82a3102689c4ea", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL", "topics": ["grounding"], "updated_at": null}, "text": "GUI-Actor 7B Qwen2-VL\nResource type: model\nReview status: curated\nScope: core\n\nGUI-Actor checkpoint fine-tuned from Qwen2-VL-7B-Instruct, with about 8.4B BF16 parameters, whose config names a pointer variant of the Qwen2-VL class rather than the stock one, under MIT.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL/tree/main\", \"gated\": false, \"hub_id\": \"microsoft/GUI-Actor-7B-Qwen2-VL\"}\n\nPrimary source: https://huggingface.co/microsoft/GUI-Actor-7B-Qwen2-VL\nCatalog record: https://awesomecua.github.io/resources/url-e3e5dd82a3102689c4ea/"}
{"id": "url:e554295a7f3f5c9ad0d6", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/osunlp/UGround-V1-7B", "citation_url": "https://awesomecua.github.io/resources/url-e554295a7f3f5c9ad0d6/", "details": {"artifact_type": "model-checkpoint", "disabled": false, "download_url": "https://huggingface.co/osunlp/UGround-V1-7B/tree/main", "gated": false, "hub_id": "osunlp/UGround-V1-7B", "languages": ["en"], "modalities": [], "size_category_reported": null}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:50:49Z", "role": "primary", "sha256": "080c9245276e9ffe5911551a66acd7e91b478f85b96c795d6079465d92580f18", "supports": ["details", "licenses.weights"], "url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-16T06:50:49Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card", "declared": "apache-2.0", "retrieved_at": "2026-10-16T06:50:49Z", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/osunlp/UGround-V1-7B"}], "name": "UGround V1 7B", "publication": null, "published_at": null, "record_sha256": "a826d915088c7d5ca975e9b620d42ca0aa867484ade3a60c7d5bddd49fd8595f", "relations": [{"evidence_url": "https://huggingface.co/api/models/osunlp/UGround-V1-7B", "target_id": "github:osu-nlp-group/uground", "type": "variant_of"}], "resource_id": "url:e554295a7f3f5c9ad0d6", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/osunlp/UGround-V1-7B", "topics": ["grounding"], "updated_at": null}, "text": "UGround V1 7B\nResource type: model\nReview status: curated\nScope: core\n\nUGround-V1 checkpoint fine-tuned from Qwen2-VL-7B, with about 8.3B BF16 parameters in the standard Qwen2-VL architecture, provided as safetensors and a PyTorch file under Apache-2.0.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_type\": \"model-checkpoint\", \"disabled\": false, \"download_url\": \"https://huggingface.co/osunlp/UGround-V1-7B/tree/main\", \"gated\": false, \"hub_id\": \"osunlp/UGround-V1-7B\", \"languages\": [\"en\"]}\n\nPrimary source: https://huggingface.co/osunlp/UGround-V1-7B\nCatalog record: https://awesomecua.github.io/resources/url-e554295a7f3f5c9ad0d6/"}
{"id": "url:eadc6cba9ea46c38dd1d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data", "citation_url": "https://awesomecua.github.io/resources/url-eadc6cba9ea46c38dd1d/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"], "formats": [], "modalities": ["images", "prompts", "actions", "responses"], "notes": "The card reports a cleaned size of 3.12M without defining the counting unit, so no sample total is inferred. The data card declares noncommercial share-alike terms, separately from the model-weight license.", "splits": [], "supervision": ["web-world-model prediction"]}, "download_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/tree/main", "modalities": ["image", "text"]}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "e656246f09340fed08136cf41a8b79880e19e1397ddd176410688ddee4e6365c", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "e656246f09340fed08136cf41a8b79880e19e1397ddd176410688ddee4e6365c", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:09:28Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "cc-by-nc-sa-4.0", "retrieved_at": "2026-10-16T11:09:28Z", "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data"}], "name": "Dreamer V1 Data", "publication": null, "published_at": null, "record_sha256": "407d71ea470d5a7df83472c9de0087a6a7bf05a7e647f9b5a9ba66df13bf9900", "relations": [{"evidence_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md", "target_id": "arxiv:2411.06559", "type": "documented_by"}], "resource_id": "url:eadc6cba9ea46c38dd1d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/osunlp/Dreamer-V1-Data", "topics": ["world-models", "data-generation"], "updated_at": null}, "text": "Dreamer V1 Data\nResource type: dataset\nReview status: curated\nScope: core\n\nWeb-world-model training data with prompts, images, action descriptions, and responses, released with the WebDreamer project.\nPlatforms: web\nTopics: world-models, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/raw/main/README.md\"], \"formats\": [], \"modalities\": [\"images\", \"prompts\", \"actions\", \"responses\"], \"notes\": \"The card reports a cleaned size of 3.12M without defining the counting unit, so no sample total is inferred. The data card declares noncommercial share-alike terms, separately from the model-weight license.\", \"splits\": [], \"supervision\": [\"web-world-model prediction\"]}, \"download_url\": \"https://huggingface.co/datasets/osunlp/Dreamer-V1-Data/tree/main\", \"modalities\": [\"image\", \"text\"]}\n\nThe card reports a cleaned size of 3.12M without explicitly defining its counting unit. The catalog leaves the exact sample count unknown. Its data license differs from the model-weight license.\n\nPrimary source: https://huggingface.co/datasets/osunlp/Dreamer-V1-Data\nCatalog record: https://awesomecua.github.io/resources/url-eadc6cba9ea46c38dd1d/"}
{"id": "url:edf38d43800be8c955ae", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/yiye2023/GUIChat", "citation_url": "https://awesomecua.github.io/resources/url-edf38d43800be8c955ae/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "single-turn portion", "qualifier": "approximately", "unit": "QA pairs", "value": 44000}, {"basis": "reported", "evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "population": "multi-turn portion", "qualifier": "approximately", "unit": "dialogues", "value": 6000}], "evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"], "formats": ["JSON", "Parquet"], "modalities": ["screenshots", "text", "bounding boxes"], "notes": "Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.", "splits": [], "supervision": ["single-turn question answering", "multi-turn dialogue"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/yiye2023/GUIChat/tree/main", "gated": false, "hub_id": "yiye2023/GUIChat", "license_review": {"data": {"evidence_urls": ["https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"], "reason": "Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict."}}}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:01Z", "role": "primary", "sha256": "0968273442bfb9a6478fe78f8b9a6247a3bb8e6d028880939b432fd009961e98", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details", "licenses.data"], "url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "3ccb4a9ded4f45aa334f0df5dc9a0f8963c645857e8442c46efe1d9cf4a4360b", "supports": ["identity", "release_status", "details"], "url": "https://huggingface.co/api/datasets/yiye2023/GUIChat"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "4bcf87ecfbbb8e07a01b21415a970c8b53a5283bf6872b657040d3f45c9241f7", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:36:01Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/yiye2023/GUIChat"}], "name": "GUIChat", "publication": null, "published_at": null, "record_sha256": "1816ec49d496637437817a5d950342e67eb06f71495d5e9c26b801fe247ebd05", "relations": [{"evidence_url": "https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md", "target_id": "arxiv:2406.11317", "type": "documented_by"}], "resource_id": "url:edf38d43800be8c955ae", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/yiye2023/GUIChat", "topics": ["grounding", "data-generation"], "updated_at": null}, "text": "GUIChat\nResource type: dataset\nReview status: curated\nScope: core\n\nConversational GUI supervision with text-rich images and bounding boxes, organized as single-turn questions and multi-turn dialogues.\nPlatforms: unspecified\nTopics: grounding, data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"population\": \"single-turn portion\", \"qualifier\": \"approximately\", \"unit\": \"QA pairs\", \"value\": 44000}, {\"basis\": \"reported\", \"evidence_url\": \"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"population\": \"multi-turn portion\", \"qualifier\": \"approximately\", \"unit\": \"dialogues\", \"value\": 6000}], \"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\"], \"formats\": [\"JSON\", \"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"bounding boxes\"], \"notes\": \"Quantities are rounded publisher descriptions, not downloaded sample counts. Data-license declarations conflict across first-party sources; no permissive license is inferred.\", \"splits\": [], \"supervision\": [\"single-turn question answering\", \"multi-turn dialogue\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/yiye2023/GUIChat/tree/main\", \"gated\": false, \"hub_id\": \"yiye2023/GUIChat\", \"license_review\": {\"data\": {\"evidence_urls\": [\"https://raw.githubusercontent.com/yiye3/GUICourse/HEAD/README.md\", \"https://huggingface.co/datasets/yiye2023/GUIChat/raw/main/README.md\"], \"reason\": \"Conflicting first-party dataset declarations: the GUICourse project README assigns CC-BY-4.0 to the datasets, while this exact Hub card declares Apache-2.0. The code MIT declaration does not resolve the data conflict.\"}}}\n\nPublic JSON annotations and Parquet image packages are present. The project README declares CC-BY-4.0 for datasets, while this Hub card declares Apache-2.0; these conflicting declarations are retained without asserting a single confirmed data license. The project code’s MIT declaration is separate.\n\nPrimary source: https://huggingface.co/datasets/yiye2023/GUIChat\nCatalog record: https://awesomecua.github.io/resources/url-edf38d43800be8c955ae/"}
{"id": "url:ef244017b1460269fc68", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://browser-use.com/", "citation_url": "https://awesomecua.github.io/resources/url-ef244017b1460269fc68/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "69e029d01dda3806f88abfaee01da08c671adcb3555f1314e975ab8942a686ae", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://browser-use.com/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://browser-use.com/"}], "name": "Browser Use Cloud", "publication": null, "published_at": null, "record_sha256": "d730789f90d0174473f0fb2c81727dbc8271a5cb1ad5e97e92bbfdac4ca2e24b", "relations": [{"evidence_url": "https://browser-use.com/", "target_id": "organization:browser-use", "type": "published_by"}], "resource_id": "url:ef244017b1460269fc68", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://browser-use.com/", "topics": ["web-navigation"], "updated_at": null}, "text": "Browser Use Cloud\nResource type: product\nReview status: curated\nScope: core\n\nHosted browser-agent platform from Browser Use.\nPlatforms: web\nTopics: web-navigation\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://browser-use.com/\nCatalog record: https://awesomecua.github.io/resources/url-ef244017b1460269fc68/"}
{"id": "url:f0584842138a52158f51", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://openai.com/index/introducing-operator/", "citation_url": "https://awesomecua.github.io/resources/url-f0584842138a52158f51/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:27Z", "role": "primary", "sha256": "474a27cf644ac5f8400a023086fb8d356e6148de71d1851337215f88624852d9", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://openai.com/index/introducing-operator/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:27Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://openai.com/index/introducing-operator/"}], "name": "Operator", "publication": null, "published_at": null, "record_sha256": "a80107b257838d4750b58d0b5b37fc5c70e5e6f7e55591f9124f9f907d0bfb59", "relations": [{"evidence_url": "https://openai.com/index/introducing-operator/", "target_id": "organization:openai", "type": "published_by"}], "resource_id": "url:f0584842138a52158f51", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://openai.com/index/introducing-operator/", "topics": ["web-navigation", "history"], "updated_at": null}, "text": "Operator\nResource type: product\nReview status: curated\nScope: core\n\nHistorical OpenAI browser-agent research preview.\nPlatforms: web\nTopics: web-navigation, history\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nHistorical product entry; the official page describes its integration into ChatGPT agent. Do not infer current standalone availability.\n\nPrimary source: https://openai.com/index/introducing-operator/\nCatalog record: https://awesomecua.github.io/resources/url-f0584842138a52158f51/"}
{"id": "url:f13474d7aa734a30624d", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1", "citation_url": "https://awesomecua.github.io/resources/url-f13474d7aa734a30624d/", "details": {"disabled": false, "download_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/tree/main", "gated": false, "hub_id": "LZXzju/Qwen2.5-VL-3B-UI-R1"}, "environments": ["web", "desktop", "android"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "6a206702a1a4faef428460e07f52aaf8fd1f1a144d711d807ca42479b248f697", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/LZXzju/Qwen2.5-VL-3B-UI-R1"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:36:00Z", "role": "primary", "sha256": "f55ac6a0c50d298d46c5f7083a0879959fd7cd3c1d91496af295ebf2afcc13a3", "supports": ["environments"], "url": "https://raw.githubusercontent.com/lll6gg/UI-R1/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:03:33Z", "role": "primary", "sha256": "620fbee20a6002bbc13562030ce4988d41ef17b0c033ec6872d83741f597adb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:03:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "mit", "retrieved_at": "2026-10-17T02:03:33Z", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1"}], "name": "UI-R1 3B", "publication": null, "published_at": null, "record_sha256": "b4ed755b4740419f200dcf13a5a72266ea4eb0866cb6f0d2424358d4e3dd90eb", "relations": [{"evidence_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/raw/main/README.md", "target_id": "arxiv:2503.21620", "type": "documented_by"}], "resource_id": "url:f13474d7aa734a30624d", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1", "topics": ["grounding", "reinforcement-learning", "action-prediction"], "updated_at": null}, "text": "UI-R1 3B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL checkpoint trained with reinforcement learning for GUI action prediction and evaluated on mobile, desktop and web grounding tasks.\nPlatforms: web, desktop, android\nTopics: grounding, reinforcement-learning, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1/tree/main\", \"gated\": false, \"hub_id\": \"LZXzju/Qwen2.5-VL-3B-UI-R1\"}\n\nTwo safetensors shards and tokenizer/configuration files are publicly ungated. The model card declares MIT; upstream base-model terms are separate. The mutable card compares v1, v2 and UI-R1-E results without pinning every result to a commit, so this record identifies the Hub release and does not import an unversioned score.\n\nPrimary source: https://huggingface.co/LZXzju/Qwen2.5-VL-3B-UI-R1\nCatalog record: https://awesomecua.github.io/resources/url-f13474d7aa734a30624d/"}
{"id": "url:f35cfcbaeb79663eabeb", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/Hcompany/Holo1-7B", "citation_url": "https://awesomecua.github.io/resources/url-f35cfcbaeb79663eabeb/", "details": {"disabled": false, "download_url": "https://huggingface.co/Hcompany/Holo1-7B/tree/main", "gated": false, "hub_id": "Hcompany/Holo1-7B"}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "66076674ebbdcf9faeb408ebd137ac1283fb56b1011acd7037ebb73e09ba749d", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/Hcompany/Holo1-7B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "a7e185380bb02150d761651420cccc9c4a57f67c70a5dec3b7f85af63fb8773b", "supports": ["licenses.weights"], "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:44:50Z", "url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/Hcompany/Holo1-7B"}], "name": "Holo1 7B", "publication": null, "published_at": null, "record_sha256": "2cffca5be8d8f10d74a463a01acf7f840e13254c7f00cdb3b0266da62308f988", "relations": [{"evidence_url": "https://huggingface.co/Hcompany/Holo1-7B/raw/main/README.md", "target_id": "arxiv:2506.02865", "type": "documented_by"}], "resource_id": "url:f35cfcbaeb79663eabeb", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/Hcompany/Holo1-7B", "topics": ["web-navigation", "grounding", "planning"], "updated_at": null}, "text": "Holo1 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-based action model released for the Surfer-H web-agent system, supporting policy, interface-localization and validation roles.\nPlatforms: web\nTopics: web-navigation, grounding, planning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/Hcompany/Holo1-7B/tree/main\", \"gated\": false, \"hub_id\": \"Hcompany/Holo1-7B\"}\n\nFour public safetensors shards and example navigation/localization files are present. The model card declares Apache-2.0. Reported Surfer-H results depend on the assembled policy/localizer/validator configuration and are not imported as standalone checkpoint scores.\n\nPrimary source: https://huggingface.co/Hcompany/Holo1-7B\nCatalog record: https://awesomecua.github.io/resources/url-f35cfcbaeb79663eabeb/"}
{"id": "url:f70aa7c95824a8729516", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/showlab/ShowUI-pi", "citation_url": "https://awesomecua.github.io/resources/url-f70aa7c95824a8729516/", "details": {"disabled": false, "download_url": "https://huggingface.co/showlab/ShowUI-pi/tree/main", "gated": false, "hub_id": "showlab/ShowUI-pi"}, "environments": ["unspecified"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "64993ab5dc8205e25bda2576055b7e252104c67d4e6eeecb33a1a9742415f67e", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:50Z", "role": "primary", "sha256": "ae12ece91946d79c5c6b578d6db07b5f095822be1ca3bc14bb3dc04a771bb010", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/showlab/ShowUI-pi"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:44:50Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/showlab/ShowUI-pi"}], "name": "ShowUI-π checkpoint", "publication": null, "published_at": null, "record_sha256": "61856215c8b9a8bfd6f327582ad092f110e8b2dd49ef40df5ebde9272e149b4d", "relations": [{"evidence_url": "https://huggingface.co/showlab/ShowUI-pi/raw/main/README.md", "target_id": "arxiv:2512.24965", "type": "documented_by"}], "resource_id": "url:f70aa7c95824a8729516", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/showlab/ShowUI-pi", "topics": ["grounding", "action-prediction"], "updated_at": null}, "text": "ShowUI-π checkpoint\nResource type: model\nReview status: curated\nScope: core\n\nA compact visual-action checkpoint that predicts continuous GUI drag trajectories from a screenshot and a natural-language instruction.\nPlatforms: unspecified\nTopics: grounding, action-prediction\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/showlab/ShowUI-pi/tree/main\", \"gated\": false, \"hub_id\": \"showlab/ShowUI-pi\"}\n\nA model.safetensors file and policy pre/postprocessor configuration are publicly ungated. The card describes a SmolVLA-based GUI drag-and-drop model despite its generic robotics pipeline tag. No weight-license declaration or specific operating-system coverage is provided.\n\nPrimary source: https://huggingface.co/showlab/ShowUI-pi\nCatalog record: https://awesomecua.github.io/resources/url-f70aa7c95824a8729516/"}
{"id": "url:f8566bd0394d36ed4f8e", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data", "citation_url": "https://awesomecua.github.io/resources/url-f8566bd0394d36ed4f8e/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "instructions", "value": 2900000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "web subset", "qualifier": "approximately", "unit": "images", "value": 173000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "instructions", "value": 1100000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "mobile subset", "qualifier": "approximately", "unit": "images", "value": 104000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "instructions", "value": 150000}, {"basis": "reported", "evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "population": "desktop subset", "qualifier": "approximately", "unit": "images", "value": 7800}], "evidence_urls": ["https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"], "formats": ["JSON annotations", "image archives"], "modalities": ["screenshots", "instructions", "element captions"], "notes": "The card identifies deprecated desktop files with screenshot-mapping errors and names corrected releases. Counts are rounded publisher figures for separate platform subsets; instructions and images are not interchangeable.", "splits": [], "supervision": ["element grounding", "element descriptions"]}, "download_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/tree/main", "modalities": ["image", "text"]}, "environments": ["web", "android", "desktop", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-16T10:42:13Z", "role": "primary", "sha256": "8713ca2a6b1962d5dce7da132e1135898d467053dc6b53b4463a0d43a6433bcb", "supports": ["licenses.data"], "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T10:42:13Z", "licenses": {"code": null, "data": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-16T10:42:13Z", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md"}, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data"}], "name": "Aria-UI Data", "publication": null, "published_at": null, "record_sha256": "7878d12e8697b59eee82ceddb56e2f86f22388dd314ed8cbbee4ab1f37c3a902", "relations": [{"evidence_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md", "target_id": "arxiv:2412.16256", "type": "documented_by"}], "resource_id": "url:f8566bd0394d36ed4f8e", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/Aria-UI/Aria-UI_Data", "topics": ["grounding", "screen-understanding"], "updated_at": null}, "text": "Aria-UI Data\nResource type: dataset\nReview status: curated\nScope: core\n\nGUI element descriptions and grounding instructions spanning web, mobile, and Ubuntu desktop interfaces.\nPlatforms: web, android, desktop, linux\nTopics: grounding, screen-understanding\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"web subset\", \"qualifier\": \"approximately\", \"unit\": \"instructions\", \"value\": 2900000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"web subset\", \"qualifier\": \"approximately\", \"unit\": \"images\", \"value\": 173000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"mobile subset\", \"qualifier\": \"approximately\", \"unit\": \"instructions\", \"value\": 1100000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"mobile subset\", \"qualifier\": \"approximately\", \"unit\": \"images\", \"value\": 104000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"desktop subset\", \"qualifier\": \"approximately\", \"unit\": \"instructions\", \"value\": 150000}, {\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\", \"population\": \"desktop subset\", \"qualifier\": \"approximately\", \"unit\": \"images\", \"value\": 7800}], \"evidence_urls\": [\"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/raw/main/README.md\"], \"formats\": [\"JSON annotations\", \"image archives\"], \"modalities\": [\"screenshots\", \"instructions\", \"element captions\"], \"notes\": \"The card identifies deprecated desktop files with screenshot-mapping errors and names corrected releases. Counts are rounded publisher figures for separate platform subsets; instructions and images are not interchangeable.\", \"splits\": [], \"supervision\": [\"element grounding\", \"element descriptions\"]}, \"download_url\": \"https://huggingface.co/datasets/Aria-UI/Aria-UI_Data/tree/main\", \"modalities\": [\"image\", \"text\"]}\n\nThe publisher flags original desktop files as deprecated because of screenshot-mapping issues, and identifies corrected versions in the dataset card.\n\nPrimary source: https://huggingface.co/datasets/Aria-UI/Aria-UI_Data\nCatalog record: https://awesomecua.github.io/resources/url-f8566bd0394d36ed4f8e/"}
{"id": "url:f976dec3d1c5e0bedbd7", "metadata": {"aliases": ["url:a6db151f939a6748ff76", "showlab/ShowUI-desktop", "https://huggingface.co/datasets/showlab/ShowUI-desktop"], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K", "citation_url": "https://awesomecua.github.io/resources/url-f976dec3d1c5e0bedbd7/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md", "population": "train split declared in the dataset card", "qualifier": "exact", "unit": "examples", "value": 7496}], "evidence_urls": ["https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md"], "formats": ["Hub table with an image column"], "modalities": ["screenshots", "text", "bounding boxes", "points"], "notes": "The 8K release name is not an exact record count; the card declares 7,496 examples. Images and original annotations come from OmniAct and are augmented with generated descriptions. Examples are not asserted to be unique screenshots.", "splits": ["train"], "supervision": ["instruction-to-element grounding"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/tree/main", "gated": false, "hub_id": "showlab/ShowUI-desktop", "languages": [], "modalities": ["image", "text", "timeseries"], "size_category_reported": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "84bd2ef05ea837eaeebb248002b2942dd5c72f072beb29a9bd70eda1ea433a31", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:02:38Z", "role": "primary", "sha256": "577a78efa03cde05141c23033222014094681cb5d04786dbb4a698df8543b3df", "supports": ["details.dataset_profile", "relations"], "url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-16T06:38:26Z", "role": "primary", "sha256": "84bd2ef05ea837eaeebb248002b2942dd5c72f072beb29a9bd70eda1ea433a31", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:57Z", "role": "primary", "sha256": "57eaa720c20e9e3a4428c8292c150e1949f3372c75c6890e424ba3c6600a5e5d", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop"}, {"method": "huggingface_repository_identity", "retrieved_at": "2026-10-17T19:14:58Z", "role": "primary", "sha256": "57eaa720c20e9e3a4428c8292c150e1949f3372c75c6890e424ba3c6600a5e5d", "supports": ["aliases"], "url": "https://huggingface.co/api/datasets/showlab/ShowUI-desktop-8K"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T06:38:26Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K"}], "name": "ShowUI Desktop", "publication": null, "published_at": null, "record_sha256": "b413c80d4d5e53af7ef86311f6d4d9e753ca9b54540232cb4bb42071674b6cea", "relations": [{"evidence_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md", "target_id": "github:showlab/showui", "type": "component_of"}], "resource_id": "url:f976dec3d1c5e0bedbd7", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/showlab/ShowUI-desktop-8K", "topics": ["data-generation"], "updated_at": null}, "text": "ShowUI Desktop\nResource type: dataset\nReview status: curated\nScope: core\n\n7,496 desktop grounding examples built from OmniAct screenshots, with annotations rewritten by GPT-4o to describe element appearance, spatial relations and function.\nRecorded aliases: url:a6db151f939a6748ff76, showlab/ShowUI-desktop, https://huggingface.co/datasets/showlab/ShowUI-desktop\nPlatforms: desktop\nTopics: data-generation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md\", \"population\": \"train split declared in the dataset card\", \"qualifier\": \"exact\", \"unit\": \"examples\", \"value\": 7496}], \"evidence_urls\": [\"https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/raw/main/README.md\"], \"formats\": [\"Hub table with an image column\"], \"modalities\": [\"screenshots\", \"text\", \"bounding boxes\", \"points\"], \"notes\": \"The 8K release name is not an exact record count; the card declares 7,496 examples. Images and original annotations come from OmniAct and are augmented with generated descriptions. Examples are not asserted to be unique screenshots.\", \"splits\": [\"train\"], \"supervision\": [\"instruction-to-element grounding\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/showlab/ShowUI-desktop-8K/tree/main\", \"gated\": false, \"hub_id\": \"showlab/ShowUI-desktop\", \"modalities\": [\"image\", \"text\", \"timeseries\"]}\n\nPrimary source: https://huggingface.co/datasets/showlab/ShowUI-desktop-8K\nCatalog record: https://awesomecua.github.io/resources/url-f976dec3d1c5e0bedbd7/"}
{"id": "url:fcc98b149272bbf31b05", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation", "citation_url": "https://awesomecua.github.io/resources/url-fcc98b149272bbf31b05/", "details": {"dataset_profile": {"counts": [{"basis": "reported", "evidence_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md", "population": "Hub split named test", "qualifier": "exact", "unit": "rows", "value": 3630}], "evidence_urls": ["https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md", "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"], "formats": ["Parquet"], "modalities": ["screenshots", "text", "action history"], "notes": "This is the checklist-generation release rather than the separate preference-pair collection. No unique-task count or data permission is inferred.", "splits": ["test"], "supervision": ["checklist generation", "subgoal analysis"]}, "disabled": false, "download_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/tree/main", "gated": false, "hub_id": "LangAGI-Lab/WebPRMCollection_checklist_generation", "languages": [], "modalities": ["image", "text"], "size_category_reported": null}, "environments": ["web"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "2e4142675f2cb994c21a7620a5188b10147ab03326a2047946886b5916789ec2", "supports": ["name", "kind", "identity", "release_status", "details", "summary"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "3693cb307065a6ced63a4789a7f592b5eba0ad6a5ffe1a3caf0649decd5cb6e8", "supports": ["details"], "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:05Z", "role": "primary", "sha256": "647d9e7bdad5fe2c03a8cfcd34d93f3b2bf7c1553b517a4a843bc0113fa47456", "supports": ["identity", "summary"], "url": "https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:44:04Z", "role": "primary", "sha256": "07431058d8b9a71b80e75b7a74a40a8346c1503eec240628459cc4d4e2f0802f", "supports": ["summary", "environments", "topics", "scope"], "url": "https://raw.githubusercontent.com/kyle8581/Web-Shepherd/HEAD/README.md"}, {"method": "huggingface_api", "retrieved_at": "2026-10-17T04:01:33Z", "role": "primary", "sha256": "2e4142675f2cb994c21a7620a5188b10147ab03326a2047946886b5916789ec2", "supports": ["details", "licenses.data"], "url": "https://huggingface.co/api/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-17T04:01:33Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation"}], "name": "WebPRM checklist-generation data", "publication": null, "published_at": null, "record_sha256": "edbf2e7b2e469d43c065ae151e69c492744af56fc47873446ab74b66d51f3758", "relations": [], "resource_id": "url:fcc98b149272bbf31b05", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation", "topics": ["reward-models", "data-generation", "web-navigation"], "updated_at": null}, "text": "WebPRM checklist-generation data\nResource type: dataset\nReview status: curated\nScope: core\n\nWeb-task context, starting observations and action history paired with structured checklists and subgoal analyses for checklist generation.\nPlatforms: web\nTopics: reward-models, data-generation, web-navigation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [{\"basis\": \"reported\", \"evidence_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md\", \"population\": \"Hub split named test\", \"qualifier\": \"exact\", \"unit\": \"rows\", \"value\": 3630}], \"evidence_urls\": [\"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/raw/main/README.md\", \"https://huggingface.co/api/collections/LangAGI-Lab/web-shepherd-advancing-prms-for-reinforcing-web-agents-682b4f4ad607fc27c4dc49e8\"], \"formats\": [\"Parquet\"], \"modalities\": [\"screenshots\", \"text\", \"action history\"], \"notes\": \"This is the checklist-generation release rather than the separate preference-pair collection. No unique-task count or data permission is inferred.\", \"splits\": [\"test\"], \"supervision\": [\"checklist generation\", \"subgoal analysis\"]}, \"disabled\": false, \"download_url\": \"https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation/tree/main\", \"gated\": false, \"hub_id\": \"LangAGI-Lab/WebPRMCollection_checklist_generation\", \"modalities\": [\"image\", \"text\"]}\n\nFive public Parquet shards are present. This collection provides checklist-generation supervision and is distinct from the previously cataloged chosen/rejected preference data. The Hub calls its split test; that label is preserved. No data license is declared.\n\nPrimary source: https://huggingface.co/datasets/LangAGI-Lab/WebPRMCollection_checklist_generation\nCatalog record: https://awesomecua.github.io/resources/url-fcc98b149272bbf31b05/"}
{"id": "url:fda52d641c6b482bda7b", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://e2b.dev/", "citation_url": "https://awesomecua.github.io/resources/url-fda52d641c6b482bda7b/", "details": {"access": "publisher-managed", "availability": "not independently tested", "pricing_url": null}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T06:38:28Z", "role": "primary", "sha256": "6232364481bbae45a3fcec1d1aca858493baff0693ececa98b09b610183e6796", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://e2b.dev/"}], "identifiers": {}, "kind": "product", "last_verified_at": "2026-10-16T06:38:28Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://e2b.dev/"}], "name": "E2B", "publication": null, "published_at": null, "record_sha256": "7f43141806559e5641afb08119448b229e5c9821365dc166bc3499ea804fccbd", "relations": [{"evidence_url": "https://e2b.dev/", "target_id": "organization:e2b", "type": "published_by"}], "resource_id": "url:fda52d641c6b482bda7b", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://e2b.dev/", "topics": ["sandbox"], "updated_at": null}, "text": "E2B\nResource type: product\nReview status: curated\nScope: core\n\nSandbox infrastructure for executing agent workloads.\nPlatforms: desktop\nTopics: sandbox\n\nRecorded specifications: {\"access\": \"publisher-managed\", \"availability\": \"not independently tested\"}\n\nPrimary source: https://e2b.dev/\nCatalog record: https://awesomecua.github.io/resources/url-fda52d641c6b482bda7b/"}
{"id": "url:fe43af1801032be708cc", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag", "citation_url": "https://awesomecua.github.io/resources/url-fe43af1801032be708cc/", "details": {"dataset_profile": {"counts": [], "evidence_urls": ["https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"], "formats": ["Parquet", "episode files"], "modalities": ["screen images", "instructions", "previous state", "continuous actions"], "notes": "Raw and Parquet folders are alternative views of the same data. Each sample describes a dragging task over multiple timesteps; action vectors encode button state and coordinates.", "splits": ["train", "test"], "supervision": ["continuous GUI dragging"]}, "download_url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/tree/main", "modalities": ["image", "text", "continuous-actions"]}, "environments": ["desktop"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:12:56Z", "role": "primary", "sha256": "39c7c028f7ef5ca5c75f79b59a107c7e25444776095a2a08bb7535c4f523f7fa", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details.modalities", "details.download_url", "details.dataset_profile"], "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-16T11:09:28Z", "role": "primary", "sha256": "88df552ccbe80efd7020905d77d526f4f9efe2d09a11fed2714b083988f9dcef", "supports": ["relations"], "url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md"}], "identifiers": {}, "kind": "dataset", "last_verified_at": "2026-10-16T11:12:56Z", "licenses": {"code": null, "data": null, "document": null, "weights": null}, "links": [{"rel": "primary", "url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag"}], "name": "ScreenDrag", "publication": null, "published_at": null, "record_sha256": "3377ce54e2db717b2a0223e3f59390ff8073de4c572b152607e4e96ec488f27d", "relations": [{"evidence_url": "https://raw.githubusercontent.com/showlab/showui-pi/HEAD/README.md", "target_id": "github:showlab/showui-pi", "type": "component_of"}], "resource_id": "url:fe43af1801032be708cc", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/datasets/h-siyuan/ScreenDrag", "topics": ["grounding", "evaluation"], "updated_at": null}, "text": "ScreenDrag\nResource type: dataset\nReview status: curated\nScope: core\n\nGUI dragging data with screen observations, prior state, and continuous pointer actions, provided as trajectory files and Parquet representations.\nPlatforms: desktop\nTopics: grounding, evaluation\n\nRecorded specifications: {\"dataset_profile\": {\"counts\": [], \"evidence_urls\": [\"https://huggingface.co/datasets/h-siyuan/ScreenDrag/raw/main/README.md\"], \"formats\": [\"Parquet\", \"episode files\"], \"modalities\": [\"screen images\", \"instructions\", \"previous state\", \"continuous actions\"], \"notes\": \"Raw and Parquet folders are alternative views of the same data. Each sample describes a dragging task over multiple timesteps; action vectors encode button state and coordinates.\", \"splits\": [\"train\", \"test\"], \"supervision\": [\"continuous GUI dragging\"]}, \"download_url\": \"https://huggingface.co/datasets/h-siyuan/ScreenDrag/tree/main\", \"modalities\": [\"image\", \"text\", \"continuous-actions\"]}\n\nRaw and Parquet folders are alternative views of the same data and should not be counted as independent samples. The inspected card did not state a license.\n\nPrimary source: https://huggingface.co/datasets/h-siyuan/ScreenDrag\nCatalog record: https://awesomecua.github.io/resources/url-fe43af1801032be708cc/"}
{"id": "url:fec1c5aa8fd93cb78923", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B", "citation_url": "https://awesomecua.github.io/resources/url-fec1c5aa8fd93cb78923/", "details": {"artifact_inspection": "Public file-manifest inspection only; weights were not downloaded.", "base_model": "Qwen/Qwen3.5-2B", "gated": false, "hub_revision": "21e684b39c33e7efaf055f000374d74f37556954", "listed_weight_files": ["model.safetensors", "phase1_mask_decoder.safetensors"]}, "environments": ["cross-platform"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:49:48Z", "role": "primary", "sha256": "7b3242c3943537b6a26d8650a78c45044dce06a031ca2ce8bc67d5e482a332ce", "supports": ["details.gated", "details.hub_revision", "details.listed_weight_files"], "url": "https://huggingface.co/api/models/jcwang0602/Seg2Click-Qwen3.5-2B"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T01:48:10Z", "role": "primary", "sha256": "08d14b9dc3010365e543cbb6707b67c5d792118451bbc2d0b70f1f0280cb20e1", "supports": ["licenses.weights"], "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T01:48:10Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T01:48:10Z", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B"}], "name": "Seg2Click Qwen3.5 2B", "publication": null, "published_at": null, "record_sha256": "6f45d56148a8d4833864625e63b1bd3fd103ef84e2df9bd03dd27601e0cf2ace", "relations": [{"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "github:jcwang0602/seg2click", "type": "variant_of"}, {"evidence_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B/raw/main/README.md", "target_id": "arxiv:2610.16234", "type": "documented_by"}], "resource_id": "url:fec1c5aa8fd93cb78923", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B", "topics": ["grounding"], "updated_at": null}, "text": "Seg2Click Qwen3.5 2B\nResource type: model\nReview status: curated\nScope: core\n\nSeg2Click checkpoint using a 2B Qwen3.5 backbone to predict a mask and select an interface click.\nPlatforms: cross-platform\nTopics: grounding\n\nRecorded specifications: {\"artifact_inspection\": \"Public file-manifest inspection only; weights were not downloaded.\", \"base_model\": \"Qwen/Qwen3.5-2B\", \"gated\": false, \"hub_revision\": \"21e684b39c33e7efaf055f000374d74f37556954\", \"listed_weight_files\": [\"model.safetensors\", \"phase1_mask_decoder.safetensors\"]}\n\nThe public Hub manifest lists weight files; the catalog did not download weights or reproduce inference. License declarations apply to this checkpoint separately from code and training data. The checkpoint includes a merged backbone and mask decoder; a generic text-generation pipeline alone does not implement mask-to-click inference.\n\nPrimary source: https://huggingface.co/jcwang0602/Seg2Click-Qwen3.5-2B\nCatalog record: https://awesomecua.github.io/resources/url-fec1c5aa8fd93cb78923/"}
{"id": "url:ff7b1873a3eead89607a", "metadata": {"aliases": [], "arxiv_snapshot": null, "authors": [], "canonical_source_url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B", "citation_url": "https://awesomecua.github.io/resources/url-ff7b1873a3eead89607a/", "details": {"disabled": false, "download_url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/tree/main", "gated": false, "hub_id": "OpenGVLab/ScaleCUA-7B"}, "environments": ["web", "desktop", "android", "linux"], "evidence": [{"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["name", "kind", "summary", "environments", "topics", "scope", "relations", "details"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T02:45:19Z", "role": "primary", "sha256": "bad2617fd64633212a30b78d99fb053a662378ab50b42722a04af4104ac2248f", "supports": ["identity", "release_status", "details", "licenses.weights"], "url": "https://huggingface.co/api/models/OpenGVLab/ScaleCUA-7B"}, {"method": "editorial_primary_review", "retrieved_at": "2026-10-17T01:54:55Z", "role": "primary", "sha256": "0538701c860afedd5d63c3839039699d8e5a14977d5c339b187ed13ad666c8fa", "supports": ["environments", "relations"], "url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md"}, {"method": "huggingface_card", "retrieved_at": "2026-10-17T02:45:20Z", "role": "primary", "sha256": "e45116466b8b9ef3e93992a1efca8fa5b921a8b041297933210c9f552767ffb8", "supports": ["licenses.weights"], "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}], "identifiers": {}, "kind": "model", "last_verified_at": "2026-10-17T02:45:20Z", "licenses": {"code": null, "data": null, "document": null, "weights": {"basis": "publisher model/dataset card metadata; third-party assets retain their own terms", "declared": "apache-2.0", "retrieved_at": "2026-10-17T02:45:20Z", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B/raw/main/README.md"}}, "links": [{"rel": "primary", "url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B"}], "name": "ScaleCUA 7B", "publication": null, "published_at": null, "record_sha256": "1695d6f065c579857ec7cf7f0c86b9b9c35ec2797c8b6396ccaed961b11e15fb", "relations": [{"evidence_url": "https://raw.githubusercontent.com/OpenGVLab/ScaleCUA/HEAD/README.md", "target_id": "arxiv:2509.15221", "type": "documented_by"}], "resource_id": "url:ff7b1873a3eead89607a", "review_status": "curated", "schema_version": 1, "scope": "core", "source_url": "https://huggingface.co/OpenGVLab/ScaleCUA-7B", "topics": ["grounding", "action-prediction", "planning"], "updated_at": null}, "text": "ScaleCUA 7B\nResource type: model\nReview status: curated\nScope: core\n\nA Qwen2.5-VL-based computer-use checkpoint with direct grounding and agent action modes across desktop, mobile and web interfaces.\nPlatforms: web, desktop, android, linux\nTopics: grounding, action-prediction, planning\n\nRecorded specifications: {\"disabled\": false, \"download_url\": \"https://huggingface.co/OpenGVLab/ScaleCUA-7B/tree/main\", \"gated\": false, \"hub_id\": \"OpenGVLab/ScaleCUA-7B\"}\n\nFour public safetensors shards and model/tokenizer configuration are present. The card declares Apache-2.0. The model card’s Paper link points to the code repository, so the arXiv association uses the repository’s explicit citation rather than treating that link as a paper URL.\n\nPrimary source: https://huggingface.co/OpenGVLab/ScaleCUA-7B\nCatalog record: https://awesomecua.github.io/resources/url-ff7b1873a3eead89607a/"}